Skip to content
Comparison

One answers what happened. One answers what caused it.

These are not competing methods for the same job. They answer different questions, and most measurement arguments are really two people answering different ones at each other.

Our bias, declared

Use attribution to allocate inside a channel and incrementality to decide whether a channel deserves budget at all. Attribution is cheap, continuous and observational — it tells you which touchpoints preceded conversions, which is genuinely useful for tuning campaigns against each other. It cannot tell you what would have happened without the touchpoint, and no model sophistication changes that, because the counterfactual was never observed. Incrementality is expensive, periodic and causal. Running only the first is the default failure; running only the second is impractical.

Side by side

Attribution vs incrementality testing comparison
FactorAttributionIncrementality testing
Question it answersWhich touchpoints preceded the saleWhat the channel actually caused
Cost to runEffectively free, continuousReal revenue withheld
Speed of feedbackDailyWeeks to a quarter
GranularityCampaign, ad set, keywordChannel, at best
Establishes causationNoYes
Survives brand-search biasNo — over-credits it heavilyYes
Works with degraded trackingIncreasingly poorlyUnaffected — measures totals
Usable for daily biddingYesNo
Defensible to a CFOContestedHard to argue with
Requires separable marketsNoYes — a real constraint

Choose attribution when

You are allocating inside one channel

Given a fixed budget, which campaigns and keywords get it. Attribution does this well and a holdout cannot do it at all — you are not going to run a geo test per ad set.

You need a signal to optimise against daily

Bidding algorithms need conversions fed back continuously. That is attribution's job and there is no substitute, which is why the answer is never to abandon it.

The decision is small and reversible

Most day-to-day media decisions do not justify the cost of causal measurement. Reserve that for the allocations large enough to matter.

Choose incrementality testing when

You are about to move or cut a large budget

Any decision big enough to change the shape of the business deserves a causal answer, because attribution's bias is systematic rather than random and points the same way every time.

The channel looks suspiciously good

Branded search, retargeting and email to engaged lists report returns that would be extraordinary if real. These are exactly where attribution over-credits, and where a holdout most often changes minds.

Tracking has degraded and you know it

Holdouts compare total revenue between regions. They do not care about cookies, consent rates or cross-device paths, which makes them more robust precisely as attribution gets less reliable.

Marketing and finance disagree about what worked

This is the argument attribution cannot settle, because both sides can produce a defensible model. A holdout produces a number neither side chose the methodology for.

What each one actually costs

Attribution's cost is not the tooling, which is usually already paid for. It is the decisions made on it — and specifically the systematic misallocation caused by its bias toward demand capture. That cost is invisible, recurring and frequently much larger than any measurement budget.

Incrementality's cost is explicit and uncomfortable: you withhold advertising from a set of markets and lose the revenue it would have produced. The useful framing is that this cost is proportional to how well the channel works — a test on a channel doing nothing is nearly free, and a test on a channel doing a great deal is expensive precisely because you have found something valuable.

The mistake worth avoiding is treating this as a budget line at all. One holdout a year on your largest channel is small against the allocation it informs, and it makes every attribution number you read for the following twelve months more interpretable — because you now know roughly how far off the model is and in which direction.

Related questions

No, and the reason is structural rather than technical. Every attribution model — last-click, data-driven, position-based — distributes credit among touchpoints that were observed. None of them can observe what would have happened if the touchpoint had not existed, because that world does not exist in the data. Sophistication improves how credit is divided; it cannot manufacture a counterfactual.

Other comparisons

Marketing agency vs in-house teamNot budget, and not company size. It depends on whether you already know which channel works — because that single fact changes the right answer completely.Marketing agency vs freelancerWe say this on sales calls regularly. Under roughly $10,000 a month in media, agency fees eat too much of the budget to be honest value.How to choose a digital marketing agencyMost agency selection processes evaluate the pitch rather than the agency. These are the questions that produce genuinely informative answers.SEO vs PPCSEO compounds and pays back slowly. PPC starts immediately and stops the moment you do. Which one you can afford to wait for is the whole question.Shopify theme vs headless commerceIt buys genuine speed and flexibility, and it buys them with permanent operational complexity most stores don't need.Retainer vs project pricingEach model creates a different incentive, and matching the model to the work matters more than negotiating the rate.Google Ads vs Meta AdsIf people already search for what you sell, start with Google. If they don't know it exists, Google has nothing to capture.Brand vs performance marketingPerformance work is measurable and brand work mostly isn't, so budget flows to what can be defended in a meeting rather than to what works.Performance Max vs Standard ShoppingThe performance difference between them is smaller than the reporting difference. What you are really choosing is how much you are willing to not be able to see.Last-click vs data-driven attributionAttribution does not measure causation and was never able to. The question is not which model is accurate — it is which set of errors you can work with.Email vs SMS marketingThat single difference drives every other decision — what you send, how often, and how quickly you can burn a list you spent a year building.Full rebrand vs brand refreshThe symptom is usually that the brand looks dated. The diagnosis is almost never that the brand is wrong — and those two problems have very different price tags.Webflow vs a custom buildMost companies pick the platform once, at the point of least information, and live with it for four years. The better question is which one fits the next eighteen months.Organic social vs paid socialThe comparison is usually framed as free versus paid, which is why it produces bad decisions. Both cost money. They just buy completely different things.Google Ads vs Amazon AdsThe difference is not intent or cost per click. It is whether you end up with a customer you can contact again, or a transaction Amazon owns the relationship for.Creator content vs studio productionThe comparison is usually framed as cheap versus expensive, which is why it goes wrong. Compare cost per asset that actually gets used and the gap narrows sharply.In-house studio vs production partnerWhether to build a studio is decided by how many assets you need every month and how predictable that number is — not by how much you care about the work.Native vs cross-platform appsThe technology argument is largely settled and mostly irrelevant. What decides this is how many engineers you have and how close your product sits to the hardware.

Services referenced

Still deciding

Get the audit before you decide anything.

Thirty minutes, free, no obligation. We'll tell you honestly whether you need an agency at all — including when the answer is that you don't.