The attribution-versus-incrementality conversation has been running for years and it’s not done. In every CMO I’ve worked with through 2024-26, there’s been the same uncomfortable moment. The attribution dashboard says one thing about which channels are driving conversions. The incrementality testing — when it gets done properly — says something different. The reconciliation work is hard, the political dynamics are real, and the operational pressure to keep spending against the attribution numbers is constant.

This is a working view of how CMOs are actually handling the tension in May 2026.

The disagreement is structural

The reason attribution and incrementality keep disagreeing is structural, not a flaw in either methodology. They measure different things.

Attribution describes the conversion paths customers actually took. It says which channels got credit in the path-based attribution model, which deserves credit by some last-touch or multi-touch logic. It doesn’t tell you what the customer would have done in the absence of any specific channel.

Incrementality describes the lift caused by a specific intervention. It says, of the conversions in the test cell that got the intervention, how many would have happened anyway in the control cell that didn’t. It tells you the causal contribution of the intervention.

These two are answering different questions. They will not always agree. The disagreement is informative, not a problem to solve.

What attribution is actually good at

Attribution is good for operational measurement at the channel level when the question is “did this channel deliver the volume we expected at the cost we expected.” For paid channels with clear bidding logic, attribution metrics are actionable and the operational team can run the channel against them.

Attribution is good for path analysis. Understanding the typical sequence of touchpoints a converting customer experiences. This is useful for content strategy, sequencing, and creative.

Attribution is good for diagnostic alerts. A channel whose attributed performance suddenly drops is a signal worth investigating. It might mean the channel actually weakened. It might mean tracking broke. Either way, the alert is useful.

Attribution is poor for budget allocation across channels with overlapping causal effects. The reason it’s poor is that the attribution credit is divided among the channels that touched the customer’s path, but the actual causal contributions don’t sum to the conversion. The math doesn’t work for budget decisions even though the dashboard makes it look like it does.

What incrementality is actually good at

Incrementality is good for budget allocation across channels and tactics where the question is “what would I lose if I cut this.” For major spending decisions, this is the right question.

Incrementality is good for evaluating new channels and tactics where the historical attribution data doesn’t exist or is unreliable. Lift testing in a new channel tells you whether the spend is actually doing anything.

Incrementality is good for periodic recalibration of media mix. Marketing mix modelling, properly run, gives you a view of what’s actually driving outcomes that’s much more reliable than rolled-up attribution.

Incrementality is poor for day-to-day campaign optimisation. The signal is too coarse and the time-to-results too long. You can’t tell a channel manager “wait two months while we run a holdout test” — they need decision support today.

The CMO reconciliation

The CMO reconciliation that actually works in 2026 has a few specific features.

Use attribution for tactical decisions and incrementality for strategic decisions. The day-to-day campaign management runs on attribution metrics. The quarterly budget allocation across channels runs on incrementality and MMM results. The two methodologies are explicitly assigned to different decision types and not asked to do each other’s jobs.

Run incrementality testing on a regular cadence, not as a one-off. The CMOs who run incrementality testing once and then leave it for years end up disconnected from the underlying causal picture. The CMOs who treat it as ongoing operational discipline maintain alignment between strategic direction and tactical execution.

Train the team on what each methodology is and isn’t. The team that doesn’t understand the difference will keep arguing about it. The team that does understands when each is the right tool.

Push back on the channel teams that complain about the strategic decisions. The channel teams will always advocate for the attribution numbers that flatter their channel. The CMO has to be willing to allocate budget against incrementality even when attribution shows a different picture.

The political dynamics

The political dynamics of attribution-versus-incrementality conversations are real. Channel teams are evaluated on attributed performance. The vendors selling each channel are evaluated on attributed performance. The internal advocates for each channel argue from attribution.

The CMOs who handle this well do a few things specifically.

They share the incrementality results transparently with the channel teams, including when the results are uncomfortable. The channel teams that learn the results matter even when they’re not flattering tend to engage constructively with the methodology. The teams that only see results when they’re flattering get sceptical of the whole exercise.

They protect the team running incrementality from political pressure. The team that runs the lift tests has to be insulated from the channel teams that have an interest in particular results. Otherwise the methodology gets compromised.

They invest in education across the marketing function and into the broader organisation. The CFO who doesn’t understand why attribution and incrementality disagree will assume one of them is wrong. The CFO who understands the structural difference can engage constructively with the strategic decisions.

The vendor landscape

The vendor landscape for both attribution and incrementality has continued to mature through 2024-26. The attribution platforms have continued to refine their methodology, particularly in handling of cookieless and SKAN-style measurement. The incrementality platforms — including the lift testing tools and the modern MMM platforms — have continued to expand their capability.

The CMOs that have invested in modern incrementality tooling have generally produced better strategic decision-making than the CMOs who relied entirely on attribution. The investment is real but the case for it is straightforward when the alternative is making major budget decisions on data that doesn’t actually answer the budget allocation question.

For organisations building this capability internally rather than relying entirely on vendor platforms, engaging an analytics consultancy for the methodology and tooling design has been a productive pattern. The internal team owns the operational execution; the consultancy brings the methodology depth.

The privacy environment

The privacy environment continues to shape both attribution and incrementality work. The deprecation of third-party cookies, the iOS App Tracking Transparency framework, the broader regulatory environment around tracking — all of these have made attribution harder while making incrementality more important.

The shift is reasonably clear. Channel-level attribution is less reliable than it used to be. Aggregate measurement of marketing effect is more reliable, partly because it depends less on user-level tracking. The structural pressure favours incrementality and MMM over user-level attribution.

CMOs that have anticipated this and invested in incrementality and MMM capability are positioned well. CMOs that have continued to invest only in user-level attribution platforms are increasingly disadvantaged.

What’s not changing

A few aspects of the landscape are unlikely to change.

Channel teams will continue to advocate from attribution. The compensation and evaluation frameworks for channel managers reward attributed performance. The structural incentive isn’t going away.

The vendor ecosystem will continue to encourage attention to attribution. The SaaS subscriptions and the broader vendor economics are easier to defend with attribution dashboards than with incrementality summaries.

The pressure to make decisions quickly will continue. Incrementality tests take time. The pressure to allocate budget today will continue to push toward attribution.

These structural realities are not solvable. They’re features of the environment. The CMOs that handle them best are the ones who acknowledge them and design their decision-making process to work around them.

Where this goes

The attribution-versus-incrementality reconciliation isn’t a problem that gets solved. It’s a tension that gets managed. The CMOs who manage it well are the ones who use both methodologies for what they’re good at, who run incrementality on a regular cadence, who educate their teams and stakeholders on the structural differences, and who are willing to make strategic decisions against incrementality even when attribution argues differently.

The CMOs who don’t manage it well end up making strategic decisions on tactical data, end up confused about why their numbers don’t add up, and end up rotating through measurement vendors looking for the one that finally produces clean answers. There isn’t one.

The tension is the price of operating in a complex multi-channel environment. It’s worth paying. The reconciliation work, done properly, produces better marketing decisions than either methodology alone.