Attribution is the most expensive faith in B2B go-to-market. We built elaborate dashboards that promise a causal answer, funded them like scripture, and never audited whether the god actually exists.
Here is the heresy. Multi-touch attribution does not measure what you think it measures. It takes correlational data, the touches that happened to be trackable, and dresses them up as a causal story about what drove revenue. Then a CFO signs the renewal because a confident number feels safer than an honest judgment call.
A recent Measured piece puts it bluntly. Multi-touch attribution is not just outdated. It was built on an assumption that was always shaky and is now almost completely false, and the brands still treating MTA as their primary measurement framework are making budget decisions on data that ranges from incomplete to actively misleading.
That is not a tooling problem. That is a theology problem.
The data the religion depends on is gone
The original pitch was seductive. Multi-touch attribution promised to show you exactly which ads drove revenue. It never quite delivered, and now the data it relies on is disappearing.
Look at the mechanics and the faith collapses. Industry surveys suggest that 30 to 40 percent of B2B buyer touchpoints occur in untracked channels: analyst calls, peer referrals, review sites without UTM parameters, LinkedIn DMs, and Slack community discussions. Your model does not measure reality. It measures the trackable slice of reality and then confidently credits it, which means MTA models allocate zero credit to these invisible influences, systematically overfunding trackable channels.
It gets worse under the hood. Attribution window configuration, 7-day versus 90-day, can shift channel credit allocation by 20-plus percentage points, so sensitivity testing is mandatory. Translation: you can change which channel looks like your hero by changing one setting. That is not a measurement system. That is a Ouija board with a spreadsheet attached.
And the platforms are all lying to you individually. Google Ads, Meta, and LinkedIn all use different attribution windows and methodologies, and they will each claim credit for the same conversion. You are being pitched salvation by five vendors who each take full credit for your best quarter.
The CODN question nobody is asking
Here is where the Cost of Doing Nothing framework matters, because it reframes the entire budget conversation.
Attribution asks: which touch deserves the credit for this deal? Wrong question. The only question that pays rent is: if I did nothing, if I killed this channel for a quarter, would pipeline actually move? That is incrementality, and it is the exact thing attribution structurally cannot answer.
The smart money already figured this out. Incrementality runs real experiments, holding a channel out from a control group, and answers the one question attribution structurally cannot: would this conversion have happened anyway. And the mistake is not subtle. The biggest error is using attribution for budget allocation, because attributed return ignores baseline conversions and channel saturation entirely.
The market is voting with its wallet. In a July 2025 survey from EMARKETER and TransUnion, 27.6 percent of US marketers named marketing mix modeling their single most reliable measurement methodology, the top answer, and 46.9 percent planned to invest more in it.
What to measure when you accept attribution is directional
Stop trying to make attribution a budget oracle. Demote it to what it actually is: a fast, in-channel optimization signal. Attribution is the tactical layer, the fast operational signal for optimizing within a channel, while the real value is calibration: incrementality results become priors that correct the MMM, the MMM tells you which channels are worth testing, and attribution optimizes execution inside the boundaries the other two set.
That is the stack I ship with GTMify clients. Three layers, three jobs.
Run holdout tests to prove lift. Pick a channel, go dark in a matched segment, and measure what pipeline does without it. If nothing moves, you just found your CODN, and it is zero. Reallocate.
Use marketing mix modeling for the macro view, especially if you have a long cycle. For B2B SaaS companies with 6 to 18 month sales cycles, MMM provides the macro channel view that user-level attribution can no longer deliver consistently.
Then track pipeline velocity and influenced pipeline by segment as your learning loop, not your verdict. The point is not perfect numbers. It is a decision that changes.
Every measurement system should either change a budget decision or get killed. If your attribution model only decorates a QBR slide, you are not running a growth engine. You are tithing. Burn the idol, fund the experiments, and start measuring whether you could do nothing and survive.
Originally published at scottwueschinski.com.