The agency pitch is theatre. Buy the team, not the show.

The people who win the pitch are rarely the people who run the account. If you are choosing an agency on the room, you are choosing wrong.

I have sat on both sides of the pitch table. On the agency side, the pitch team is a specialist unit. Senior, rehearsed, brilliant in a room. On the client side, the committee scores the room. Then the account starts and a different, younger team shows up.

That is not deceit. It is how the industry is built. But it means the thing you are evaluating is not the thing you are buying.

What to evaluate instead

Ask for the named account team, with hours per month by person, before the final round. Put those names in the contract. Score the pitch on the working session, not the presentation: give them a real problem, sit in the room while they work it, and watch who actually thinks.

Then benchmark the financial proposal against the staffing plan. The fee should be explainable from the people on the account and their market rates. If it is not, ask why.

The part most clients skip

Reference the account team, not the agency. Call two clients the same people have served in the last three years. Ask what happened after month six.

A good search is a procurement decision with a creative component, not a beauty parade. Run it that way and you will keep the agency for five years instead of two.

How much should a mid-market company pay its media agency?

There is no single right number, but there is a right way to get to one. Here is how agency fees are actually built, where the hidden margin sits, and what a fair deal looks like in 2026.

Every CMO and CFO asks this at some point, usually right after seeing the agency invoice next to the media plan and noticing the fee has not moved while the plan has shrunk. The honest answer is that it depends on scope, spend, and how the agency makes money from you beyond the fee. But "it depends" is not useful, so here is the actual structure.

The three ways a media agency gets paid

A media agency typically earns from you in three places, and only one of them is on the invoice.

The first is the fee: a retainer, a percentage of spend, a fixed project cost, or a blend. This is the visible part. For a mid-market brand spending between $5M and $50M a year on media, the fee usually lands somewhere between 4% and 12% of media spend, depending on how much planning and strategy is bundled in, how many channels are involved, and how senior the team is.

The second is the agency's technology and trading margin. Programmatic media passes through the agency's platform seat, its data partnerships, and sometimes its own trading desk. Each of those can carry a margin you do not see as a separate line. In audits we regularly find this layer adds 10 to 25% on top of the working media cost.

The third is rebates and incentives from media owners and platforms. Depending on your contract, some, all, or none of this comes back to you.

When executives compare fees, they compare the first one. The second and third are where the real money is.

What a fair fee looks like

Start from the work, not the percentage. Ask the agency for a staffing plan: who is on the account, at what level, for how many hours a month. Price that against market hourly rates by role and region. That gives you a cost-of-service number. A fair fee is that number plus a reasonable agency margin, typically 15 to 25%.

If the fee the agency proposed is well above that, one of three things is true. The team is more senior than you think, the scope is bigger than you think, or the fee is padded. It is worth finding out which.

Then deal with the invisible layers. Your contract should give you full transparency on platform and technology fees, the right to audit, ownership of your log-level data, and clarity on rebates. Without those clauses, the fee negotiation is a sideshow.

What has changed since 2023

AI has cut the time agencies spend on planning, reporting, and a large part of production. Most agencies have kept those gains. If your fee was set before that shift, it reflects a cost base that no longer exists. That is not an accusation. It is a reason to reopen the conversation.

The other change is that mid-market brands now have access to independent benchmarks and audit capability that used to be reserved for global advertisers. You can see what you are paying for. Most agencies are prepared for that conversation. The ones that are not tend to be the ones with something to protect.

A practical sequence

Pull the contract, the last 12 months of invoices, and the media plans. Ask for the staffing plan. Benchmark the rates. Request platform fee transparency. Then sit down with the agency and reset the fee to the work, with a model that fits how you actually operate.

Done properly, this takes six to eight weeks and does not damage the relationship. Agencies respect clients who understand the economics. What they exploit, quietly and without malice, is clients who do not.

What a decade of selling to CMOs taught me about how marketing money is actually spent

I spent years on the vendor side of the table, inside ad tech and media businesses. Here is what that vantage point shows you about where a marketing budget really goes.

Before I ran a consultancy, I ran revenue for technology companies that sold to marketers. Demand-side platforms, video, data. Our customers were agencies and brands. Our job was to get as much of the media budget flowing through our pipes as possible, at as high a margin as the market would bear.

Nothing about that is sinister. It is what every vendor does. But it gives you an education that most marketing leaders never get, because they only ever see the vendor from the outside.

Three things the vendor side teaches you

The rate card is a starting position. Every platform and every agency has a list price and a floor. The distance between them is the negotiation. Clients who know the floor exists pay a fraction of what clients who do not pay. Most do not.

Margin hides in layers. Fees are not taken in one place. They are taken in six small places: platform fee, data fee, technology fee, trading margin, rebate retained, and the fee on the fee. Each one is defensible on its own. Together they can be a third of the working budget.

Incentives are aligned to volume, not outcomes. A platform is paid on spend. An agency on a percentage of spend is paid on spend. Nobody in the chain is paid for the client spending less and getting more. That job has to belong to the client, or to someone the client hires.

What I would tell a CMO

You are not being cheated. You are being sold to by professionals, and the commercial side of the relationship is not where your expertise lies, nor should it be. Your job is the brand, the growth, the team.

But someone on your side of the table has to know the floor, see the layers, and hold the incentives straight. If that is not you, and it is not procurement, and it is not the agency, then it is nobody. And nobody is expensive.

The upside

When the commercial side is run properly, the marketing gets better. Money that was leaking into intermediaries goes back into working media. Senior agency people come back onto the account because the contract asks for them. Measurement gets honest because the model is not being marked by the people it grades.

That is the whole thesis of the work we do. It came from having sold the other side of it.

What does operational due diligence actually cover?

Financial diligence tests the numbers. Commercial diligence tests the market. Operational diligence tests whether the business can deliver the plan, and where the cost base leaks. Here is what a proper one covers.

Most mid-market deals get financial diligence, legal diligence, and some form of commercial review. Operational diligence is often lighter, done late, or folded into a management presentation. That is a mistake, because the operating cost base is where the value creation plan will succeed or fail.

The question it answers

Financial diligence asks: are the numbers real? Commercial diligence asks: will the market support the plan? Operational diligence asks: can this business actually execute, and what is it spending that it does not need to?

That second half is where most of the recoverable value sits in a mid-market business. Not headcount. Contracts.

What a proper review covers

Vendor and contract base. Every material supplier agreement: term, renewal mechanics, pricing structure, escalation clauses, exit rights. In most businesses that have not been through a sponsor before, a meaningful share of these were signed years ago and never revisited. That is addressable cost, and it should be sized in the deal model.

Technology spend. Software licenses, cloud infrastructure, IT services. License counts against headcount. Cloud commitments against actual workloads. Support contracts against what is actually used. This is where acquisitive businesses leak most, because each acquisition brought its own stack and nobody consolidated.

Marketing and agency effectiveness. Agency contracts, fee structures, media transparency, production costs. Marketing is often the largest discretionary budget in the business and the least commercially governed. In diligence, we look at whether the spend is buying growth or activity.

Procurement maturity. Is there a function, a process, and an owner? Or does every department buy what it wants? The answer tells you how fast savings can be captured post-close and how much of the improvement will stick.

Revenue operations. Pipeline hygiene, forecast accuracy, CRM data quality, sales and marketing handoff. If the growth plan depends on the sales engine and the sales engine cannot forecast within 20%, the plan is at risk.

Management reporting. How long does it take to close the month? How much of the board pack is assembled by hand? Slow, manual reporting is a leading indicator of a business that will struggle to run a 100-day plan.

What the output should look like

Not a report. A set of deal model inputs: addressable savings by category, risk-adjusted, with a timeline for capture. Plus a draft 100-day plan, so that Day One is a start, not a planning exercise.

Done this way, operational diligence pays for itself several times over before close, because it changes the price, the plan, or both.

When to do it

Earlier than you think. The findings are most valuable while price and structure are still open. A two to three week review during the exclusivity period is usually enough for a mid-market business, provided the data room is reasonably complete.

The CFO is in the marketing meeting now. Good.

Finance scrutiny of marketing spend has gone from annual to constant. The CMOs who treat that as an opportunity are winning budget. The ones who treat it as an intrusion are losing it.

For most of my career, marketing budgets were set once a year in a negotiation with finance and then defended, more or less, until the next one. That has changed. The CFO now wants to see the return on every major line, every quarter, and sometimes every month.

Some marketing leaders resent this. I think it is the best thing that has happened to the function in a decade.

Why

Money that cannot be explained gets cut. Money that can be explained gets protected, and often increased. The CMOs who can show the CFO exactly what the agency fee buys, what the media supply chain costs, what production returns, and how the technology stack earns its keep are the ones whose budgets survive a bad quarter.

That means treating marketing as a cost base with the same discipline finance applies to everything else. Benchmarked fees. Contracts with audit rights. Media measured on outcomes rather than impressions. A technology inventory that matches licences to users.

The uncomfortable part

Most marketing organizations cannot do this today. Not because they are careless, but because commercial governance was never their job and nobody built the capability. Procurement helps with the paperwork and stops at the door of the creative relationship.

So the CFO asks a fair question, the CMO does not have the number, and the conversation becomes adversarial. Both sides lose.

What good looks like

A marketing leader walks into the finance review with a one-page view of the cost base: agencies, media, production, technology, team. Each line benchmarked. Each line with an owner and a plan. The savings from last quarter's renegotiation flowing back into media that performs.

The CFO stops asking whether marketing is efficient and starts asking how much more it can do. That is the conversation you want.

Should we bring programmatic media in-house?

Every few years the in-housing debate comes back. The answer for most mid-market brands is not yes or no. It is: take control of the parts that matter, and stop paying for the parts that don't.

The in-housing conversation usually starts with a fee. Someone looks at what the agency charges for programmatic, compares it to the cost of two or three hires, and concludes the brand could do it cheaper. Sometimes that is right. More often the math is incomplete and the real question is being missed.

What you are actually buying from the agency

Programmatic delivery through an agency bundles several things: platform access and seat costs, trading expertise, data partnerships, planning and strategy, reporting, and the agency's margin on each. When brands talk about in-housing, they usually mean taking over the trading. But trading is the part most likely to be automated and least likely to be where the value sits.

The value sits in control: of the platform contract, of the data, of the supply path, of the measurement. You can have all of that without hiring a trading team.

The hybrid most brands should run

Own the platform contracts. Your Trade Desk, DV360, or Amazon DSP seat should be in your name, with your data in it. If the agency leaves, the history stays.

Own the data. Log-level data, first-party data, and the measurement stack belong to the brand. Write it into the contract.

Own the supply path decisions. Curated deals, private marketplaces, and direct publisher relationships should be set up on your terms, not the agency's. In one recent audit we found 53 supply-side platforms routing the same inventory. Nobody at the brand had chosen that.

Then decide who trades. For most mid-market brands, an agency or specialist trading partner working inside your seat, on a transparent fee, is more efficient than a full in-house team. The cost of senior programmatic talent is high and the retention risk is real.

When full in-housing makes sense

When programmatic is a core competency of the business, when spend is large enough to justify a team of five or more, and when the brand has the operating discipline to run measurement and supply path management itself. That describes some retailers, some direct-to-consumer brands, and some platforms. It does not describe most mid-market companies.

The mistake to avoid

Bringing trading in-house while leaving the agency's platform contracts, data, and supply path in place. That gives you the headcount cost without the control, which is the worst of both.

A useful way to decide

List what you want to control. List what you want to stop paying for. Then look at what the agency is actually charging for each element. The right structure is usually obvious once the bundle is unpacked.

Principal media, explained for the people who sign the invoices

Agencies buying media as principal and reselling it to clients is now mainstream. It is not automatically bad. It is automatically opaque. Here is what to ask.

If your agency group has offered you a "proprietary inventory" or "principal-based" media product, you have been offered principal media. The agency buys inventory at one price, sells it to you at another, and keeps the difference. The margin is not disclosed. The product is presented as a deal.

Is it bad?

Not inherently. Sometimes the agency's scale buys inventory you could not access at that price. Sometimes the product is fine and the price is fair.

The problem is that you cannot know, because the model is built on not telling you. Your agency, which is supposed to be your agent, is now also your vendor, with a margin incentive to put your money into its own product. That is a structural conflict, and it should be managed as one.

Five questions before you say yes

What is the disclosed margin, or at minimum, what is the guaranteed price versus the market price for comparable inventory?

Can you opt out per campaign, or is it bundled into the plan?

Do you retain log-level data and the right to audit delivery?

Is performance measured by a party who does not profit from the product?

Does your master services agreement cover principal transactions, or was it written when the agency was purely an agent?

What we do about it

We do not tell clients never to buy principal media. We tell them to buy it with their eyes open: a contract that names it, a price test against the open market, an audit right, and a measurement model the agency does not control. Most agencies accept those terms when a client asks clearly. The ones that do not are answering the question for you.

How do you audit a marketing budget?

A marketing audit is not a review of the creative. It is a commercial review of where the money goes, what it buys, and whether anyone is governing it. Here is the sequence that works.

Most marketing budgets have never been audited in the way finance would audit any other cost center of the same size. The reasons are cultural. Marketing is seen as creative, subjective, and hard to measure, so the commercial questions do not get asked. That is expensive. Here is the sequence we use.

Step one: build the full picture

Pull every marketing-related cost into one view: agency fees, media, production, technology subscriptions, freelancers, events, sponsorships, and the internal team. Most brands have never seen this as a single number. When they do, it is usually larger than anyone expected.

Step two: pull the contracts

Every agency agreement, every platform contract, every production supplier. Note the term, the fee model, the renewal mechanics, the audit rights, and the data ownership clauses. This is where you learn what you can and cannot change quickly.

Step three: benchmark the fees

Compare agency rate cards and fee structures against the market by agency type, role level, and region. Compare production costs against industry rates. Compare technology costs against comparable stacks. Independent benchmarks exist. Use them.

Step four: check scope against delivery

For each agency, compare what the scope of work says with what has actually been delivered over the last 12 months. Look for scope that expanded informally, deliverables that were paid for but not produced, and senior people from the pitch who are no longer on the account.

Step five: follow the media money

For media, go past the agency reporting. Look at supply path fees, platform charges, inventory quality, and the attribution model. Agency reporting tends to measure activity. You need to know what the spend did for the business.

Step six: assess the technology stack

List every marketing tool. Match licenses to users. Identify duplicates and tools that are paid for but barely used. The average mid-market stack has 30 to 40 tools and uses about a third of them properly.

Step seven: look at the operating model

Who decides what? How does work flow between the internal team and the agencies? Is there a procurement or marketing operations function, or does everyone buy what they want? Structure explains most of what you found in steps three to six.

What comes out

A prioritized list of actions with a value attached to each: renegotiate this fee, consolidate these two agencies, cut these tools, reset this contract, restructure this team. Then a governance model so the audit does not need repeating in three years because the problems came back.

The whole exercise, done properly, takes six to ten weeks for a mid-market brand. It routinely identifies savings equal to several times its cost, and the performance improvement is usually worth more than the savings.

Your AI pilot is a licence, not a strategy

Most companies we meet have three AI subscriptions, two pilots, and no number. That is not a strategy. It is a cost line waiting to be noticed.

Ask a leadership team what their AI strategy is and you will usually hear a list of tools. A licence for the writing assistant. A pilot with a vendor. A team in one department doing something interesting. Everyone agrees it matters. Nobody can say what it has returned.

That is not a criticism. It is where most of the market is. But it is worth being honest that a set of subscriptions is not a strategy, and a pilot without a metric is not a pilot. It is a licence.

The question that fixes it

Which workflow, with what volume, costing what today, will cost what after? If nobody can answer that for a given tool, the tool is not deployed. It is installed.

Start with the process and the number. Then choose the tool. Most companies do it the other way round, which is how you end up paying for three things that overlap and none that pay.

What we see working

One function. One high-volume, judgment-heavy workflow. A baseline measured before anything is built. A tool chosen for that job, not for the logo. An owner. Sixty to ninety days to production, adoption included. Then the next one.

It is not glamorous. It works. And it usually pays for itself before the second workflow starts, partly because the first thing it finds is the licences you can cancel.

Will AI reduce our agency fees?

Only if you ask. AI has cut the time agencies spend on planning, production, and reporting. Most have kept the savings. The fix is in the contract, not the budget.

This is the question we hear most often from CMOs and CFOs right now, and the answer is uncomfortable for both sides. Yes, AI has materially reduced the cost of producing a lot of agency work. No, that has not shown up in most clients' fees. The gap is the agency's margin, and it will stay there until the client reopens the conversation.

What has actually changed

Media planning that took a team a week now takes a day with the right tools. Reporting is largely automated. Creative production, particularly adaptations, versions, and content at volume, has collapsed in cost. Strategic thinking, senior judgment, and relationships have not changed much. The agency's cost base has shifted toward the expensive people and away from the hours.

If your fee was built on a staffing plan and hourly rates from 2022 or 2023, it reflects a world that no longer exists.

Why fees have not moved

Agencies are businesses. They have absorbed AI tooling costs, and they are under pressure from holding company targets. Nobody voluntarily gives back margin. It is not deceptive. It is what any business does when its costs fall and its prices are fixed.

The other reason is that most clients have not asked. Fee reviews happen at pitch or at renewal, and both are infrequent. In between, the agency's productivity gains accumulate quietly.

What to do

Reopen the staffing plan. Ask how many hours, at what level, the work now takes. Compare that to the plan the fee was built on. The difference is your conversation.

Move production to output-based pricing. If adaptations, versions, and content pieces are largely automated, pay per unit at a rate that reflects that. Retainers built on hours make no sense for automated work.

Write AI into the contract. A productivity clause that shares efficiency gains, a requirement to disclose where AI is used, and clarity on ownership of AI-generated assets and data.

Keep paying for judgment. The senior strategist, the planner who knows your category, the account lead who tells you the truth: that is still worth the money. Pay for it directly rather than through a blended rate that also covers work a machine now does.

What a good agency will say

The good ones will engage. They have already restructured their own cost base and they would rather have an honest conversation than lose the account at the next pitch. Some will propose new models themselves. The agencies that resist transparency on this are telling you something.

The bigger point

AI does not automatically make marketing cheaper. It makes it cheaper for whoever controls the contract. Make sure that is you.

Agentic workflows: what is real in 2026, and what is a demo

Agents that take actions across systems are the most oversold and the most valuable thing in enterprise AI right now. Both at once. Here is how to tell the difference.

Every vendor now sells agents. Software that does not just answer but acts: reads the contract, updates the CRM, drafts the renewal notice, files the report. Some of it is real and producing measurable results inside businesses we work with. Some of it is a demo that has never survived contact with a real system.

What is real

Agents that operate inside one well-defined workflow with clean inputs and a human checkpoint. Contract review that extracts terms, flags risks, and drafts the renegotiation brief for a person to send. Agency deliverable checking against scope. Pipeline hygiene that flags stale deals and drafts the follow-up. Month-end reporting that assembles the pack from source systems and leaves a person to write the commentary.

These work because the task is bounded, the data is accessible, the failure mode is visible, and a person still owns the outcome.

What is a demo

Agents that promise to run a function end to end across six systems with no checkpoint. They look extraordinary on a stage. Inside a business with messy data, inconsistent processes, and permissions nobody has mapped, they break in ways that are hard to see until something has gone wrong.

Not because the technology cannot get there. Because the process underneath was never cleaned up, and an agent running a broken process runs it faster.

How to buy

Insist on a bounded first use case with a measured baseline. Insist on a human checkpoint until the error rate is known. Insist on knowing which systems the agent touches, with what permissions, and who owns the outcome. Choose the tool for the job, not the vendor for the roadmap.

Then, once the first one is running and measured, do the next. The companies that are getting real value from agents in 2026 got there one workflow at a time. The ones with the impressive demos are still in the pilot.

Why Your Media Agency Review Is Long Overdue (And How to Run One Without Blowing Up the Relationship)

Most brands go years without a proper agency review. Here's how to run one that's honest, structured, and doesn't poison the relationship in the process.

The Review You've Been Putting Off

Most brands haven't done a proper agency review in three or four years. The last one was probably triggered by a crisis, a missed campaign, a budget blowout, a change in CMO, rather than by any kind of deliberate governance. And so the relationship has drifted. The agency has gotten comfortable. You've stopped asking hard questions. Performance has become something you describe in percentages that look fine on a slide and mean almost nothing in practice.

That's not a dig at agencies. It's a description of what happens when there's no structured accountability. The work fills the space available, the fees creep, and both sides learn to manage expectations downward rather than have an uncomfortable conversation.

The review isn't about finding a reason to fire them. It's about understanding what you're actually getting.

What a Real Review Looks At

Start with the commercial structure. Pull the contract, the rate card, and the last two years of invoices. What did you pay for? What was delivered? Where did the scope expand without a formal conversation? Most brands discover at least one or two line items they can't fully explain. That's not fraud, it's just what happens when there's no one on the client side managing the commercial relationship properly.

Then look at performance. Not the metrics the agency put in the monthly deck. Go back to the business outcomes you were trying to drive and ask whether the work moved them. If you can't connect agency activity to business results, that's a structural problem, either you didn't brief them properly, or they've been optimizing for the wrong things, or both.

Finally, look at the team. The senior people who won the pitch, are they the ones actually working on your account? This is one of the oldest complaints in the industry and one of the most consistently ignored. You're paying for expertise. You should know whose expertise you're actually paying for.

Running It Without Blowing Things Up

The relationship is real and it has value. Running a review doesn't mean you're about to leave. The agency doesn't need to feel like a defendant.

Be direct about what you're doing and why. "We're doing a structured review of all major vendor relationships this year" is honest and neutral. It signals seriousness without signalling the exit. Most agency leaders, when treated like adults, respond well. They want to know where they stand. The ones who don't are usually the ones with something to hide.

Bring a clear scorecard. Define what good looks like before you walk into the room. That keeps the conversation factual and reduces the chance that it becomes a personality conflict.

If you're considering going to market, say so. Pretending you aren't when you are is a waste of everyone's time and tends to come out anyway.

What Comes Out the Other Side

A good review ends with one of three outcomes: a renegotiated, tighter relationship; a competitive pitch that either reaffirms the incumbent or brings in someone better; or a managed transition. All three are legitimate. The one outcome you want to avoid is what you have now, a relationship that's never been properly examined, operating on autopilot, with no one sure whether it's actually working.

Your media budget is likely one of the largest line items on your marketing P&L. It deserves the same rigour you'd apply to any other major vendor relationship. The review doesn't have to be adversarial. But it does have to happen.

Private equity’s next lever is the vendor stack, not headcount

Headcount is the visible cost. Vendors, agencies, software, and cloud are the invisible one, and in most mid-market portfolio companies they are where the recoverable EBITDA sits.

When a value creation plan needs EBITDA quickly, headcount is where most people look first. It is visible, controllable, and the maths is simple. It is also the lever with the highest cost to morale, capability, and the growth story you will need to tell at exit.

The lever most plans under-use is the vendor stack. Agencies, software, cloud, IT services, telecom, professional services. In a typical mid-market business these lines have grown for years without a commercial owner, and they are almost always over market.

Why it is under-used

Nobody in the portfolio company owns the total. Marketing owns the agencies, IT owns the software, engineering owns the cloud, finance sees the bills. Each function defends its own vendors. The sponsor sees a cost base that is "mostly people" because that is how the P&L is presented.

And the work is unglamorous. Benchmarking a renewal is not a strategy offsite.

What it returns

In our experience, a vendor-first program in a mid-market portfolio company recovers a meaningful percentage of the addressable spend inside the first year, with most of it visible by month four. Agency fees reset to market. Software licences matched to headcount and renewals renegotiated with pricing data. Cloud commitments re-cut. Intermediaries removed from the media supply chain.

None of it requires a restructuring. Most of it improves performance, because the money goes back into things that work.

How to run it

Put it in diligence, so the deal model reflects it and the price does. Build the 100-day plan around the largest addressable lines. Give the program one owner with the sponsor's backing, because functions will defend their vendors. Track it against EBITDA every fortnight. Document it for the exit.

Then, when you do need to look at headcount, you will be doing it from a business that is already leaner where it does not hurt.

The RFP Process Most Brands Get Wrong, And How to Fix It

An RFP is supposed to help you find the best partner. Most of them do the opposite, they select for whoever writes the best pitch.

A Process That Selects for the Wrong Things

The standard agency RFP process is essentially a writing competition. Agencies spend weeks crafting a response designed to impress a committee of people who probably don't agree on what they're looking for in the first place. The prettiest deck wins. Or the most familiar name. Or the one whose account director had the best lunch.

That's not entirely cynical. Procurement teams are busy, evaluation criteria are often vague, and the people scoring responses are usually looking for a reason to feel confident rather than a method to measure fit. The result is a process that's thorough in appearance and fairly random in outcome.

If you've ever awarded a major agency contract and had buyer's remorse within six months, the problem started in the RFP.

Where Most Briefs Fall Apart

The brief is where the process lives or dies. Most briefs are too long, too vague, or both. They describe the business at length without saying what problem needs solving. They list objectives that could apply to any brand ("drive awareness, grow consideration, convert customers"). They ask for case studies that demonstrate broad capability when what they should be asking for is specific evidence of work in comparable situations.

The agencies reading a poor brief will fill the gaps with whatever makes them look strongest. You're not learning about them, you're learning about how they pitch.

A good brief is honest about the actual challenge. It names the tension in the business. It tells prospective agencies what success looks like in concrete, measurable terms. It sets scope clearly enough that you can compare responses apples-to-apples. And it is short enough that an agency principal actually reads it rather than delegating it to the team who'll write the response.

Fix the Evaluation Before You Send the Brief

Most organizations design the RFP and then figure out how to score it. It should be the other way around. Before the brief goes out, the selection committee needs to agree on what they're actually buying: capability, chemistry, capacity, price, or some specific combination. Those priorities need to be weighted. And the people scoring need to be the people who will manage the relationship, not just procurement and legal.

Two-stage processes work better than one. An initial submission to shortlist, followed by a live working session with the finalists, tells you far more than two rounds of paper. The working session isn't a presentation. It's a task, give them a real problem from your business, with two days to respond. Watch how they think, who shows up, and whether the work is actually good.

Chemistry matters and it's fine to say so. A five-year agency relationship that's purely transactional is a miserable way to work. You're allowed to factor in whether you'd want to be in a room with these people for extended periods. Just don't let it substitute for evaluating the work.

What Good Looks Like

A well-run RFP takes longer to design than it does to execute. The brief gets reviewed internally until it's honest. The evaluation criteria are fixed before the process starts. The scoring is done by the people with skin in the game. And the winning agency understands exactly what they're walking into, not a cleaned-up version of it.

That last point matters more than most procurement guides admit. An agency that wins a properly structured RFP and then delivers badly has no excuses. An agency that wins a vague brief and then underdelivers can always point to the gaps. Give them no gaps.

Where the money actually is

Twenty percent of the cost base holds eighty percent of the recoverable value. Here is where it tends to sit in a mid-market company, and why nobody has found it.

The name of this firm is not an accident. In every cost base we have worked on, a small number of lines hold most of the recoverable value. Not because the rest is efficient, but because the big money concentrates in a few places, and those places are the ones with the weakest governance.

The usual suspects

Agency and media spend. Large, discretionary, governed by relationships rather than contracts, with fees that predate AI and supply chains nobody has mapped.

Enterprise software and SaaS. Renewals signed at the vendor's best offer. Licences above headcount. Acquired businesses with duplicate stacks.

Cloud. Commitments bought at a scale the business no longer runs at, or never bought at all.

IT services and telecom. Contracts nobody has re-bid in years, on the assumption that they are too complicated to change.

Professional services. The consultants, advisors, and specialists that accumulate around a growing business.

Why nobody has found it

Each line has an owner who defends it and nobody who challenges it. The finance team sees the bill and not the contract. The function sees the vendor and not the market price. Procurement, if there is one, is downstream of the decision.

The money is not hidden. It is unowned.

The eighty twenty of the work

Find the five largest addressable lines. Benchmark each against the market. Renegotiate, consolidate, or replace, in that order. Put an owner and a renewal date against every contract that survives. Repeat quarterly.

That is most of the value, in most businesses, most of the time. The rest is detail.

Programmatic Advertising: Where Your Budget Goes and Why a Chunk of It Shouldn't

Programmatic can be one of the most efficient channels in your mix. It can also be one of the most opaque. Here's what's actually happening to your budget.

The Supply Chain Nobody Explained Properly

Programmatic advertising works. That's the first thing to say, because the critique that follows isn't an argument for going back to direct buys and insertion orders. It works. The targeting is real, the scale is real, the efficiency gains over traditional media buying are real.

But the supply chain between your media budget and the actual impression delivered to an actual human is longer, murkier, and more expensive than most advertisers understand. Money moves through DSPs, SSPs, data vendors, verification layers, ad networks, and exchanges, each one taking a cut, before anything reaches a publisher. In a reasonably well-managed programmatic setup, you're lucky if 50 cents of every dollar ends up as working media. In a poorly managed one, it's less.

This is not a secret. It's been documented repeatedly. It just tends to get buried in dashboards full of CPMs and click-through rates that look fine, so nobody asks where the rest went.

What's Actually Eating Your Budget

The three biggest culprits are fees, fraud, and mismatched inventory.

Fees first. The tech stack costs money. Your DSP takes a percentage. Your data segments cost per thousand. Your verification vendor costs per impression. Your agency's trading desk may be taking a margin on top of that, which may or may not be disclosed depending on your contract. By the time all of that clears, the media you're buying is a fraction of what you budgeted for media.

Fraud is real and persistent. Industry estimates suggest somewhere between 10% and 20% of programmatic impressions are non-human. Bot traffic, domain spoofing, and made-for-advertising sites are well-documented problems. Brand safety and verification tools help, but they're not free, and they don't catch everything.

Then there's inventory quality. Cheap CPMs often mean cheap inventory, ads appearing on pages that nobody reads, in placements that nobody sees, against content that you'd never want your brand associated with if you actually knew what it was. Low cost per impression is not a metric. Cost per impression that was seen by a real person in a context relevant to your brand is closer to a metric.

What You Can Actually Control

Start with transparency. Ask your agency or trading desk for a complete breakdown of where fees sit in the supply chain and who captures them. This conversation can be uncomfortable. Some partners won't like it. That's useful information.

Look at your tech stack. Are you paying for three different data vendors who are largely overlapping? Are you running through more intermediary layers than you need to? Simpler supply paths deliver more working media. That's just arithmetic.

Apply inclusion lists rather than exclusion lists. Blocking bad inventory is fighting a rearguard action. Specifying the inventory you want, publishers, environments, formats, is a more controlled approach. Yes, it constrains scale. Scale at low quality is not scale.

Invest in attention metrics and viewability, not just reach. An impression that wasn't seen didn't happen.

The Conversation Worth Having

This isn't about blaming anyone. Most of the fee structures in programmatic advertising are industry standard, disclosed in contracts that nobody read carefully enough, and operating exactly as designed. The design just happens to not be especially aligned with the advertiser's interests.

Knowing that is the starting point. From there, the conversation with your partners, agency, DSP, data vendors, becomes specific and commercial rather than vague and uncomfortable. "Help me understand the total take rate in this supply path" is a reasonable question. The answer will tell you a lot.

Having a procurement team is not the same as having procurement

Most mid-market companies with a procurement function still buy their largest categories without it. The gap is not people. It is mandate.

"We have procurement." I hear it in most first conversations, and it is usually true and usually beside the point. The team exists. It handles purchase orders, supplier onboarding, and the categories it was set up for. It is not in the room when the agency contract is renewed, the software enterprise agreement is signed, or the cloud commitment is made.

Those are the largest addressable lines in the business. They are bought by the function that uses them, negotiated by people who do it once every few years, against vendors who do it every day.

Why the gap exists

Procurement in a mid-market business grows up around direct spend and operational purchasing. Marketing, technology, and professional services are seen as specialist, relationship-driven, and too important to slow down. So they stay outside the process, and the process stays where it is.

Everyone is behaving sensibly. The result is that the categories with the most money and the least governance are the ones with no commercial ownership at all.

What fixes it

Not a bigger team. A clearer mandate, backed by the CFO, that says: nothing above this threshold gets signed without benchmarking, and here is who does it. Category expertise for the big lines, whether in-house or on call. A renewal calendar that surfaces contracts ninety days out. And a quarterly review where the largest vendors are looked at as a portfolio, not as a set of individual relationships.

That is a governance change, not a headcount change. It is also, in our experience, one of the highest-return decisions a CFO can make in a year.

Agency Roster Bloat Is Costing You More Than You Think

Most marketing organizations have more agencies than they need. The hidden costs aren't just financial, they're operational, strategic, and slow.

How It Happens

Nobody sat down and decided to work with twelve agencies. It accumulated. A new CMO brought in their preferred creative shop. A regional team hired a local digital agency because the global AOR didn't want to handle smaller markets. Someone needed a PR firm for a product launch and the one they used is still on the roster three years later. An agency was kept on a retainer "just in case" because ending it felt like burning a relationship.

Add it up and the average mid-size marketing organization has between eight and fifteen agency relationships. Plenty of them have more. The number is rarely the result of deliberate design.

The Costs That Don't Show Up in the Invoice

The obvious cost is fees, retainers, project fees, and the overhead of maintaining commercial relationships with a dozen external partners. That's real and usually large enough to justify a review on its own.

The less obvious costs are harder to quantify and often worse.

Briefing time is the first one. Every agency relationship requires ongoing input, briefs, feedback, reviews, approvals. If your marketing team is spending meaningful hours every week managing agency admin across a large roster, that's internal capacity being eaten by coordination rather than productive work.

Consistency is the second. Twelve agencies produce twelve different creative interpretations of your brand. Even with a brand guide, the output fragments. Work that should be building cumulative brand equity ends up feeling scattered. You can manage this, but managing it takes more effort than most teams realize.

The third is accountability. With one agency, performance is easy to attribute. With twelve, everyone has a plausible explanation for why results weren't their responsibility. Accountability diffuses in direct proportion to the number of parties involved.

What Rationalization Actually Looks Like

The goal isn't to cut to the bone. It's to have a roster that's the right size for what you're actually trying to do, with each relationship well-managed and clearly scoped.

Start by mapping what each agency is doing and what it's costing, total cost, not just the headline retainer. Include internal time. Then ask, honestly, which relationships are producing work you couldn't get from an existing partner with a slight scope expansion. You'll usually find two or three that are effectively redundant.

Then look at your core work and ask whether your lead agencies have the full capability you need, or whether you've been using specializts because your main partners weren't configured correctly. Sometimes the problem isn't too many agencies, it's a lead agency that underdelivered and got supplemented rather than fixed.

Consolidation conversations are easier to have than people expect. Most agencies would rather expand scope than lose the relationship. Frame it as a genuine opportunity and most of them will engage seriously.

The Relationship You Actually Want

The goal is a small number of agency relationships that are high trust, well-compensated, and genuinely accountable. Partners who understand the business deeply, who you brief properly and who brief you back when the strategy needs challenging. That kind of relationship takes time to build and requires some intentional investment.

It's harder to have twelve of them. Usually impossible. Most organizations with large rosters don't have any of them, just a large collection of transactional relationships, each one wondering whether it's going to survive the next budget review.

Auto-renewal is a business model

The clause you did not read is the vendor’s most reliable revenue line. A renewal calendar is the cheapest fix in the company.

Software vendors do not rely on you loving the product. They rely on you not noticing the date.

Auto-renewal with a notice window, a price escalator, and a term that resets on renewal is standard in enterprise agreements. It exists because it works. A meaningful share of enterprise software revenue renews without a negotiation, at a higher price, for a longer term, because the notice window closed while everyone was busy.

What it costs

A licence count that has drifted above headcount since the last true-up. An escalator compounding on a price that was never benchmarked. A term that now runs past the point where you might have consolidated the tool. Multiply by the number of contracts nobody owns.

The fix

A renewal calendar. Every contract, its notice window, its owner, and a reminder ninety days out. It is a spreadsheet. It is also, in most companies we work with, worth more than any single negotiation, because it turns every renewal back into a decision.

Then use the decision. Benchmark the price. Match licences to users. Ask for the term you want. The vendor expected you to miss the date. Surprise them.

How to Negotiate an Agency Contract Without Losing the Agency

Agency contracts are almost always negotiable. Here's how to get better commercial terms without turning the relationship into a standoff.

The Awkward Truth About Agency Contracts

Agency contracts are built for agencies, not clients. That's not a conspiracy, it's just what happens when one side of a negotiation has a standard template and the other side doesn't have a procurement specializt in the room. Rate cards get accepted. Scope definitions stay vague. IP ownership sits in a clause nobody read. Audit rights are absent. And the whole thing rolls over at the end of the year because renegotiating feels uncomfortable and switching costs seem high.

None of this is inevitable.

What's Actually Negotiable

Most commercial terms in agency contracts are negotiable, and the agency expects some pushback. The question is knowing where to push.

Rates are the obvious place. Most agencies have a rate card and then a real rate, and the gap between them is meaningful. Market benchmarking, getting a sense of what comparable agencies are charging for comparable talent, gives you a credible basis for the conversation. You don't need to threaten a pitch. You need to be able to say, with some evidence, that the rates are above market. That's usually enough.

Scope definition is more important than rate. Vague scopes are expensive because everything outside them becomes a change order. A contract that defines deliverables clearly, including what's included in the retainer versus what gets scoped separately, saves real money and prevents the low-level friction that erodes agency relationships. Both sides benefit from clarity. If the agency resists it, ask why.

Payment terms, audit rights, IP ownership, notice periods, and benchmarking provisions are all areas most clients don't bother to negotiate. They should. An audit right, for instance, costs nothing to put in a contract and materially strengthens your position if questions arise about what you were billed for. The fact that most clients never exercise it doesn't mean it's not valuable to have.

Running the Negotiation Without Wrecking the Relationship

The relationship matters. If you go into this as an adversarial exercise, you'll win on paper and lose in practice, the agency will find ways to recover margin, they'll deprioritize your account, and the best talent on their team will move to a client they'd rather work for.

The frame that works is mutual fairness. You want a commercial structure that's sustainable for both sides. An agency that's losing money on your account will either understaff it or exit the relationship at the first opportunity. A client who's overpaying will eventually run a review. Neither outcome is good for anyone.

Come with specific asks. "Your rates are too high" is a complaint. "Your senior strategist rate is 20% above what we're seeing at comparable agencies and we'd like to align it to market" is a negotiation. The more specific you are, the more professional the conversation feels, and the easier it is for the agency to say yes without feeling like they've surrendered.

After the Contract, the Real Work

A better contract is a starting point, not an outcome. The value of a renegotiation comes from what you do with it: tighter governance, clearer performance measurement, regular commercial check-ins that keep both sides honest.

Contracts that sit in a drawer tend to drift. The people managing the relationship day-to-day don't know what's in them, and the terms that were carefully negotiated go unenforced. Build a management cadence that uses the contract as an active document rather than a historical artefact. Review scope against delivery quarterly. Look at the billing. Ask the questions.

It doesn't take much to stay on top of a well-structured agency contract. It takes quite a lot to recover ground after you've let it run unchecked for three years.

Retail media is the new programmatic. Same margin, new name.

Retail media networks are the fastest-growing line in many consumer budgets and the least governed. The lessons from a decade of programmatic apply directly.

If you sell through large retailers, their media networks are now a significant and growing line in your budget, often funded from trade spend and managed by a different team from the one that runs your brand media. It is measured by the retailer, planned by the retailer, and priced by the retailer.

We have seen this movie. It was called programmatic, and it took a decade for advertisers to get transparency, independent measurement, and fair pricing. Retail media is at the beginning of the same curve.

The same problems, earlier

Measurement controlled by the seller. Attribution that credits the network for sales that would have happened anyway. Pricing with no benchmark. Inventory quality that varies enormously between placements that cost the same. Fees taken by the technology layer between you and the retailer. Agencies with their own incentives layered on top.

What to do now

Bring retail media into the same governance as the rest of the media budget. One owner across brand and trade. Independent incrementality testing, even if the retailer resists it. A price benchmark across networks. Contracts with the retailer and the technology partner that specify data access and audit rights.

And be honest about the strategic trade. Some retail media spend is the cost of shelf position dressed as advertising. That may be a fine decision. It should be a decision.

Advertisers who apply what they learned from programmatic will get to fair terms in two years instead of ten. The ones who treat retail media as a trade-marketing relationship will pay for the lesson twice.

Your SaaS Stack Has a Leak: The Vendor Rationalization Guide

Most organizations are paying for software they don't use, software that duplicates other software, and software nobody can quite explain. Here's how to find the leak.

The Problem With Software That's Easy to Buy

The SaaS model changed how organizations buy software, and not entirely for the better. When procurement was hard, a capital expenditure, a formal evaluation, a lengthy IT review, there was a natural filter. Only serious requirements made it through.

Now anyone with a corporate card can spin up a subscription. That's genuinely useful for moving quickly. It's also how you end up with forty-seven SaaS tools, three project management platforms, two contract management systems, and a marketing automation tool that nobody uses because the one the sales team preferred was also purchased and they never integrated.

The average mid-market organization spends meaningfully more on software than its finance team knows, because spend is fragmented across departments, cost codes, and credit cards. The spend that is tracked is often treated as fixed, renewal reminders come in, they get approved by the person who approved the original purchase, and the cycle continues.

What a Rationalization Actually Involves

First, you need a complete picture. Pull all SaaS subscriptions, from IT-managed systems, department budgets, and anything on corporate cards. Include contracts with renewal dates and actual cost, not just the headline subscription fee (add-ons, professional services, and extra seats add up).

Then look at usage. Most SaaS platforms will give you utilization data, active users, feature adoption, login frequency. When you run this analysis, the pattern is consistent: somewhere between 20% and 40% of licenses are either unused or barely used. People got access, never onboarded properly, and the license kept renewing.

Cross-reference for overlap. You'll find tools doing similar things. It's rarely malicious, different teams bought what made sense for them, or a merger brought in a second instance of something you already had. But the overlap is real and it costs real money.

The Conversations You Need to Have

Rationalization runs into resistance because people attach to their tools. The team that fights hardest to keep a platform that costs £8,000 a year is usually the team that's built workflows around it and fears the disruption of changing. That's a legitimate concern, not obstruction.

The conversation that works is one that's honest about total cost and honest about what's being replaced. "We're going to retire this and move to the platform the rest of the organization uses, and here's how we'll make that transition manageable" lands better than "we're cutting software to save money."

Where there's genuine functional need, the tool does something that nothing else in the stack does, the users are genuinely dependent on it, and the cost is proportionate, keep it. Rationalization isn't about reaching a target number of tools. It's about making sure every tool earns its place.

What the Savings Look Like

A well-run rationalization of a 200-person organization's SaaS stack typically finds 15% to 30% cost reduction without removing anything that's actually being used. On a stack that's reached seven figures, that's meaningful.

The less obvious return is operational. Fewer tools means fewer integrations to manage, fewer security surface areas, and less of your IT team's time spent on vendor management. It also means data is less fragmented, which matters every time someone tries to build a report that should take twenty minutes and takes three days because the relevant information lives in four systems that don't talk to each other.

The operator’s test for any consultant: will they run it?

Recommendations are cheap. The question to ask anyone who wants to advise you is whether they will stay for the negotiation.

I have hired consultants, been one, and worked alongside the large firms on client sites. Most of them are smart. Most of the work is competent. And most of it stops at the point where the hard part begins.

The deck says: renegotiate the agency contract, consolidate the software stack, restructure the cloud commitment, rebuild the forecast process. Correct. Then the engagement ends and the client is left to do it with the same team, the same bandwidth, and the same vendors who now know a change is coming.

The test

Ask: will you sit in the negotiation? Will you run the RFP? Will you build the reporting? Will you be here in month four when the vendor pushes back and the internal owner has gone quiet?

If the answer is a version of "we advise, you implement," you are buying a document. Sometimes a document is what you need. Usually it is not.

Why it matters more now

Execution is where the value is, and it has become harder. Vendors are more sophisticated. Contracts are more complex. AI has changed the cost base of agencies and software faster than most contracts have adapted. Doing the work requires people who have done it before, against these counterparties, recently.

That is a different skill from analysis. It looks less impressive in a proposal. It is the only thing that shows up on the P&L.

What we chose

We built the firm around delivery. We take the mandate through to the result, and where it fits, we tie part of our fee to the savings. It keeps us honest and it keeps us in the room.

Ask anyone who wants to advise you the same question. The answer tells you what you are buying.

How to Stop Overpaying for Cloud Without Firing Your Entire DevOps Team

Cloud spend is the fastest-growing cost line in most technology budgets and the least understood. The good news is that most of the waste is findable without a full infrastructure overhaul.

Why Cloud Costs Are Hard to Control

Cloud pricing is deliberately complex. AWS, Azure, and Google Cloud each have pricing models that require a specializt to navigate, hundreds of service types, multiple pricing dimensions, discount structures that only apply under specific conditions, and billing that changes month to month based on consumption patterns you may not fully understand.

That complexity isn't accidental. It makes comparison harder, switching costs higher, and over-spending easier to miss. A 15% increase in your cloud bill looks like normal growth. It might be normal growth. It might also be idle resources, over-provisioned instances, and reserved capacity commitments you forgot to use.

Most organizations find out they've been overspending when a CFO asks a question that nobody can answer properly.

Where the Waste Usually Lives

The four most common sources of cloud waste are idle and over-provisioned resources, reserved instance gaps, orphaned storage, and data transfer costs.

Idle resources are the easiest to find and the easiest to cut. Instances that were spun up for a project, a test environment, or a migration and were never turned off. Resources running at 5% utilization that were provisioned for peak capacity three years ago. Every cloud provider gives you the tools to identify these. The reason they persist is usually that nobody is explicitly responsible for looking.

Over-provisioning is subtler. An application running on infrastructure sized for a load it never reaches. A database that's allocated ten times the storage it uses. These aren't waste in the same sense as idle resources, but right-sizing them reduces costs meaningfully without touching anything that matters.

Reserved instances and savings plans are where organizations leave real money. On-demand pricing is expensive. Committing to capacity in advance, one year or three years, brings significant discounts. But you need to commit to the right things, and you need to actually use what you commit to. Many organizations have reserved capacity they're not fully using alongside on-demand spend they could be covering with reservations. Cleaning this up requires some analysis, but it's not complicated.

Orphaned storage and data transfer costs are smaller individually but add up fast. Objects in cloud storage that belong to resources that no longer exist. Data moving between regions that could be avoided with a different architecture. These are the things that don't show up in any single review but compound over time.

The Governance Gap

Most cloud cost problems are governance problems rather than technical problems. The resources weren't reviewed. The commitments weren't tracked. The tagging wasn't done, so you can't attribute cost to the teams or applications generating it. The alerts weren't set.

FinOps, a growing practice area for cloud financial management, addresses this. The core of it is straightforward: give teams visibility into what they're spending, make them accountable for it, and build a process for reviewing and optimizing on a regular basis. It doesn't require a large team. It requires someone who owns the question and has the access to act on the answers.

What Good Looks Like

Cloud spend optimized. Reserved coverage matched to actual usage. Cost attributed to the applications and teams generating it. Anomalies caught in near-real-time. A regular review cadence that keeps the number from creeping.

This isn't a one-time project. Cloud costs change as architectures evolve and usage patterns shift. The goal is a steady state of reasonable efficiency, not a perfect optimization achieved once and then ignored.

Most organizations with a serious effort find 20% to 35% waste that can be cut without any change to what their infrastructure actually delivers. That's not a small number.

The 100-Day Plan Most PE-Backed Companies Don't Actually Execute

A hundred-day plan is a fixture of private equity value creation. Most of them are also aspirational documents that lose contact with reality around day thirty.

Why 100-Day Plans Fail

The idea is sound. Acquisition closes, new ownership takes over, and the first hundred days are an opportunity to establish direction, identify the highest-value levers, and start moving on the ones that don't require long preparation. Use the mandate that comes with new ownership before it dilutes into day-to-day management reality.

In practice, the plan fails for predictable reasons.

It was built on pre-acquisition data that didn't capture operational reality. The due diligence was financial and the plan reflects financial assumptions, cost reduction here, revenue improvement there, without the operational depth to know whether they're achievable. The people responsible for executing the plan are the same people who inherited a different set of priorities and a business that has immediate needs. And the plan was built to impress an investment committee, not to operate a business.

None of this is unique to private equity. It's what happens when planning is treated as a deliverable rather than a management process.

What a Credible Plan Looks Like

The first thirty days should be diagnostic, not execution. Walk the operations. Talk to the people doing the actual work, not just the leadership team. Understand what the business is currently good at and where the friction is. Form a view on what the pre-acquisition analysis got right and where it was optimiztic.

This is often resisted, investors want to see action, and a thirty-day diagnostic period can feel like delay. It isn't. It's the difference between executing the right plan quickly and executing the wrong plan expensively.

Days thirty to sixty should be focused on a small number of high-confidence, high-impact initiatives. Not everything on the opportunity list. The two or three things where you know what needs to happen, the path to execution is clear, and the management team is capable of delivering them now. Getting these moving builds credibility, creates momentum, and starts generating the returns that fund the harder work later.

Days sixty to hundred should establish the operating cadence, the rhythm of management that determines how the business will run going forward. Performance reviews, financial reporting, commercial governance, people decisions. This is less visible than a major initiative but more durable in its impact. A business with a strong operating cadence will outperform a business with a weak one at almost every level over time.

The Execution Gap

Most 100-day plans have more initiatives than any management team can realiztically execute simultaneously. Prioritization isn't the only problem, it's also the assumption that the existing team has the capacity and capability to execute the plan on top of running the business.

PE-backed companies often need external resource for the 100-day period. Not to replace management, but to provide focused project capacity on the specific initiatives that are time-critical. This is a legitimate and often underused lever.

The plan should also be a living document, reviewed weekly, updated as reality diverges from assumption, and ruthlessly honest about what's on track and what isn't. A plan that nobody looks at after day fifteen isn't a plan. It's a record of good intentions.

Log-level data: the clause worth more than the discount

Advertisers negotiate the fee and forget the data. The right to your own transaction records is what makes every future negotiation possible.

When brands negotiate with a media agency or platform, the conversation is about the fee. Percentage points, retainer levels, rate cards. Important, but it is the smaller prize.

The larger one is a clause: the brand owns, and can access at any time, the log-level data on every impression bought in its name. Where it ran, through which platforms, at what price, with which fees taken along the way.

Why it matters more than the fee

Without it, you cannot audit. You cannot see the supply path. You cannot test the attribution model the agency reports against. You cannot move agencies without losing the history. Every future negotiation starts from the agency's version of the facts.

With it, the audit is a matter of running the numbers. The supply path is visible. The intermediaries are countable. The fee negotiation is informed by what was actually bought.

In one recent review, log-level data showed 53 supply-side platforms routing the same inventory. That finding alone was worth several million dollars a year. It was only possible because the clause existed.

What the clause needs

Ownership of the data, not just access. Delivery on request, in a usable format, within a defined period. Coverage of every platform and partner in the chain, including the agency's own technology. Survival of the clause after the relationship ends.

Get that into the contract, then negotiate the fee. You will negotiate it better.

EBITDA Improvement Without Cutting People: The Vendor-First Playbook

When investors push for margin improvement, headcount is usually the first thing they look at. It shouldn't be. The vendor and contract side of the business almost always has more accessible upside.

The Default That Isn't Actually Optimal

Headcount is the biggest cost in most businesses. It's visible, it's controllable, and it produces immediate results in the P&L. When a business needs to improve EBITDA quickly, cutting people is often the first lever that gets pulled.

It's also, frequently, the wrong one, not for sentimental reasons, but for commercial ones. People create revenue. Cutting too deeply in the wrong places creates a performance hole that shows up eighteen months later, when the EBITDA improvement has already been claimed. By then, the damage is done and the same business that was cut to improve margins is now underperforming on the top line.

Vendors, on the other hand, are pure cost. No vendor ever generated a sale. Reducing vendor spend doesn't reduce capacity to earn revenue. Which makes it the first place to look, not the second.

Where Vendor-Side EBITDA Lives

There are four primary sources of vendor-side EBITDA improvement in most mid-market businesses.

Pricing normalisation: contracts that were signed at non-competitive rates, or rates that were competitive three years ago and haven't been benchmarked since. Enterprise software vendors, in particular, put significant rate increases into renewal proposals and capture them from organizations that aren't paying attention. Systematic benchmarking and renegotiation is a straightforward exercise.

Scope right-sizing: paying for vendor capacity or service levels that exceed actual need. IT managed services contracts scoped for a business twice the current size. Insurance coverage that was never revisited after a strategic pivot. Marketing retainers for agencies doing work that could be done for a fraction of the cost with a different structure.

Duplicate or redundant vendors: the accumulation problem. Multiple vendors delivering overlapping value, each rationalized individually at the time and collectively redundant.

Structural commercial misalignments: contracts where the incentive structure doesn't serve the business. Agency fee-for-service models that reward activity rather than outcomes. IT support contracts without performance standards. Professional services arrangements without clear scope boundaries.

The Process

A vendor cost review for EBITDA purposes should run eight to twelve weeks for a mid-size business. The output is a prioritized set of commercial actions, renegotiations, consolidations, terminations, and restructures, with a clear view on expected savings and implementation timeline.

Sequencing matters. Start with the contracts closest to renewal, those give you the most natural bargaining position. Run the renegotiations in parallel where possible. Be explicit with vendors about what you're doing: "we're reviewing commercial terms across our vendor base as part of a management improvement initiative" is honest and signals that this isn't a crisis, which keeps the conversations professional.

Savings in the range of 8% to 15% of total vendor spend are realiztic in a first pass. For a business with meaningful vendor cost, that can move EBITDA by multiple points without touching a single headcount line.

The Longer Game

Vendor cost management is not a one-time exercise. Businesses that build the governance to manage it continuously, with regular benchmarking, clear ownership of renewals, and commercial discipline embedded in the operating cadence, sustain the improvement. Those that treat it as a project tend to drift back.

The investment thesis that relies on cost improvement to drive returns should look here first. It's faster to execute than restructuring and far less disruptive.

Transformation starts to slip long before the technology fails

Digital transformation stalls when operating discipline is missing. Why programs slow down, how to restore ownership and decision cadence, and why mid-sized firms can capture big gains from today's pace of change.

Most digital transformation programs do not stall because the technology is wrong. They stall because the business does not change how it works around the technology.

The pattern is familiar. A company picks a platform, brings in a partner, launches a program, and the early stages look productive. Activity picks up. Workstreams multiply. Progress gets reported. But the business itself does not feel materially different. That is the warning sign.

Transformation is not implementation. It is not a roadmap, a governance structure, or a series of workstreams moving through a plan. It is a real change in how the business performs.

Faster cycle times. Lower cost to serve. Fewer manual handoffs. Better customer experience. More reliable operations. If those outcomes are not clear from the start and managed tightly, programs drift into activity without creating enough real change.

When leaders talk about a stalled transformation, they rarely describe a software problem. They describe slow decisions, unclear ownership, expanding scope, competing priorities, and security, compliance, or data issues showing up too late. They describe partners starting to shape the program instead of supporting it. They describe the day job taking over while transformation gets attention only when something slips.

That is not a technology failure. It is an operating failure.

Why Programs Stall

That distinction matters because it changes where leaders need to focus. Most stalled programs do not need another strategy deck, a revised target state, or broader governance. They need a tighter way of running the work.

One accountable owner. A clear decision cadence. Real trade-offs. Clear business outcomes. Scope discipline that holds. Adoption treated as a management job, not a communications exercise.

This matters even more now because the pace of technology is moving faster than many organizations are set up to absorb. Capabilities that once required a major platform overhaul can often be delivered through better configuration, stronger automation, and lighter integration. That should make transformation easier, but it only does so if the business can make decisions faster and turn those tools into practical operating change.

That is why this is not just a large enterprise issue. Mid-sized and smaller businesses face the same challenge, often with less room for waste and more to gain from getting it right. Modern tools can remove manual work, improve customer experience, and tighten operations without adding headcount. But the value does not come from the technology alone. It comes from the discipline to implement it in a way the business can absorb and sustain.

What Leaders Should Focus On

The strongest programs are usually simpler than the weaker ones. They have one leader with clear accountability and the authority to make trade-offs. They run on a regular cadence that clears blockers and forces decisions. They deliver in waves the business can actually feel, rather than managing toward abstract milestones. And they measure success through a small number of operational and commercial outcomes, not the volume of activity completed.

They also build internal capability as they go. That is one of the clearest signs that a program is creating real value. A transformation effort that stays permanently dependent on outside partners usually means the business implemented new tools without building the discipline to sustain them.

In 2026, the risk is not failing to invest in transformation. The greater risk is investing heavily and still not realizing value because execution discipline was never built into the program.

Most organizations do not need more strategy. They need clearer ownership, tighter scope, faster decisions, and outcome metrics tied to real operating change. If transformation is moving more slowly than it should, or the business is not feeling the impact, the issue is usually not the technology. It is the operating model around the program. That is where leaders should focus.

The board pack in an afternoon: what automation actually changed

Management reporting was the first workflow we automated in most businesses last year. What changed was not the speed. It was the arguments.

Every mid-market company has a version of the same ritual. In the last week of the month, finance and operations pull numbers from six systems into spreadsheets, reconcile the ones that do not agree, build the deck, and send it to leadership a day before the meeting. Then the meeting is spent arguing about whether the numbers are right.

We started automating this workflow because it is high-volume, repetitive, and painful. The speed improvement was expected. The other change was not.

What changed

The numbers arrived earlier, from the source systems, the same way every month. Nobody re-keyed anything. The reconciliation happened in the workflow, with the exceptions flagged rather than silently fixed. The commentary was written by a person, on time, because the person had the numbers three days earlier than before.

The meeting stopped being about the numbers and started being about the business. That is the return.

What it took

Less than most expect. Access to the source systems. A definition of every metric that everyone agreed on, which was the hardest part and the most valuable. A workflow that pulls, reconciles, and assembles. A person who owns the output.

Sixty days, in most cases. No new platform. The tools were chosen for the job after the process was defined, not before.

Where it leads

Once leadership trusts the pack, the same workflow feeds the board, the lender, and the sponsor. The finance team stops being a reporting function and starts being an analysis function. And the next workflow, whichever one it is, starts from a business that has already seen automation work.

Fewer agencies, more senior people. Why the small roster wins.

A big roster feels like coverage. It is usually overlap, junior staffing, and a coordination cost that nobody has counted.

The largest agency rosters we audit were never designed. They accumulated. A specialist for social. A local shop for a market. A production partner for a launch. A second media agency after a bad year. Each decision reasonable. The total, unmanaged.

What a big roster actually costs

Overlap: two or three agencies paid for versions of the same scope. Coordination: internal marketing time spent managing agencies rather than the brand. Dilution: every agency staffs the account with the seniority the fee supports, and a fee split ten ways supports juniors. Leakage: nobody sees total agency spend, so nobody negotiates it as a whole.

Why the small roster wins

Consolidating to fewer agencies with larger scopes changes the economics. The fee is large enough to justify senior people. The contract is important enough to the agency to be governed properly. The internal team manages two relationships instead of ten. And the negotiation happens at the level of total spend, with benchmarks, rather than agency by agency.

The work usually gets better. Senior people, clear scopes, and less coordination produce better output than a roster of specialists competing for attention.

How to do it without breaking things

Audit the roster against the business need, not against the org chart. Identify the overlap. Design the target structure first, then run the consolidation as a search where the incumbents can compete. Reset the contracts with account team seniority written in. Then govern it quarterly.

We have run this for global consumer companies and regional businesses alike. The pattern holds. Fewer, bigger, more senior, better governed.

Revenue Is Stalling and It's Not a Sales Problem

When revenue stalls, the first instinct is to fix sales. Often the problem is upstream, in strategy, pricing, positioning, or the go-to-market structure itself.

The Sales Team Gets the Blame First

Revenue misses create immediate pressure on sales leadership. The logic seems straightforward: revenue comes from selling, selling is done by the sales team, therefore underperformance is a sales problem. New quota, new territory structure, new head of sales. Sometimes a new commission plan.

This is sometimes right. Sales execution genuinely underperforms. Reps who can't qualify, managers who can't coach, a forecasting discipline that's aspirational. When the diagnosis is correct, the fix is correct.

But it's frequently wrong, and it's worth understanding why. Sales teams close deals that the rest of the business creates conditions for. Winning requires a product that solves a real problem, at a price the market will pay, communicated to people who need it, in a way that's credible and timely. If any part of that chain breaks, sales struggles, regardless of how talented the team is.

A great sales team cannot sell a product the market doesn't value, at a price it won't pay, into a segment that's been poorly defined.

Diagnosing the Actual Problem

The diagnostic starts with pipeline, not close rates. Where in the funnel does revenue stall? If leads are plentiful but qualify poorly, the problem is targeting, you're attracting the wrong audience. If qualified opportunities stall in the middle of the funnel, the problem might be pricing, product gaps, or competitive positioning. If opportunities stall at the close, that's closer to a sales execution problem.

Each of these has a different solution. Fixing close rates when the problem is targeting is expensive and ineffective. You train harder, coach more, and watch the same deals fall apart for the same underlying reason.

Talk to the customers who didn't buy. Win/loss analysis is one of the most underused tools in commercial diagnostics. The reasons that appear in internal CRM notes, "price," "went with competitor", are usually incomplete. A structured conversation with a lost prospect tells you far more.

Then look at the customer base. What's working? Which customers are renewing, expanding, and referring? That profile is the one to double down on. Often a revenue stall is partly a targeting problem, the business has drifted away from the segment where it's genuinely differentiated.

The Structural Problems That Look Like Sales Problems

Pricing is the first. A price point that worked during growth, when the market was less competitive, may now be wrong. Not necessarily too high, sometimes too low, which affects perceived value and attracts the wrong buyer. Pricing deserves a regular strategic review, not just a reaction to deals lost.

The marketing-to-sales handoff is the second. Leads that arrive too cold, with too little information, handed off too early, require more sales effort than they should. A better handoff, better qualified, better informed, better timed, makes the same sales team more effective without changing anything on the sales side.

The third is product-market fit drift. Markets change. Competitors enter. The problem you built a product to solve evolves, or your solution has been replicated at a lower price point. Recognising this requires honesty that's harder inside a business than outside it.

What the Fix Actually Looks Like

Revenue stalls are usually multi-causal. The fix is usually multi-part. Sales execution may need work and so may targeting and so may pricing and so may the handoff. The mistake is fixing only one and expecting the whole problem to resolve.

That's a leadership conversation, not a sales conversation. It requires the CEO and the commercial leadership to look at the full system. Sales is one part of it.

The most boring million in the company

Licence counts, renewal dates, and commitment levels. Nobody wants to own them. That is exactly why they are worth so much.

Ask a leadership team where the next million of savings is and they will talk about strategy, structure, maybe headcount. Ask the finance team where the software and cloud bills are going and you will get a shrug and a number that grows every quarter.

The million is in the shrug.

Where it hides

Licences assigned to people who left. Premium tiers for users who need the basic one. Cloud instances running at five percent because they were sized for a peak that never came. Reserved capacity bought for a workload that moved. Three tools doing one job because three acquisitions brought three stacks. Telecom lines to offices that closed.

None of it is dramatic. All of it is cash.

Why it stays hidden

It is boring, it is nobody's job, and the vendors are structured to keep it that way. Auto-renewals, true-ups that only go one direction, pricing that nobody benchmarks. Finance sees the bill; the function sees the vendor; procurement sees the purchase order.

How to get it

Inventory everything. Match every licence to a user and every commitment to a workload. Put a name and a renewal date on every contract. Benchmark the largest ten. Renegotiate in the notice window with data.

Then keep doing it, quarterly, because the drift starts again the day you stop. It is the least glamorous work in the company. In our experience it is also among the highest return per hour of anything a finance leader can sponsor.

When cloud spend starts eating margin: the operator's playbook

Cloud spend is hitting margins as stacks get heavier and costs rise without proportional gains. What operators are doing to regain control through consolidation, smarter workload placement, and stronger discipline.

Cloud is no longer just an IT line item. It is a margin lever, a resilience question, and increasingly, a test of operating discipline.

Across leadership teams in 2026, the same pattern is playing out. Cloud bills continue to rise. Technology estates are becoming more complex. Yet many businesses still struggle to show a clear relationship between spend and value, whether in performance, agility, resilience, or commercial return.

That does not usually happen because of one bad decision. It happens gradually.

A new tool is approved for one function. Another platform comes in through a vendor relationship. Development environments stay live longer than planned. Contracts roll forward without challenge. Over time, what started as flexibility becomes embedded cost. By the time it reaches the executive agenda, the conversation is rarely about architecture. It is about economics and control.

Why is cloud run-rate growing faster than revenue? Why does product delivery still feel slower than it should? Why are there so many overlapping vendors, tools, and contracts? And if a major provider experiences an incident, how concentrated is our exposure?

These are no longer technical questions. They are operating model questions.

The Resilience Conversation Has Moved On

Over the last two years, senior operators have had repeated reminders that even world-class infrastructure is not immune to disruption.

-

A Microsoft Azure authentication outage temporarily prevented organizations from accessing systems that rely on Azure identity services.

-

A Google Cloud service disruption affected enterprise applications running on its platform.

-

The global CrowdStrike incident in 2024 disrupted airlines, banks, hospitals, and enterprise systems tied to interconnected infrastructure.

-

And in 2025, an AWS outage disrupted multiple websites and services relying on its cloud infrastructure.

None of this suggests hyperscale platforms are fundamentally unreliable. It reflects a simpler reality: complex systems fail, and concentrated dependency amplifies the impact when they do.

That is why the resilience discussion is changing.

The most effective operators are not pursuing complexity for its own sake. They are becoming more deliberate about where concentration risk is acceptable, and where it is not. Critical workloads, core data, identity layers, and customer-facing services are being reassessed through a different lens: not just cost and performance, but business continuity. The result is not a blanket move to multi-cloud. It is a more selective strategy. Diversify where failure would materially damage the business. Simplify everywhere else. That is a far more practical position than treating resilience as a theoretical architecture debate.

Where the Best Operators Are Finding Value

The strongest leadership teams now assess cloud through three practical lenses:

-

Performance - Are systems delivering the speed and reliability the business needs?

-

Resilience - Can the organization recover quickly when systems or providers fail?

-

Cost discipline - Can leadership explain what is driving spend and how it supports growth?

When those answers are vague, cloud stops looking like an investment and starts looking like silent margin erosion. Cloud infrastructure should support growth, resilience, and speed, not become an unchecked operating burden.

The biggest improvements rarely come from large-scale transformation. They come from a small number of disciplined decisions that materially improve cost, resilience, and performance.

1. Stop overpaying for standard workloads

Not every workload needs premium services or maximum redundancy. Many just need reliable performance and the right controls. The opportunity is to match infrastructure to business criticality.

2. Reduce tool and vendor sprawl

Cloud costs extend well beyond infrastructure. Overlapping tools, platforms, contracts, and vendors create unnecessary cost and complexity. In many cases, the problem is duplication.

3. Make cloud economics visible

Cloud spend should be measured in business terms: per workload, environment, transaction, or customer. That visibility turns cloud from a finance issue into an operating discipline.

4. Be more precise with AI infrastructure

AI workloads vary widely. Treating them the same drives overspend. Compute should be aligned to the actual use case, performance need, and business value.

5. Treat resilience as a business decision

Continuity, recovery, and concentration risk are not just technical issues. They have direct commercial impact and should be managed accordingly.

What This Means for Leadership

Most organizations do not need a full cloud reset. They need sharper visibility, tighter governance, and a workload strategy aligned to cost, performance, and risk. The goal is not less cloud or more complexity. It is a technology estate that is commercially sound, resilient, and built to scale. That is the difference between cloud as a growth enabler and cloud as a margin drain.

Measurement is a governance problem, not a data problem

Every platform will happily measure itself. The question is who owns the model that decides where the next dollar goes.

Marketing teams talk about measurement as a data challenge. Better attribution, cleaner pipes, more signals. Those matter. But in most businesses the real problem is simpler and harder: the parties who report performance are the parties who profit from it.

The platform reports what the platform delivered. The agency reports on the plan the agency built. The attribution model was configured by whoever set up the account, often years ago, and it credits the channels that are easiest to credit.

Then the budget follows the report.

What that produces

In one recent audit, the legacy first-touch model was crediting programmatic with roughly twice its actual contribution. The team was not wrong to trust it; it was the only number they had. But the budget had been shaped by it for years.

The governance fix

Someone on the client side owns the measurement model, and it is not the agency. The model is written down, with its assumptions. Every major platform's reported numbers are reconciled against it. Incrementality is tested where it matters, even when a partner would rather it were not. And the budget decision is made from the reconciled view, not from the deck.

That is not a technology project. It is a decision about who is allowed to mark their own homework. Make it, and the data problems become tractable. Skip it, and no amount of data will help.

What a good fifteen-minute call sounds like

We start every engagement with a short conversation. Here is what we are listening for, and what you should expect back.

Most consulting relationships begin with a pitch. Ours begin with fifteen minutes and a question: what is not working?

We are not selling in that call. We are trying to find out, quickly, whether the problem you have is one we solve, and whether the way we work fits your situation. About a third of the time the honest answer is that it is not, and we say so.

What we are listening for

A cost line or a function that has grown faster than the governance around it. A number that leadership does not trust. A vendor, agency, or platform relationship that nobody has looked at commercially in years. A deal, a 100-day plan, or an exit with an operational component. A leadership team that wants the work done, not described.

What you should expect back

A straight answer on whether we can help. If we can, a view of what the engagement would look like: scope, timeline, what we would need access to, and how we charge. Usually a first step that is small: a diagnostic, a single renewal, an audit of one budget.

If we cannot, a suggestion of who might.

What we do not do

We do not send a deck. We do not follow up with a sequence of emails. We do not need a discovery phase to tell you whether the pattern matches. We have seen it before, or we have not.

That is the whole call. It is the best fifteen minutes we know how to spend, and most of the people who take it tell us the same.

Outdated marketing contracts are quietly draining performance

Outdated marketing contracts inflate costs and weaken accountability. Where value leaks, how to reset agency and platform models, and practical steps to make marketing spend easier to defend.

Most companies do not overspend on marketing. They overpay because contracts outlive their value.

Too often, agreements remain in place for four, five, or six years with only minor adjustments along the way. In some cases, they have gone much longer. That matters because marketing has changed quickly. Measurement is more precise, channels shift faster, automation is now embedded in delivery, and many agency models still reflect an earlier operating reality.

When contracts are not revisited every two to three years, costs compound. Scope expands quietly. Legacy structures stay in place long after the value has faded. What looks manageable in isolation becomes a recurring drag on both performance and profitability.

With global advertising spend projected to reach $1.3 trillion in 2026 (WARC), even small inefficiencies now carry real financial weight.

Marketing Leaders Are Being Held to a Higher Standard

Senior marketers are expected to defend spend with increasing precision.

One CMO at a financial institution described the budget process as relentless. To secure a 6% increase, he had to justify initiatives, channel mix, agency scope, headcount, tools, measurement plans, and expected business impact. That level of scrutiny is reasonable. The problem is that the same standard often does not extend across the partner setup supporting the work.

Agencies, platforms, and suppliers do not always provide enough clarity on what is driving results, what has been automated, where senior talent is applied, or how effort translates into business outcomes. As a result, accountability becomes harder than it should be.

This is where value leakage begins. Not in one dramatic decision, but in a series of arrangements that are no longer aligned to how modern marketing actually works.

Where Value Leaks First

1. Paying for effort instead of outcomes

Many contracts still reward volume. Retainers, rate cards, and staffing models often incentivize activity, whether or not that activity creates incremental value. If the commercial model rewards output, output is what it will generate. That does not always mean the business is buying impact.

The stronger models now tie economics more closely to outcomes with better visibility into what is automated, what requires senior judgment, and where strategic value is actually being created.

2. Temporary scope becomes permanent cost

Support that was meant to be short term often becomes embedded.

Additional reporting, production layers, specialist projects, and platform support are added over time and rarely removed with the same discipline. The result is a larger retainer, more complexity, and limited challenge around whether the work still matters.

3. Too much spend on execution, not enough on leadership

Execution is becoming easier to scale. Senior strategic thinking, creative leadership, and accountable stewardship are not. Yet many models still overfund delivery layers while underinvesting in the senior talent that drives better decisions and stronger commercial outcomes. That is where contract value often falls out of line with business need.

4. Tool sprawl compounds quietly

The LUMAscape shows how crowded the setup has become. Every platform and partner has a cost, but not all deliver proportional value. With AI-native providers emerging and capabilities evolving quickly, the stack now needs more frequent review than it did in the past.

What Better Contracts Look Like

The strongest marketing contracts do four things well:

-

Clear accountability. Define a small number of meaningful outcomes and measure them consistently.

-

Transparent economics: Make scope, pricing, and cost triggers explicit.

-

Flexibility: Allow the contract to evolve with changing priorities.

-

Governance: Review spend, performance, and decisions together on a regular cadence.

None of this is complicated. But it does require discipline, and that discipline is often missing when long-standing contracts are allowed to roll forward by default.

Where the Best Gains Usually Show Up

The biggest gains usually come from a small number of practical moves:

-

Consolidating overlapping agencies, platforms, and vendors.

-

Repricing agreements that have drifted away from market reality.

-

Resetting performance accountability.

-

Stopping always-on activity that no longer drives impact.

-

Reducing tool overlap and reinvesting in what is actually used.

-

Shifting spend toward senior leadership and away from low-leverage delivery layers.

These are not dramatic changes. They are commercial corrections. But they often unlock meaningful savings and better performance at the same time.

Where Leaders Should Focus Now

Most organizations do not need to rebuild their marketing setup. They need a clear view of where contract value has eroded, where scope has expanded, and where spend is no longer tied tightly enough to outcomes.

The objective is not cost reduction alone. It is a partner model that is transparent, accountable, and aligned to how marketing value is actually created today. That is the difference between contracts that support growth and contracts that quietly weaken return.