How to Justify Agency Spend to Your CFO (and Find Agencies With Real Fashion Case Studies)

Key Takeaways

• A CFO does not fund ROAS, reach or impressions — they fund incremental contribution and a defensible payback period, so build the case in those terms or lose the room.
• The four numbers that win the meeting: fully-loaded CAC, contribution margin, CAC payback, and proven incrementality (caused, not correlated). Miss any one and the case has a hole a CFO will find.
• In fashion, contribution must be calculated after returns, markdowns and discounts — a healthy ROAS routinely sits on top of an unprofitable order once those are counted.
• Incrementality is what turns a marketing claim into a finance-grade result: a holdout, a geo test or an MMM read proves the lift was caused, and a CFO has seen 'we spent more and revenue went up' fall apart before.
• Benchmarks persuade only when honest — use ranged, sourced, clearly-labelled figures and anchor on your own baseline; a defensible smaller number beats an impressive unverifiable one.
• A credible agency case study names the brand or specific category and scale, states the baseline it improved from, explains the method, reports contribution or payback rather than only ROAS, and survives a reference call. A logo, a big number and no method is marketing, not evidence.
• Justifying the spend and choosing the agency are the same discipline: the agency worth funding is the one that already thinks in your CFO's terms and can prove it — so if an agency can't support your business case, it won't deliver the result either.

Your CFO Isn't Asking About ROAS — They're Asking a Finance Question

When your CFO pushes back on the agency line item, it is tempting to hear it as skepticism about marketing. It is not. It is a finance question wearing marketing clothes, and the reason so many marketers lose this conversation is that they answer the marketing question they imagined instead of the finance question that was actually asked. The CFO wants to know three things, in order: does this spend produce more profit than it costs; is that outcome reliable and defensible enough to put in a plan the board will scrutinise; and can you prove the growth would not have happened anyway without it? Answer those three, in finance's own units, and the meeting is short. Answer with ROAS, impressions, reach, engagement or 'brand lift', and you will lose — not because those metrics are worthless, but because none of them can be entered into a model, defended upward, or reconciled to the ledger.

This is the single most important reframing in the entire discipline of justifying agency spend, so it is worth stating bluntly: you are not defending a marketing budget, you are proposing a capital allocation. A CFO allocates capital every day — to inventory, to headcount, to capex — and they do it on a consistent basis: expected return, payback period, risk, and the quality of the evidence behind the forecast. Your agency proposal is just another capital allocation competing for the same rupees, and it will be judged on exactly those criteria. The marketer who walks in with a return, a payback, a risk assessment and credible evidence is speaking the CFO's language; the marketer who walks in with a ROAS chart is speaking a foreign one and wondering why they are not understood.

The good news is that this is entirely learnable, and it is not about dumbing marketing down — it is about translating it correctly. Every genuine marketing outcome can be expressed in the CFO's terms without losing anything true; what gets lost in translation is only the stuff that was never rigorous to begin with. When you translate 'we drove a 4x ROAS' into 'we produced X of incremental contribution against Y of fully-loaded cost, with a payback of Z months, proven with a holdout', you have not weakened the claim. You have made it fundable. The rest of this guide is that translation, metric by metric, plus how to find agencies whose documented results can actually support it.

One more framing point, because it changes the emotional dynamic of the meeting. When you arrive with a finance-grade case, the conversation stops being adversarial — marketer defending, CFO attacking — and becomes collaborative, because you have handed the CFO a proposal they know how to evaluate and, if it is sound, how to approve and defend to the board. CFOs are not the enemy of marketing spend; they are the enemy of unjustifiable spend. Give them something justifiable and you convert your CFO from a gatekeeper into an ally, which is worth far more than winning any single budget request.

Why Fashion Makes This Harder — and Why That's an Opportunity

Everything above is true for any category, but fashion turns the volume up, because fashion economics are among the most unforgiving in retail and a fashion CFO has usually been burned by exactly the trap this guide helps you avoid. Fashion carries structurally high return rates — a meaningful share of what is 'sold' comes back, so reported revenue and collected revenue diverge sharply. It runs on heavy, habitual discounting — end-of-season markdowns, festival sales, first-order incentives — so the price that drove a conversion is rarely the full price. It is intensely seasonal, so a channel's performance in one month tells you little about the next. And after cost of goods, which in apparel can be a large fraction of price, the contribution left per order is often thin. Put those together and you get the defining fashion trap: a campaign can post a perfectly healthy ROAS and still lose money on every order once returns, markdowns and COGS are counted.

A fashion CFO knows this in their bones, which is why a case built on ROAS will be met not with curiosity but with justified suspicion. They have seen the brand scale a 'profitable' channel and watched the bank balance fall, because the ROAS was profitable and the contribution after returns was not. So when a marketer walks into a fashion finance meeting leading with ROAS, the CFO's internal translation is 'this person does not understand our economics, so I cannot trust their forecast'. You lose credibility before you have made your case.

But here is why that difficulty is actually your opportunity. Precisely because most marketers and most agencies argue fashion in ROAS, the marketer who argues it in contribution-after-returns and payback-against-the-season stands out immediately as someone who gets it. You are speaking the CFO's language about their specific pain, and that alone earns a level of trust that no amount of ROAS gloss can buy. In a category where the finance function is chronically skeptical of marketing, the marketer who arrives fluent in the real economics is rare enough to be memorable — and fundable. The discipline that fashion forces on you is exactly the discipline that wins the room.

The Four Numbers That Win the Room

There are four numbers a CFO actually cares about, and your entire business case should be built from them. Learn to produce each honestly and you can justify agency spend in any category; produce them for fashion, after returns, and you can justify it in the hardest category there is. Let us take each in turn, because the detail is where credibility lives.

The first is fully-loaded customer acquisition cost. Not the platform's cost-per-purchase, and not media divided by orders — the true, all-in cost to acquire a customer as finance would compute it, including the media, the agency fees, the tooling and any incentive or discount used to close the sale. Marketers routinely quote a CAC that quietly excludes the agency retainer and the tech stack, understating the real number by twenty to forty percent, and a CFO will find that gap in about ninety seconds because agency fees are a line item they already see. The moment they catch it, your credibility is gone, because it looks like you were either sloppy or hiding the cost. Lead with the honest, fully-loaded number. It will look worse than the platform's figure, and that is exactly why the CFO will trust it — a marketer who volunteers the unflattering, complete number is a marketer whose other numbers can be believed.

The second is contribution margin — the money a sale actually generates for the business after every variable cost, not the revenue it books. Contribution is revenue minus cost of goods sold, minus shipping and fulfilment, minus payment and platform fees, minus discounts, and — in fashion, non-negotiably — minus the cost of returns. This is the denominator that determines whether a given CAC is profitable, because a customer acquired for a CAC of X is only worth acquiring if they generate more than X in contribution over the relevant horizon. A CFO thinks entirely in contribution, never in revenue, so you must present the agency's impact as incremental contribution, not incremental sales. Revenue that carries no contribution is not a benefit to the business; it is often a cost, and presenting revenue growth to a CFO as if it were profit growth is one of the fastest ways to mark yourself as financially naïve.

The third is CAC payback — how many months it takes to earn back the fully-loaded CAC from the contribution a customer generates. This is frequently the number a CFO cares about most, because it is not really about profitability, it is about cash and risk. A short payback means the capital you tie up acquiring customers comes back quickly and can be redeployed, so you can scale aggressively with limited cash-flow risk. A long or unknown payback means capital is locked up for a long time and the business is exposed if the forecast is wrong, so caution is rational. Modelling payback explicitly — and, in fashion, modelling it against the season, since a customer acquired in October behaves differently from one acquired in a January sale — is what lets a CFO decide how much and how fast to fund. Give them the payback and you give them the lever they actually use to size the investment.

The Fourth Number — Incrementality — Is the One That Separates Credible From Hopeful

The fourth number is the one most marketers skip, and it is the one that turns a marketing claim into a finance-grade result: incrementality. Can you show that the growth was caused by the spend, rather than merely correlated with it? This is the question a good CFO asks that a marketer least expects, and it is devastating if you have no answer, because 'we spent more and revenue went up' is not evidence — it is a coincidence that has not been ruled out. Sales might have risen because of seasonality, a competitor's stockout, a new product, a PR moment, or simple mean reversion, and the ad spend might have taken credit for demand that already existed and would have converted anyway. A CFO has watched last-click attribution and platform-reported ROAS take credit for exactly this kind of non-incremental demand, and then watched the numbers collapse when the spend was cut and revenue did not fall. So they discount uncaused claims heavily, and they are right to.

There are three credible ways to prove incrementality, in rough order of rigour and effort. The first and cleanest is a holdout test: withhold the marketing from a randomly selected, representative slice of your audience or geography, and measure the difference in outcomes between the exposed group and the held-out control. The gap is the incremental effect, causally attributable, because the only systematic difference between the groups was the marketing. This is the gold standard, and even a modest holdout — a small percentage of budget or geography withheld for a defined window — produces a number a CFO will respect far more than any attribution model, because it is an experiment, not an inference.

The second is a geo experiment: turn a channel or campaign up (or off) in some geographies and not in matched comparison geographies, and compare the outcomes. This is a practical form of holdout that works well when audience-level holdouts are hard, and it is especially useful for testing whether a channel like paid social or CTV is genuinely additive or is cannibalising demand that search would have captured anyway. The third, for brands with enough history and channel variation, is media-mix modelling (MMM) — a statistical model that estimates each channel's incremental contribution from historical spend and outcome data, controlling for seasonality, price and other drivers. MMM is less precise than a live experiment and more susceptible to modelling assumptions, but it can read the whole portfolio at once and does not require withholding spend. The rigorous answer is usually a combination: experiments to calibrate, MMM to allocate. Whatever method you use, arriving at the CFO meeting with an incrementality plan — 'here is how we will prove, after the fact, that this spend caused the growth' — is what makes your forecast defensible, and defensibility is the whole game. This is the measurement backbone of any serious performance marketing engagement, and its absence is the single biggest tell of an agency that cannot be held to results.

It is worth being honest with your CFO about what incrementality reveals, because it often reveals that some of your reported ROAS was never incremental — that a chunk of the 'return' was demand you would have captured for free. A weaker marketer hides from this; a stronger one leads with it, because a CFO who sees you actively hunting for and removing non-incremental spend trusts you far more than one who presents an unrealistically clean story. The willingness to find and admit that some spend was not working is, paradoxically, the strongest possible signal that the spend you are defending genuinely is.

A Worked Business Case: How the Four Numbers Fit Together

To make this concrete, here is how the four numbers assemble into a single proposal — illustrative, with round figures chosen only to show the logic, not as benchmarks. Suppose you propose to invest in an agency to run and scale paid acquisition for a fashion brand. You do not walk in and say 'we want to spend more on ads and we think ROAS will be good'. You walk in with this: 'We propose an incremental media investment, plus the agency retainer and tooling, for a fully-loaded cost of A per month. Based on our current, measured unit economics, a customer generates B in contribution after COGS, shipping, fees, discounts and returns over the first purchase cycle, with repeat behaviour adding further contribution over the customer's life. At our target scale the fully-loaded CAC is C, which gives a CAC payback of D months against our season. We will prove the growth is incremental with a geo holdout in the first quarter, so within one season we will have a caused, not correlated, read on the return. Here is the sensitivity: if CAC comes in twenty percent higher than plan, payback extends to D-prime months, which is still within our tolerance; if returns run higher than the modelled rate, here is the effect on contribution. We are asking to fund it for one season with a defined kill criterion if the incrementality read is below threshold.'

Read that back and notice what it is: it is not a marketing pitch, it is a capital proposal with a return, a payback, an evidence plan, a sensitivity analysis and a downside kill criterion. That is precisely the shape of proposal a CFO approves for inventory or capex every quarter, and presenting agency spend in the same shape does two things at once. It makes the spend fundable, because it can be evaluated on the CFO's normal basis. And it makes you credible, because you have demonstrated that you understand the spend the way finance does — as risk-adjusted capital, not as a marketing entitlement. The sensitivity analysis and the kill criterion matter especially: volunteering the downside and pre-committing to stop if the evidence is bad is what separates a marketer asking for trust from a marketer who has earned it.

The deeper reason this works is that it inverts the burden. In the ROAS conversation, the marketer is asking the CFO to trust an unfamiliar metric, and the CFO's rational default is 'no'. In the capital-proposal conversation, the marketer has done the CFO's work for them — framed the decision, quantified the return and risk, and provided the evidence plan — so the CFO's rational default shifts toward 'yes, if the numbers hold'. You have not argued the CFO into approval; you have made approval the obviously correct decision given the evidence, which is the only kind of 'yes' that survives the board meeting afterward.

Using Benchmarks Honestly (So They Survive Scrutiny)

Benchmarks are where most agency business cases quietly self-destruct, because the instinct is to reach for an impressive, precise-sounding external number — 'the industry average ROAS is X' or 'good CAC for fashion is Y' — with no source and no context. A competent CFO distrusts a suspiciously specific benchmark on sight, and rightly, because they know that averages across brands with wildly different margins, price points, return rates and discounting strategies are close to meaningless as a target for any specific brand. The moment you present an unsourced benchmark and the CFO asks 'where did that come from?' and you cannot answer, every other number in your case becomes suspect by association. One weak benchmark can sink a strong proposal.

The honest way to use benchmarks is as ranged, sourced, explicitly-labelled directional context — 'published sources put typical payback for this category in this band; here is the source; our own measured payback is here, inside (or outside) that range, and here is why'. Ranges signal that you understand the variance; sources signal that you are not making it up; and labelling them as directional signals that you know an external average is context, not a target. A defensible smaller claim beats an impressive larger one every time, because the impressive one dies the instant it is questioned and the defensible one does not. In fashion specifically, be especially wary of any 'good ROAS' benchmark: presenting one as a target is exactly how brands talk themselves into scaling an unprofitable order, so if you must reference ROAS benchmarks at all, immediately reframe them in contribution-and-payback terms and caveat that your own unit economics, not an industry average, define what 'good' is for you.

The single most persuasive benchmark, though, is not external at all — it is your own baseline. Before-and-after your own measured CAC, contribution and payback, established with incrementality, is more convincing to a CFO than any published figure, because it is your data, in your model, that they can audit line by line. External benchmarks set the stage and provide sanity checks; your own baseline wins the argument. So the strongest structure is: establish your baseline rigorously, cite ranged external benchmarks only to show your baseline is not wildly out of line with reality, and then make the entire case about moving your own numbers. That is a case a CFO cannot easily attack, because there is nothing borrowed in it to attack — it is your business, measured honestly, with a plan to improve it.

How to Find Agencies With Genuinely Documented Fashion Case Studies

Once you have the framework, you need agencies whose documented results can actually support the case, and here you must be ruthless about the difference between a case study and a marketing artefact, because the two look superficially identical and are worlds apart in value. There are genuinely strong performance agencies with public retail and fashion work — among the larger performance shops, names like Tinuiti, Brainlabs, Wpromote and Media.Monks appear regularly, and among specialist e-commerce and D2C operators, agencies such as Common Thread Collective and Pilothouse are well known — all described here by their public positioning, not as endorsements or a ranking. But the agency's brand is not the evidence you need; the credibility of the specific, relevant case study is. A famous agency with a vague case study is worth less to your CFO than an unknown one with a rigorous, verifiable, category-matched result.

Where to look, in rough order of trustworthiness. Client references you obtain and call yourself are the gold standard, because a client answering your direct questions cannot be curated the way a published case can. Documented industry-awards entries (the Drum Awards, the Performance Marketing Awards, effectiveness awards like the IPA or WARC) are strong, because entries are written to a standard, judged, and often required to disclose method and results. Platform partner directories and their published success stories — Meta's and Google's official case studies — carry some weight because the platforms have reputational skin in the game, though they are naturally selective. Third-party review platforms such as Clutch and G2 give you unfiltered client voice, useful for triangulation. And the agency's own case-study library is the least trustworthy source on its own — not worthless, but to be read critically and cross-referenced, never taken at face value.

The most powerful move is cross-referencing: a result that shows up in an awards entry, a platform case study and a client reference call, all telling a consistent story, is far more trustworthy than one that lives only on the agency's homepage. Discrepancies between sources — a number that is bigger on the agency site than in the awards entry, a client who remembers the engagement differently — are exactly the signal a CFO would want you to catch before you stake budget on the agency. Do the cross-referencing before the finance meeting, not after, because 'we vetted their claims across four independent sources' is itself a line that builds your credibility with a CFO who values diligence.

The Anatomy of a Real Fashion Case Study (and the Fluff to Reject)

You need a reliable test to separate a real case study from marketing, and in fashion specifically there are five tells, each of which a fabricated or hollow case study fails. First, specificity of the subject: a real case study names the brand, or at minimum its exact category and scale ('a mid-market women's footwear brand doing X in annual online revenue'), because a result is only interpretable if you know whose result it is. A case study attributed to 'a leading fashion brand' with no further detail is telling you it cannot be verified.

Second, a stated baseline. The number that matters is not the flattering end state but the distance travelled from a defined starting point — 'improved contribution-per-order from A to B', not just 'achieved a great ROAS'. A case study that reports only the destination and hides the origin is hiding the most important information, because a great absolute number might represent a tiny improvement on an already-great baseline, or the baseline might have been artificially depressed. Third, an explained method: a credible case study tells you what the agency actually did — the specific changes, the sequence, the reasoning — so the result is explicable rather than magical. 'We rebuilt the feed segmentation, moved bidding to contribution-based value signals, and fixed the returns leak in the size guide' is evidence; 'we optimised their account and delivered results' is a sentence with no information in it.

Fourth, the right metric. A real fashion case study reports the metric that decides the business — incremental contribution, cost per profitable order, payback — or at least reports ROAS alongside an honest acknowledgement of returns and margin, rather than trumpeting a ROAS number as if it were profit. A case study whose headline is a ROAS figure with no mention of returns is either naïve or selective, and in fashion that is disqualifying. Fifth, and most decisive, a reference: the agency will connect you with the client to verify the story. An agency that produces a glowing case study but cannot or will not put you on a call with the client is showing you the case study is not solid enough to survive contact with the person it is about. Bring the case studies that pass all five tests to your CFO; leave the rest on the pitch deck where they belong. A logo wall, a wall of big percentages, and no method, baseline or reference is not a portfolio of evidence — it is a portfolio of assertions, and a CFO funds evidence, not assertions.

Red Flags in Agency Proof — and the Reference Call That Settles It

Certain patterns should make you more skeptical, not less, however impressive the surface. Be wary of case studies that lead exclusively with ROAS and never mention contribution, returns or margin — in fashion this is a tell that the agency either does not understand the economics or is choosing the flattering metric. Be wary of round, suspiciously clean numbers with no baseline and no method — real results are messy and specific. Be wary of an agency that cannot name a single thing that went wrong in an engagement, because every real engagement has setbacks, and an agency that presents an unbroken record of flawless wins is presenting a sales narrative, not a track record. And be especially wary of any reluctance to provide references, or references that are exclusively current, happy, hand-picked clients with no churned client in the mix.

Which brings us to the reference call, the single highest-value diligence step and the one most marketers skip out of politeness. Ask the agency for references, and specifically ask to speak with at least one client who has left, not only current ones — an agency confident in its execution will facilitate this, and one that resists is telling you something. On the call, ask the questions that reveal reality rather than reputation: What did the agency actually do, in concrete terms? What were the results, and how were they measured — did they prove incrementality, or report platform ROAS? What went wrong, and how did the agency handle it? Were they responsive and proactive, or did you have to chase them? Did the senior people who pitched stay involved, or did the account get handed to a junior team? Would you hire them again, and for what specifically would you not? Those answers, from a real client speaking freely, tell you more about whether the agency can support your CFO business case than any deck ever will — and the notes from that call are themselves powerful evidence to bring into the finance meeting.

Assembling the Case: Your CFO Meeting Checklist

Put the whole thing together and the meeting almost runs itself, because you have pre-built everything a CFO needs to say yes. You arrive with: the specific business outcome stated in finance terms (incremental contribution and the payback you are targeting, not ROAS or reach); your own current, measured baseline for CAC, contribution and payback, established honestly and after returns; an explicit incrementality plan describing how you will prove, after the fact, that the spend caused the growth (holdout, geo test or MMM); honest, sourced, ranged external benchmarks used only for context and sanity-checking; a sensitivity analysis showing what happens if CAC or returns run worse than plan; a pre-committed kill criterion so the CFO knows the downside is bounded; and a shortlist of agencies whose documented, cross-referenced, reference-verified fashion case studies demonstrate they can actually deliver it. That is not a marketing request; it is a capital-allocation proposal with the diligence already done, and a CFO knows how to approve those.

The deepest insight in this entire guide is that justifying agency spend and choosing the right agency are not two tasks but one, because they are governed by the same standard. The agency worth funding is precisely the one that already thinks in your CFO's terms — that manages to CAC payback and contribution rather than ROAS, that measures incrementality rather than claiming credit, that owns the measurement so the numbers are auditable, and that can prove its record with real case studies and references rather than a deck. If an agency cannot support your business case, it is not merely bad at pitching to finance; it will not deliver the result either, because the discipline that makes an agency fundable is the same discipline that makes it effective. So choose the partner your CFO would choose if they sat in the pitch and asked the hard questions — the one whose answers are specific, whose numbers are contribution-based, and whose references check out — and the justification writes itself, because you will be funding an agency that was always going to produce a fundable result.

A Note on Methodology, Sources and Fairness

A word on how to read this, because credibility is the whole point of a piece about credibility. This is an opinionated guide published by Fluxsy, not an independent ranking or a paid audit. Where we name other agencies or tools we describe them only by their genuine, publicly stated positioning and well-known strengths; nothing here is intended to disparage a competitor, and any brand named may be an excellent choice for the right client and the wrong one for another. We have deliberately not invented statistics, client results, benchmark numbers or endorsements — an invented figure is worse than no figure, because it collapses the moment a CFO or a prospect asks 'how do you know that?'. Where we give benchmark ranges they are framed as directional and sourced, and the strongest benchmark in every case is your own measured baseline.

The durable value of this guide is the evaluation framework, not a shortlist. Frameworks survive when specific vendors, prices and rankings do not: the questions to ask, the metrics that matter, and the tests that separate substance from a good pitch hold true no matter which partner you ultimately choose — including if that partner is not us. Use the framework, verify every specific claim (ours included) against primary sources and your own data, and treat any agency's confidence as a hypothesis to be tested rather than a fact to be trusted.