.svg)
Where Marketing Measurement Breaks: The 4 Gaps to Check First
When measurement isn't working, most teams reach for a new attribution model. It's almost never the fix. The problem is rarely the model you chose — it's one of four specific places measurement breaks, usually before any model even enters the picture. Marketing-measurement expert Kelly Zeitlow, who has built and rebuilt measurement systems at Booking.com, Gorillas, and HeyJobs (now VP of Marketing at Spread Group), makes exactly this point: when she walks into a new company, she starts with an audit, not a model. Here are the four gaps that audit looks for — and how to tell which one is yours.
📺 This piece draws on our fireside chat with Kelly Zeitlow. Prefer to watch? See the full conversation here →
The short version: Marketing measurement usually breaks in one of four places, not because you picked the wrong attribution model. Tracking: events fire inconsistently and no one can confirm the dashboard matches reality. Reconciliation: marketing, finance, and the ad platforms each report a different "revenue" that no one reconciles. Attribution: an inherited first-touch or last-click model quietly misreads what's driving growth. Method fit: the setup can't support where measurement has moved — multi-touch, media mix modeling, and incrementality. Figuring out which gap is yours is the real first step — which is why Kelly Zeitlow starts every engagement with an audit, not a model.
It's a gap problem, not a model problem
The instinct to swap attribution models treats a symptom. What actually determines whether you can trust your numbers is the state of the foundation underneath them — and Kelly's audit checks that on two tracks. First, she looks at the data itself: are the KPIs accurate, available when they're supposed to be, and do the dashboards marketers use daily line up with the ones executives use for strategy? Second, she interviews the people using the data: what KPI do they actually act on, and do they trust it?
The tell she's stopped being surprised by: two people at the same company give completely different answers. That gap — between what people believe about their measurement and what's actually happening — is itself the diagnosis. And it almost always lands in one of the four buckets below.
Gap 1: Tracking
The foundation gap. Events fire inconsistently, tags break silently when the site changes, and no one can confirm that the numbers on the dashboard match what actually happened. When Kelly audits data strength, this is the first thing she checks — are the KPIs accurate and available, or broken and delayed? If your tracking can't be validated, nothing built on top of it can be trusted either, no matter how sophisticated your model is. Fixing it means disciplined event tracking you can actually verify.
Gap 2: Reconciliation
The trust gap. Marketing quotes one revenue number, finance quotes another, and the ad platforms each claim their own — and no one owns reconciling them. Kelly's "million-dollar question" for any company is simply which sources do you trust as the truth? — and the answer varies wildly, company to company and person to person. When there's no single agreed definition of a metric, every cross-team conversation becomes a translation problem, and every meeting risks becoming an argument about whose number is right. The fix is one single source of truth the whole company reconciles to.
Gap 3: Attribution
The bias gap. This is the one teams think is the whole problem — and it's real, just not in the way they assume. Kelly's conclusion after working with every model (first touch, last touch, multi-touch, the platforms' own attribution, lift studies, MMM, incrementality): every model optimizes for something and quietly discounts everything else. Click-based models are trackable, which is their appeal — but they systematically favor click-driving, bottom-of-funnel channels like paid search, while undercounting view-based channels (organic and paid social, influencer), missing retention and CRM, and leaving out offline entirely.
On top of that, privacy is shrinking what any click-based model can even see — Apple blocking cookies by default, GDPR tightening across the EU. So an inherited first- or last-click model doesn't just have a bias; it's working from an increasingly incomplete picture. The fix isn't a "perfect" model — it's attribution built to reflect reality so budget follows what actually works.
Gap 4: Method fit
The gap that's newest and most strategic: your setup can't support where measurement has moved. For years, Kelly notes, the best digital teams believed they'd reach a single source of truth — granular, individual-level tracking of everything. That goal is gone, and Google itself has shifted from chasing one perfect source to triangulating multiple imperfect ones. That means three methods working together: multi-touch attribution for the trackable path, media mix modeling to capture the channels click-based models miss (and to simulate budget shifts before you make them), and incrementality testing — geo-splits that turn a channel off to reveal what you'd actually lose without it. Kelly calls incrementality the most underused and most powerful tool available, best added once you're live across four or five channels. If your measurement is one fragile number instead of a triangulated picture, that's a method-fit gap — and it's the difference between a number you can defend to finance and one you can't. (Our guide to incrementality covers the causal method in depth.)
How to tell which gap is yours, and why it matters
You diagnose it the way Kelly does: look at the data and talk to the people, and see where belief and reality diverge. The reason it's worth the effort is that measurement gaps don't stay in the dashboard — they shape strategy. Kelly's example: leadership says "we're struggling because CPCs are going up," but that's a symptom, not a cause — dig down and it's usually an under-diversified mix or too little brand investment, sitting on top of a measurement model that can't see them. And when budget gets tight and someone asks where to cut, teams with real measurement can say exactly what a 10% cut to a channel will cost. Without it, as Kelly puts it, "it's only opinions" — and the loudest voice in the room wins.
Find your gap: watch the chat, then run the audit
Kelly walks through all of this — the audit, the model biases, MMM and incrementality, and the "CPCs are a symptom" insight — in the full fireside chat.
▶ Watch "Marketing Measurement: Yesterday, Today, and Tomorrow" →
And when you're ready to find your gap, that's exactly what our Measurement Audit does — a structured read across your tracking, data layer, and attribution, with a clear picture of what you can trust and a prioritized roadmap to fix what you can't.
Speak with the Adasight team →
Frequently asked questions
What are the four places marketing measurement breaks?
Tracking (events fire inconsistently and can't be validated against reality), reconciliation (marketing, finance, and platforms each report a different revenue that no one reconciles), attribution (an inherited first- or last-click model misreads what drives growth), and method fit (the setup can't support multi-touch, media mix modeling, and incrementality). Most measurement problems trace to one of these rather than to the choice of model.
How do you know if your marketing measurement can be trusted?
Audit it on two tracks: examine the data itself (are KPIs accurate, available, and consistent across daily and executive dashboards?) and interview the people using it (which KPI do they act on, and do they trust it?). Where people's beliefs and the actual data diverge, you've found the gap.
Why isn't a new attribution model the fix?
Because the model is usually a symptom, not the root cause. If your tracking can't be validated, your teams can't agree on a definition of revenue, or your setup can't support triangulation, no attribution model will produce trustworthy numbers. Every model also carries built-in bias, so switching models just trades one blind spot for another.
What does "method fit" mean in measurement?
It's whether your measurement approach matches where the discipline has moved — from chasing one "single source of truth" to triangulating multiple imperfect methods: multi-touch attribution, media mix modeling, and incrementality testing. A method-fit gap means your setup produces one fragile number instead of a triangulated picture you can defend.
Where should you start fixing marketing measurement?
Start with an audit rather than a model, so you know which of the four gaps you actually have. Fix the foundation first (tracking and reconciliation), get a multi-touch model that reflects your real path to purchase, then add media mix modeling and incrementality as your channel count grows. And treat it as continuous — measurement goes stale as channels, tactics, and privacy rules change.
