Ask buyers who own an attribution tool what frustrates them and the complaints split along a fault line: the tools that assign credit — multi-touch attribution and the dashboards built on it — draw the trust complaints, while the tools that measure lift earn more belief but cost more and fit lumpy budgets worse. The numbers don't reconcile, the tool over-credits what it can see, the model is opaque, the data underneath is messy, the annual bill doesn't match a seasonal spend, and no single tool is the source of truth. What differs is which tool earns which complaint most — and whether a given complaint is livable, fixable, or the kind that ends with the numbers pulled from the business review. Below is the catalog, mapped and graded. For the market shift behind it, see the flagship read on why nobody believes their attribution anymore; for the pre-purchase view, what attribution-tool buyers wish they'd known; and for buying mix modeling, what MMM buyers wish they'd known.
The six complaints — graded
Each complaint below carries a verdict: is it something you can live with, something money and discipline can fix, or the kind of problem that ends with the numbers pulled from the business review. The grade is the useful part — the credit-assigning tools draw most of the trust complaints, the lift-based tools draw the cost ones, and the question is never does it have complaints but which of them are dealbreakers for me.
The numbers don't reconcile — you can't trust them anymore
The complaint at the center of the whole category. Privacy changes — iOS signal loss, cookie deprecation — broke the tracking that click-path multi-touch attribution depends on, and buyers describe the fallout bluntly: the numbers no longer match reality. One apparel brand deemed its multi-touch figures unreliable to the point of pulling them out of its weekly business review entirely; other buyers doubt the results outright. This is the complaint most likely to be fatal, because when a measurement can't be trusted, teams don't argue with it — they stop looking at it, and start an evaluation. It falls hardest on the credit-assigning tools; the lift-based methods buyers are moving toward earn more belief precisely because they sidestep the tracking that broke.
Named for this: Rockerbox · Triple Whale · Northbeam · multi-touch attribution broadly
It over-credits the paid channels it can see
The structural bias buyers learn to correct for. Click-path attribution can only credit what it observes, and after signal loss it observes paid, trackable media best — so it systematically over-credits paid and under-credits email, SMS, organic, and offline, where much of the real influence lives. Buyers describe dashboards that skew heavily toward paid and aren't built for the email or SMS marketer, quietly misallocating budget toward whatever reports well. It's fixable, but not by the tool: the correction is to triangulate — validate the dashboard's channel split against an incrementality or holdout test that measures causal lift regardless of trackability. Treat a single attribution tool's channel story as one biased input, not the answer, and the complaint becomes manageable.
Named for this: Rockerbox · Triple Whale · Northbeam
It's a black box you can't explain — so you can't defend it
The complaint that undoes rigor. The same tools praised for statistical seriousness are, by some buyers, called a "black box" — the calculations opaque, leaving users skeptical of results in a space where trust is the entire product. In measurement, methodology is the value: if you can't see how the model reaches its answer, what it assumes, and how the holdout was built, you can't stand behind the output when the CFO pushes back. Recast earns its reputation partly on transparency for exactly this reason, and is the counter-example buyers name when they want to see inside the model. Opacity is often fatal because it fails at the decisive moment — the budget defense — no matter how good the underlying math is. The complaint is retrospective in every case — buyers describe the opacity of a model already running under contract, not one they audited first. So the next part is our recommendation rather than theirs: make any vendor walk you through the model end to end, and treat "it's proprietary" as a yellow flag, not a reassurance.
Named for this: Haus · incrementality & MMM models broadly · Recast (the transparency counter-example)
It's only as good as your data — messy inputs sink it first
The complaint buyers blame on the tool least fairly. Attribution and incrementality models amplify whatever you feed them: clean tracking produces a trustworthy read, and broken tagging produces a confident, wrong one. Onboarding pain in the interviews concentrates here — messy internal data and slow integration — and it hits every tool in the category, from the dashboards to the rigorous lift-based platforms. It's fixable, but the fix is yours, not the vendor's: the buyers who got value fixed their tracking, tagging, and data pipes first, then bought the model. The tool is the last mile of measurement, not the cleanup crew — and a tool blamed for bad numbers is often just reporting bad inputs faithfully.
Named for this: Haus · Rockerbox · Triple Whale · tools broadly
The cost doesn't fit lumpy or seasonal spend
The complaint specific to the rigorous tools. The lift-based platforms are praised for exactly the trust the dashboards lack — but several buyers on seasonal or sporadic media budgets describe the annual contracts as a poor fit, paying through off-peak periods when they aren't running enough tests to justify the cost. Incrementality and geo-lift methods also need a minimum viable media budget to produce a readable signal; below that threshold the tests can't conclude and you're paying for measurement you can't act on. It's fixable — price the tool against the conclusive tests you'll actually run in a year, and negotiate the cadence if your spend is lumpy — but left unaddressed it's the gripe that makes buyers resent an otherwise-trusted tool every off-season.
Named for this: Haus · Recast · incrementality & MMM tools
No single tool is the source of truth — you'll run several
The complaint that's really the mature posture in disguise. Buyers describe fractured measurement across multiple tools without a single source of truth and wish for one number — but the buyers who succeed stopped hunting for it. They run multi-touch for in-platform optimization, incrementality to validate what actually caused sales, and mix modeling for the cross-channel and offline picture, and let the lift-based method referee when the tools disagree. It's livable — in fact it's the correct operating model — once you reframe it: you're not failing to find the one true tool, you're building a measurement stack where each tool answers the question it's good at. Budget for the triangulation, not the silver bullet. No buyer here describes asking a vendor how its numbers reconcile with the rest of the stack before signing — the divergence is found once the tools are running side by side — so the next part is our recommendation rather than theirs: ask every vendor how their numbers reconcile with the others when, not if, they diverge.
Named for this: Haus · Rockerbox · Triple Whale · Northbeam · the category broadly
Want this read against your own stack?
Get my read →The stories behind the complaints
The counter-current: some complaints are the method, not the tool
Listen closely and several of the loudest complaints aren't defects a better vendor fixes — they're properties of the method. Over-credits paid channels is what click-path attribution does after signal loss; only as good as your data is true of every model ever built; and no single source of truth is the honest reality of measuring marketing, not a failure to shop hard enough. The interviews bear this out: the buyers most at peace with their measurement stopped expecting one tool to be trustworthy, transparent, cheap, and complete all at once. They cleaned their data, triangulated the credit-assigning tools against lift-based ones, demanded transparency where they could, and matched the cost to their cadence. Before you conclude the tool is the problem, separate the complaints a vendor owns — a genuinely opaque model, a cost structure that won't flex — from the ones the method owns. Only the first kind is a reason to switch; the second follows you to the next tool.
What this means — whether you run one or you're buying one
If you already run an attribution or measurement tool, sort your own complaints with the grades above: the livable one (no single source of truth) is a posture to adopt, not a problem to solve; the fixable ones (paid-channel bias, dirty data, a cost that won't flex) are discipline-and-negotiation questions; and the fatal ones (numbers that no longer reconcile, a black box you can't defend to finance) are the signal to plan a replacement rather than renew on faith. If you're evaluating one, decide first what question you're answering — credit, causal lift, or contribution — and don't buy one expecting the others; then accept the trade the category forces: the credit-assigning tools are cheaper and faster but draw the trust complaints, while the lift-based tools earn belief but cost more and fit lumpy spend worse. There's no measurement tool without complaints. There's only the set you can live with — and the discipline, around whichever tool you pick, that decides whether finance believes the number.
Common questions
Why don't marketers trust their attribution numbers?
Three complaints recur across buyer interviews. First, the numbers don't reconcile: privacy changes — iOS signal loss, cookie deprecation — broke the tracking that click-path attribution depends on, so the figures no longer match reality, and one brand pulled its multi-touch numbers out of its weekly business review as too unreliable to report. Second, opacity: buyers describe their measurement tool as a 'black box' whose calculations they can't see or explain, which makes the output impossible to defend when leadership pushes back. Third, dirty inputs: attribution models amplify whatever data you feed them, so messy tracking produces confident, wrong answers. The pattern is that the credit-assigning tools — multi-touch attribution and the dashboards built on it — draw the trust complaints most, while the lift-based methods buyers are moving toward earn more belief because they measure causal impact rather than assign credit to touchpoints.
Which attribution tool has the worst reputation with buyers?
There isn't a single worst — the complaints sort by what the tool does, not by brand. The multi-touch tools and DTC dashboards — Rockerbox, Triple Whale, Northbeam — draw the trust complaints most: numbers that don't reconcile after signal loss, a bias toward the paid channels they can see, and blind spots on email, SMS, and offline. The incrementality and mix-modeling tools — Haus, Recast — earn more trust on the numbers but draw different complaints: cost that fits lumpy or seasonal spend poorly, onboarding on messy data, and, for one buyer, 'black box' opacity in the model. The useful question isn't which tool is worst but which complaint can I live with: a tool whose numbers you doubt, or a rigorous one you'll pay for through your off-season and have to take partly on faith.
Are these complaints reasons to switch, or can you live with them?
It depends which complaint. Some are livable or fixable: paid-channel bias is real but you correct for it by triangulating with a lift-based method rather than trusting one dashboard; poor data quality is on you to fix before you blame the tool; and a cost that fits lumpy spend badly can be negotiated on cadence. And accepting that no single tool is the source of truth — that you'll run several and let incrementality referee — is the mature posture, not a failure. But two complaints are frequently fatal: numbers that have stopped reconciling with reality, and a black-box model whose output you can't defend to finance. When the measurement can't be trusted or can't be explained, teams don't argue with it — they stop using it, and start an evaluation. Sort your own complaints into livable, fixable, and fatal before you decide whether to invest or replace.
Why do attribution tools over-credit some channels?
Because they can only credit what they can see, and after privacy signal loss they see paid, trackable channels best. Click-path multi-touch attribution assigns credit along the touchpoints it can observe, which biases it toward the platforms that still report well — typically paid media — and leaves it weak on email, SMS, organic, and offline, where much of the real influence happens. Buyers describe dashboards that skew heavily toward paid and aren't built for the email or SMS marketer, so the tool systematically over-credits the channels it tracks and under-credits the ones it can't. It's fixable, but not by the tool: the correction is to triangulate — validate the dashboard's story against an incrementality or holdout test that measures causal lift regardless of trackability. Treat a single attribution tool's channel split as one biased input, not the answer.
Get this research made for your stack
Running an attribution or measurement tool and wondering if your complaints are normal? Do a 15-minute interview about your own spend and stack and get this personalized — which of these gripes peers on your tool share, which are livable, and which are the ones that end with the numbers pulled from the business review.
Get my personalized read — 15-min interviewNo password · your interview is anonymized before it ever informs a page like this one.