Marketing and finance are using the word revenue to mean different things
Before you can argue about attribution you have to agree what a sale is, when it happened, and how much it was worth. Most businesses have never had that conversation, which is why the marketing number and the finance number never match and everyone assumes somebody is wrong.
A marketing team presents attributed revenue of nine point one million. The finance pack for the same quarter says eight point four. Nobody has done anything wrong, nobody can immediately explain the gap, and the next forty minutes are spent on reconciliation instead of on marketing. This happens in most businesses, repeatedly, and it is treated as a data quality problem when it is almost always a vocabulary problem.
Marketing and finance do not disagree about the numbers. They disagree about what the word revenue means, and neither has noticed that the other is using it differently.
Settling this is not glamorous work and it is the precondition for everything else. An attribution figure that cannot be reconciled to the accounts will never be used to allocate budget, however good the underlying analysis.
The six places the definitions diverge
Marketing and finance revenue figures diverge on timing, refunds, tax, discounts, contract treatment and entity scope. Each is individually small and they compound quickly.
| Dimension | Marketing usually means | Finance usually means |
|---|---|---|
| Timing | The date the deal was won | The date revenue is recognised |
| Refunds | Ignored, or handled inconsistently | Netted off in the period of the reversal |
| Tax | Often included in the deal value | Always excluded from revenue |
| Discounts | Sometimes list price | Always net of discount |
| Multi-year contracts | Total contract value | Spread across the term |
| Entity scope | Everything the team worked on | Only the reporting entity |
Any two of those can produce a gap large enough to derail a meeting. All six together routinely produce differences of ten to twenty percent, which is comfortably larger than the effects most marketing analysis is trying to detect.
The tax row deserves a special mention because it is the most embarrassing and the most common. A marketing report built from deal values that include sales tax will overstate revenue by the tax rate, uniformly, and it is remarkable how long that can go unnoticed when nobody is reconciling.
Why finance should win
The definitional authority belongs to finance, because their figure is the one the business is externally reported and internally managed on. Marketing adopting the finance definition costs almost nothing and removes the entire class of dispute.
This occasionally meets resistance on the grounds that the finance definition is less useful for measuring marketing, and there is something to that. Recognised revenue spread over three years is genuinely awkward for evaluating a campaign that ran in March. But the resolution is a stated bridge, not a competing total.
The practical arrangement that works is to analyse on bookings, because bookings sit closest to the marketing event, and to reconcile bookings to recognised revenue explicitly on the same page. That gives marketing the analytical basis it needs and gives everyone else a line they recognise.
What does not work is marketing maintaining its own revenue number and defending it. That position is lost in advance, because the chief executive will believe the finance figure, and the effort spent defending the alternative is spent discrediting the analysis attached to it.
The conversation to have, once
Ask finance for their definition of a closed sale, the date field they use, the treatment of refunds, tax and discounts, and the contract convention. Write the answers down. It takes one meeting and does not need repeating.
- Which system is the source of truth for closed revenue, and which date field defines the period?
- Is the figure gross or net of refunds, and in which period does a reversal land?
- Is tax included, and are discounts applied before the figure we would see?
- How are multi-year and instalment contracts valued for management reporting?
- Which entities and business units are in scope, and are there any exclusions we should know about?
Five questions, one meeting, and a page of notes that will settle every future argument. The reason this rarely happens is not difficulty; it is that neither side realises there is anything to ask, because each assumes revenue is a word with one meaning.
Bookings, recognised revenue, and the bridge between them
Analyse marketing performance on bookings and present a reconciliation to recognised revenue. The bridge is usually three or four lines and it converts a suspicious discrepancy into an explained one.
A bridge is simply a list of the differences: bookings in the period, less amounts not yet recognised, plus amounts recognised from prior periods, less refunds, less tax. Ending at the finance figure. It takes a few lines on a slide and it demonstrates that you understand their number rather than competing with it.
It also protects the analysis when the two diverge for legitimate reasons. A strong booking quarter that will not appear in recognised revenue for six months is a good news story that looks like nothing at all in the finance pack, and the bridge is how it gets told.
The date field problem
Most CRMs contain several fields that could reasonably be called the close date, and they can differ by weeks or months. Picking the wrong one shifts revenue between periods and makes every trend unreliable.
The candidates are typically the date the opportunity stage changed to won, the date the contract was signed, the date the invoice was raised and the date payment was received. In a well-run business these cluster; in most businesses they do not, particularly where deals are marked won optimistically before paperwork completes.
The rule is to use whichever field finance uses, even if it is the least convenient one. If they recognise on invoice date, attribution should use invoice date, and the resulting misalignment with the marketing event is handled by the cohort approach rather than by choosing a different field.
Refunds, and why gross ranking misleads
Channels differ systematically in how durable their revenue is. Reporting gross revenue by channel will rank a source producing high-reversal customers above one producing loyal ones, and the ranking will be confidently wrong.
The effect is largest in categories with meaningful return, cancellation or churn rates, where the cheapest acquisition often produces the least durable customers. A discount-led channel can look excellent on gross bookings and poor on net revenue twelve months later, and only one of those views should be driving budget.
The mechanics are straightforward once the decision is made: net the reversal against the channel that generated the original sale, in whichever period finance books it. What makes this hard is not the arithmetic but remembering to ask for the reversal data at all, since it is rarely on the standard export.
What finance actually wants from marketing
Finance is not looking for sophistication. They want a figure that ties to theirs, a method that does not change, an explicit statement of what is not measured, and the ability to trace a specific number to source.
This is worth hearing clearly, because marketing teams frequently prepare for that conversation by adding analytical depth, which is not what is being asked for. A simple number that reconciles will be accepted where a sophisticated one that does not will be rejected, every time.
The traceability point is the one that converts sceptics. Being able to take any figure on the page, drill to the specific sales behind it, and show the source record for each, is the standard finance applies to its own numbers, and meeting it earns a level of trust that no methodology explanation will.
CloseRev is built to be drilled into: every channel figure opens to the specific sales behind it, and every match shows the source record it was joined to. That is the property that makes a marketing number survive a finance review.
Marketing cost is contested too
The denominator is as disputed as the numerator. Whether agency fees, salaries, tooling, content production and sponsorships count as marketing investment changes every efficiency ratio the business quotes, and most organisations have never settled it.
The pattern is familiar. Media spend is always included. Agency fees usually are. Marketing salaries sometimes are. Tooling, research, events, sponsorship and content production are included by whoever is making the argument and excluded by whoever is not. The resulting ratios are incomparable between periods, teams and companies, and are quoted as though they were not.
The resolution is the same as for revenue: adopt finance's cost centre definition, state what is in it on every report, and hold it constant. If you want a media-only ratio as well, publish both and label them, rather than moving between them depending on which is more flattering.
Be aware that this decision has a political dimension. Including salaries roughly doubles reported marketing investment in many mid-market businesses and therefore halves every efficiency ratio. That is the honest number, and a team that adopts it voluntarily is in a much stronger position than one that has it imposed during a cost review.
The lifetime value complication
Channels differ in the long-term value of the customers they produce, so first-purchase revenue can rank them wrongly. Introducing lifetime value is correct in principle and introduces a forecast into a report that was previously all fact.
This is a real tension and worth handling explicitly rather than resolving by preference. First-purchase revenue is verifiable and understates channels that acquire loyal customers. Lifetime value captures the difference and depends on assumptions about retention that nobody can check today.
The workable compromise is to report observed revenue by cohort at fixed horizons — first purchase, twelve months, twenty-four months — rather than a modelled lifetime figure. Every one of those is a fact rather than a forecast, they become available with time rather than assumption, and they capture most of the durability difference between channels.
Finance will accept observed cohort revenue far more readily than a lifetime value model, because it is the same kind of evidence they use themselves. If a modelled figure is genuinely needed, present it separately and labelled as a forecast, never blended into the measured table.
Two books, and how to avoid keeping them
The failure state is a business where marketing quotes one revenue figure and finance quotes another, both are used, and nobody reconciles them. It is more common than it sounds and it quietly disqualifies marketing from budget conversations.
It arises without anybody deciding on it. Marketing builds reporting from the CRM because that is what they have access to; finance builds from the accounting system because that is authoritative; the two are never compared because they appear in different meetings. By the time somebody puts them side by side, both have a year of history and neither will yield.
The remedy is a single reconciliation, done once, and then a standing line on the marketing report showing the total revenue figure alongside the attributed portion. That line is doing quiet work every month: it demonstrates that marketing is reporting against the same universe as finance, and it makes any future divergence visible immediately rather than a year later.
If you find yourself in the two-books situation already, the fastest route out is to adopt the finance figure wholesale for one quarter, absorb whatever that does to your reported numbers, and explain the restatement once. It is briefly uncomfortable and it ends a problem that otherwise compounds indefinitely.
Getting finance into the room early
Involve finance when the measurement is being designed rather than when it is being presented. A definition agreed in advance is a shared decision; the same definition explained afterwards is marketing's assumption.
The political difference is enormous and the work is identical. A finance director who was consulted on the revenue basis will defend it in a board meeting. One who sees it for the first time on a slide will interrogate it, and will be right to.
It is also the cheapest possible route to the data. Finance generally knows exactly which system holds the authoritative figures, which date fields are reliable, and which known quirks will otherwise consume a week of investigation.
There is a practical sequencing point worth adding. Bring finance in at the stage where you are choosing the export, not after the first report exists. The export decisions — which date field, gross or net, which entities — are exactly the ones finance has authoritative answers to, and getting them right at that moment costs one conversation. Getting them wrong means the first report is built on a basis that has to be redone, and a redone report is remembered as a report that was wrong.
It also changes how the work is described internally. A measurement exercise that finance helped design is a company project; the same exercise presented to them at the end is a marketing project seeking approval. The second framing invites scrutiny of the motive as well as the method, and that is a much harder room.
Every hour spent agreeing definitions with finance before the analysis saves a day of arguing about them afterwards, and the argument is one marketing loses by default.
Questions people actually ask
- Why do marketing and finance revenue numbers never match?
- Because they are measuring different things with the same word. Marketing usually counts bookings at the date a deal was won; finance counts recognised revenue net of refunds and tax, sometimes spread over a contract term. Both are correct within their own convention.
- Which revenue figure should attribution use?
- Whichever one finance uses, because that is the figure the business is managed on. Attribution that reconciles to the management accounts is trusted; attribution that produces its own revenue total will spend every meeting defending the difference.
- Should attribution use bookings or recognised revenue?
- Bookings are closer to the marketing event and therefore more useful for measuring campaigns, but the totals must still reconcile to recognised revenue with the bridge explained. Use bookings for the analysis, and show how they tie to the accounts.
- How do you handle refunds and cancellations in attribution?
- Net them off against the channel that generated the sale, in the period your finance team uses for the reversal. A channel producing revenue that later reverses is not performing as well as its gross figure suggests, and the gross figure will rank channels wrongly.
- What about multi-year contracts?
- Pick a convention and hold it: total contract value, first-year value, or annualised value. Total contract value flatters long deals and makes periods incomparable when contract lengths vary, so first-year or annualised is usually the safer default.
- Who should own the revenue definition?
- Finance owns the definition and marketing adopts it. Any other arrangement produces two sets of books, and the one the chief executive believes will be the one finance produced.