Add up what every team in a large company claims to have driven, and the total will exceed the company's actual revenue — sometimes by a lot. Nobody is lying. Each team's dashboard is configured to claim the same conversions, and the meeting where those claims collide decides budgets based on who argues best, not on what actually worked.
TL;DR
- The same conversion gets counted by multiple teams at once. Marketing's dashboard, product's dashboard, and sales' dashboard each claim credit for the identical purchase — so the sum of attributed contributions can exceed 100% of real revenue.
- This is a political problem wearing an analytics costume. Each team's measurement is configured, in good faith, to maximize its own visible contribution. The tools aren't wrong; they're just all pointed at the same conversions.
- Budget follows the loudest, best-instrumented argument — not the channel with the highest true incremental contribution. That misallocates spend toward whoever built the most flattering dashboard.
- The fix is organizational, not technical: one shared definition of incremental contribution that every team reports against, so credit reconciles to reality instead of summing past it.
| What each team's dashboard says | What the company actually has |
|---|---|
| Marketing: "we drove this conversion" | One conversion |
| Product: "our onboarding drove it" | One conversion |
| Sales: "our follow-up closed it" | One conversion |
| Sum of claims | ~300% of a single sale |
When three teams each book the same sale, the organization's attributed revenue is fiction — and it's the fiction that budgets get set against.
The measurement is fine; the incentives are the problem
This looks like an attribution problem, and there's a real attribution problem underneath it — attribution models are genuinely broken in ways worth understanding on their own. But the thing that makes this failure so durable isn't the measurement math. It's that every team has a rational incentive to configure its measurement to claim as much credit as the tooling allows.
Marketing sets a generous attribution window, so any conversion within 30 days of an ad impression counts as marketing's. Product instruments the onboarding flow, so any conversion by an activated user counts as product's. Sales logs a touch, so any closed account counts as sales'. None of these is fraudulent. Each is a defensible measurement choice made by a team that will be evaluated, funded, and promoted on the number it produces. The predictable result: the same conversion is claimed three times, and no single dashboard is "wrong" in isolation. They're only incoherent in aggregate — and no one owns the aggregate.
The dashboards don't disagree because someone made an error. They disagree because each was built to win a budget argument, and they're all pointed at the same conversions.
Why the org can't see the double-count
The double-counting stays invisible because of how large companies are structured. Each team reports its own number up its own chain. Marketing's VP sees marketing's contribution. Product's VP sees product's. Nobody's default view is the reconciliation — the pooled question of whether the sum of claimed contributions bears any relationship to total revenue. That reconciliation isn't any single team's job, and the teams that would look worse under it have no incentive to build it.
So the numbers roll up in parallel, each internally consistent, and collide only in the rare cross-functional meeting where two leaders present the same conversion as their own win. In that room, the disagreement gets read as a data-quality issue ("our numbers don't tie out") rather than what it is: a structural consequence of every team optimizing its own visible credit. And because it's framed as a data glitch rather than an incentive problem, the "fix" is usually another dashboard — which just adds a fourth claimant.
This is the political layer sitting on top of the measurement failure. The measurement question — did this channel actually cause incremental revenue — is the domain of channel cannibalization and incrementality testing. This article is about the layer above it: even when you could measure incrementality, the organization's incentives push every team to claim the credit anyway, and budgets get set on the claims rather than the incrementality.
The cost: budget follows argument, not contribution
The consequence isn't just messy reporting. It's capital misallocation. When credit is contested and no shared truth exists, budget flows to the team with the most persuasive dashboard, not the channel with the highest true incremental return. A team that instruments aggressively and presents confidently captures budget from a team whose real contribution is larger but worse-marketed internally.
The canonical evidence for how large this gap can be comes from the eBay paid-search experiments. When researchers actually ran the counterfactual, they found that for branded search terms, roughly 99.5% of the traffic the ads claimed credit for would have arrived through organic search anyway — the paid channel's reported contribution was almost entirely non-incremental (Blake, Nosko & Tadelis, *Consumer Heterogeneity and Paid Search Effectiveness*, Econometrica 2015). A field experiment run through Yahoo! reached a structurally similar conclusion: a majority of an ad campaign's true incremental effect came from people who never clicked, meaning click-based credit was pointed at the wrong users entirely (Johnson, Lewis & Reiley, *Online ads and offline sales*). If a channel's self-reported contribution can be that detached from its causal contribution, then setting budget on self-reported credit is setting it on a number with almost no relationship to what actually drives revenue.
Now layer the politics on top: it's not just that each channel over-claims against reality. It's that channels over-claim against each other, and the arbiter is a stakeholder meeting rather than an experiment. That's how a company ends up over-funding its best internal marketer and under-funding its best actual driver of growth.
The fix: one shared definition of incremental credit
The failure is organizational, so the fix has to be organizational. A better dashboard for one team just escalates the arms race. What actually resolves it is a single, shared definition of incremental contribution that every team reports against — enforced above the level of any individual team, because no team will disarm unilaterally.
Concretely:
- Establish one incrementality standard, owned centrally. Credit is assigned based on measured incremental lift — the counterfactual question of what would have happened without the channel — not on last-touch or self-configured attribution windows. This is where holdout tests earn their keep: they produce a credit number that isn't a team's to configure.
- Make the reconciliation someone's explicit job. The pooled view — do claimed contributions sum to something like actual revenue? — must have an owner with the authority to say "these three claims are the same conversion." Without an owner, the aggregate stays invisible by default.
- Report contribution in incremental terms, not vanity terms. A team's claimed conversions are a vanity metric when they're not incremental; the number that should drive budget is incremental margin. Shifting the reported unit from "conversions we touched" to "incremental revenue we caused" removes most of the double-count at the source.
- Change what gets rewarded. As long as teams are evaluated on self-claimed contribution, they'll optimize self-claimed contribution. Tie evaluation to incremental contribution under the shared standard, and the incentive to over-claim evaporates.
I've been the person in the room when two teams presented the same win as their own. The path of least resistance is to let each keep its number and move on — everyone stays comfortable. The move that mattered was insisting on a single incremental standard the whole org reported against, so credit reconciled to reality instead of summing past it. It's not a popular position in the moment; the teams whose reported contribution shrinks under an honest reconciliation don't enjoy it. But it's the only version where budget follows what actually drives growth.
FAQ
Isn't some double-counting unavoidable in a multi-touch journey?
Multiple touches genuinely contributing to one conversion is real — the journey isn't single-cause. The problem isn't that several teams touched the conversion; it's that each claims the full value of it, so the sum exceeds reality. The fix isn't to pick one winner; it's to move to an incremental standard where each team's credit reflects its marginal contribution, so the parts sum to the whole instead of to three times the whole.
How is this different from channel cannibalization?
Channel cannibalization is the measurement problem — a channel taking credit for conversions that would have happened anyway. This is the political problem layered on top: even when incrementality is measurable, each team's incentives push it to claim the credit, and budgets get set on the claims. Cannibalization is about whether the number is causally real; this is about which team gets to book it and why the organization can't see that they're all booking the same one.
Who should own the reconciliation?
Someone above the competing teams — typically a central analytics or finance-partnered function — with the authority to define the incremental standard and to declare when two claims are the same conversion. It cannot be one of the claiming teams, because no team will adopt a standard that shrinks its own reported contribution. The ownership has to sit where the incentive to over-claim doesn't.
Won't teams resist an incremental standard that lowers their numbers?
Yes, and that resistance is the tell that you've found the real problem. The teams whose reported contribution was inflated by double-counting will see their numbers fall, and they'll fight it. That's exactly why the standard has to be imposed from above and tied to how teams are evaluated. The resistance isn't a sign the standard is wrong; it's a sign it's working.
Bottom line
In a large organization, the same conversion gets claimed by every team that touched it, so the sum of attributed contributions can far exceed actual revenue — and budgets get set against that inflated fiction. This isn't a data-quality bug; it's the predictable result of every team, in good faith, configuring its measurement to maximize its own visible credit. Left unmanaged, capital flows to the best internal marketer rather than the best actual driver of growth. The only durable fix is organizational: one shared definition of incremental contribution, owned above the competing teams, that every team reports against — so credit reconciles to reality instead of summing past it. Until that exists, your budget is being allocated by whoever built the most flattering dashboard.
Turning incrementality into a standard the whole org can share is exactly the gap I built GrowthLayer to close. For more on the politics of measurement inside real companies, subscribe to Lean Experiments.