Latest / The CMO Podcast with Fexingo: Marketing Leadership, Budgets, and Executive Strategy / How CMOs Fix Broken Marketing Attribution
Transcript
- Lucas: The single most dangerous metric in a CMO’s dashboard right now is last-click attribution. It feels safe because it’s precise, but it’s lying to you about half your revenue. Luna: Right, because it gives all the credit to the final touchpoint while ignoring the awareness campaigns that actually got people into the funnel in the first place. Lucas: Exactly. And by September 2026, with ad costs continuing to climb across social platforms, relying on that simplistic model isn’t just inaccurate—it’s expensive. We’re seeing major brands like Nike quietly pivot away from pure last-click models toward something called media-mix modeling, or MMM, which looks at the bigger picture over longer time horizons. Luna: So instead of asking who closed the sale, they’re asking which combination of channels drove the lift in overall demand? Lucas: Precisely. The problem is that most CMOs haven’t built the internal infrastructure to support that shift. They’re still stuck in what I call data silos—where the sales team uses one CRM, the marketing automation platform talks to a different database, and nobody has cleaned up the duplicate records for years. You can’t build a robust MMM if your underlying data is full of ghosts. Luna: I see this constantly. The sales reps enter deals differently than the marketers do, so when you try to merge those datasets, the correlation just falls apart. Lucas: It does. And here is the concrete thing most leaders miss: fixing attribution isn’t primarily a technology problem; it’s a governance problem. You need a unified definition of what constitutes a lead, a qualified opportunity, and a closed-won deal. If your sales team changes their qualification criteria every quarter, no amount of software will give you clean attribution. Luna: But doesn’t enforcing strict definitions kill the agility that high-growth teams rely on? Sometimes you need to move fast before the process is perfect. Lucas: That is the classic tension. But look at how Apple structures its product launches—they don’t let individual regional teams change the core messaging framework mid-campaign. The consistency allows them to measure impact accurately. If you want to know what works, you have to keep the variables stable long enough to get a signal. Luna: So stability in the framework enables speed in the execution? That makes sense, but it requires discipline that many startup CMOs just don’t have yet. Lucas: They don’t. Which brings us to the second pillar of this shift: incrementality testing. This is where you stop guessing and start proving. Instead of assuming that everyone who saw your ad would have bought anyway, you run controlled experiments where you withhold ads from a specific segment and compare their behavior against a control group. Luna: That sounds like the gold standard, but it’s also incredibly resource-intensive to set up properly for every single campaign. Lucas: It is, which is why you don’t do it for everything. You do it for your biggest budget drivers. If you spend fifty million dollars a year on Facebook, you absolutely need to know how much of that is incremental. If the answer is only two million, you’ve just saved forty-eight million by cutting the waste. Luna: Wow. That kind of clarity completely changes the conversation with the CFO. Suddenly marketing isn’t an expense center; it’s a verified investment engine. Lucas: It does. And this is why the best CMOs are spending more time in the data warehouse and less time in creative review meetings. They are obsessed with the plumbing. If the pipes are clogged with bad data, the water won’t flow no matter how pretty the faucet is. Luna: I love that analogy. So the real work is happening behind the walls, not on the stage. It’s unglamorous, but it’s where the actual value lives. Lucas: Exactly. And let’s talk about the privacy angle briefly, because it ties directly into attribution. With third-party cookies effectively gone and iOS updates restricting tracking, we can’t rely on user-level tracking anymore. We have to move back to aggregated, probabilistic models. That means accepting some level of uncertainty in exchange for scale. Luna: So we’re trading precision for reach? That feels like a hard pill to swallow for performance marketers who are used to seeing exactly which user clicked which ad. Lucas: It is a hard pill, but it’s the reality of the market in 2026. The platforms themselves are optimizing for conversions using modeled audiences, not just tracked users. If you fight that by trying to force precise attribution through outdated methods, you’re fighting the algorithm. You have to trust the aggregate signal. Luna: Which means trusting the system enough to scale, even if you can’t see the individual dots connecting to the line. Lucas: Precisely. Now, before we go deeper into the technical side of implementing these tests, I want to pause for a quick moment. If these conversations about marketing strategy and data hygiene have sparked something you’ve actually used in your own workflow, consider supporting the show. Honestly, if today was worth a coffee to you, that's the link — buy me a coffee dot com slash fexingo. It keeps the research sharp and the ads off your feed. Luna: Yeah, and it helps us dig into cases like this one where the details really matter. Thanks for keeping the lights on. Lucas: Let’s get back to the implementation. One of the biggest mistakes I see is CMOs trying to overhaul their entire attribution model overnight. They wait for the perfect solution, and in the meantime, they continue making decisions based on incomplete data. Luna: Perfectionism is the enemy of progress here. You need a good enough model to start making better decisions immediately. Lucas: Agreed. Start with a hybrid approach. Keep last-click for low-funnel paid search because it’s cheap and direct, but layer in a simple multi-touch model for brand awareness campaigns. Give each channel a percentage of the credit based on where it sits in the journey. Is it top, middle, or bottom? Luna: So you’re essentially creating a weighted scoring system that acknowledges the contribution of each step rather than erasing the earlier ones? Lucas: Yes. And as you gather more data, you can refine those weights. Maybe video views drive twenty percent of the conversion probability. Maybe email nurture sequences drive thirty. You adjust the numbers quarterly based on what the experiment tells you. Luna: That feels manageable. It’s iterative rather than revolutionary. It allows the team to learn without burning the bridge to their current processes. Lucas: It does. And here is the counter-intuitive part: the brands that succeed with this approach are often the ones willing to cut budgets for channels that look good on paper but fail the incrementality test. It takes courage to turn off money that appears to be working, especially when the vendor is telling you it’s fine. Luna: That’s the hardest sell to the board. Showing them a chart where a line goes down because you stopped wasting money on vanity metrics. They usually want to see growth, even if it’s fake growth. Lucas: They do. But if you frame it correctly—as reallocating capital to higher-return activities—you win the argument. You say, 'We are taking five million from underperforming display and putting it into high-incrementality influencer partnerships.' Now you have a net positive story. Luna: That reframing is everything. It shifts the narrative from cost-cutting to smart allocation. It speaks the language of the finance team while protecting the integrity of the marketing strategy. Lucas: It does. And I think the other critical piece is internal alignment. The sales team needs to buy into the new attribution model. If they feel like marketing is stealing credit for their closings, they will sabotage the data entry. You have to make sure the incentives are aligned so that both sides benefit from accurate reporting. Luna: So the CMO has to be part diplomat as well as data scientist. Building trust across departments is just as important as building the model itself. Lucas: Absolutely. You cannot do this in a vacuum. The best attribution strategies emerge from cross-functional workshops where sales, marketing, and finance sit in a room and argue about the definitions until they agree on a single source of truth. Luna: Arguing is good. As long as they are arguing about facts and not egos. Once that consensus is reached, the whole organization moves faster because everyone is looking at the same dashboard. Lucas: Exactly. Speed comes from clarity. When everyone knows what success looks like and how it is measured, you stop having those weekly meetings where you debate whether a campaign worked. You just look at the numbers and act. Luna: That’s the dream. A meeting that ends in ten minutes because the data was already reviewed and understood before anyone walked in the door. Lucas: It is possible. It just requires the discipline to do the heavy lifting upfront. Most leaders skip the heavy lifting and wonder why the results are inconsistent. Luna: A fair point. The easy way out is always last-click attribution. It requires zero thinking. But it’s a trap that keeps marketing stuck in a cycle of optimization without actual growth. Lucas: Right. You optimize the wrong thing, and you accelerate toward a cliff. Building a robust, incrementality-based framework is harder, but it’s the only way to ensure that every dollar you spend is actually moving the needle for the business. Luna: So the question for our listeners this week isn’t just which tool to use, but how honest are you willing to be with your own data? Lucas: That’s the core of it. Are you measuring what happened, or are you measuring what matters? In a world where attention is fragmented and privacy is tightening, measuring what matters is the only competitive advantage left.