It’s time to move on from multi-touch attribution

Multi-touch attribution (MTA) is a waste of time and money. This isn’t a controversial statement anymore. Few disagree except for MTA vendors, and even they have quietly hedged their bets, introducing media mix modeling (MMM) and incrementality testing. Wise.

Yet brands continue to fall into the MTA trap. I’ve watched too many marketing teams realize that last-click or platform attribution is leading them astray, only to be sold the seductive promise of MTA.

Whether it’s pitched hard by a vendor or by a CMO looking to justify their existence for a 12-month project under the guise of getting our house in order, the outcome is predictably the same: no meaningful changes are made that make a marketing program more profitable.

The promise vs. the reality

Multi-touch attribution vendors tell a compelling story that sounds almost too good to be true — because it is.

The promise

Data-driven credit distribution across all touchpoints based on sophisticated statistical modeling and comprehensive customer journey analysis across every touchpoint. Finally, you’ll move beyond crude last-click attribution to elegant, nuanced measurement that captures the full complexity of modern customer journeys.

The reality

Still click-obsessed: Despite all the sophisticated modeling and data-drivenness, MTA systems heavily weight clicked interactions because they’re the most trackable and reliable data points. You end up with a marginally better-than-last-click attribution wrapped in significantly more complexity and cost.

Built on dying data: iOS 14.5+ and other tech changes, cookie deprecation and evolving privacy regulations systematically destroy the data foundation MTA requires. 

Same optimization mistakes: Even with a theoretically perfect MTA, every touchpoint tracked and accounted for, you’ll still optimize toward attributed performance rather than actual incremental impact. The fundamental flaw remains unchanged. 

MTA may tell you that someone:

• Clicked your Facebook ad.
• Saw a display ad.
• Watched a YouTube ad.
• Got served a TikTok ad and a Reel.
• Searched for you.
• Saw a Google ad.
• Then purchased. 

It cannot tell you which of your ads influenced that purchase decision or whether the customer would have bought anyway.

Dig deeper: What your attribution model isn’t telling you

The psychology 

Beyond the technological limitations, human psychology keeps brands trapped in the MTA cycle.

The sunk cost spiral

Brands invest so much time, effort and energy (not to mention money) in making MTA work that admitting defeat becomes challenging. The more resources invested, the harder it becomes to walk away, even when the evidence clearly shows it’s not delivering value.

The comfort of familiarity

Because MTA isn’t all that different from last-click or platform attribution, it’s easier to get organizational buy-in. Marketing teams comfortable with attribution-based reporting find MTA safer than other measurement methodologies.

For years, marketing teams have justified their budgets (and value to the organization) via attribution-based reporting. Changing the measurement philosophy feels risky. Sticking with the same approach in a more sophisticated package feels safer, despite leading to the same fundamental problems.

Dig deeper: How attribution masks what’s actually driving growth

Where attribution belongs

Attribution isn’t entirely worthless. You need to know its role. Attribution provides a real-time signal for operational decision-making, but don’t take it at face value for strategic budget allocation.

The solution is straightforward: augment attribution with learnings from incrementality testing and/or well-built MMM. Apply multipliers, coefficients or adjustment factors — call them whatever you want — to bring attributed numbers closer to causal reality.

For example, through testing, bottom-funnel tactics drive many conversions that would happen regardless of ads and end up with a multiplier of less than 1. Mid- and upper-funnel tactics that do not receive appropriate click credit end up with multipliers greater than 1 to account for the halo effect not captured in attribution. 

While these adjustments don’t maintain perfect accuracy over long periods or through large-scale campaign changes, they provide a more realistic, real-time picture of actual contribution than raw attribution data.

Dig deeper: The smarter approach to marketing measurement

MTA and triangulation

I recently walked a brand through my perspective after they’d invested heavily in building an MTA system but were questioning whether to continue. As we dug into the details, it became clear that their model showed only about 10% of conversions involved multiple touchpoints. That wasn’t true, but it was all their attribution system could see. In the end, their sophisticated MTA setup was effectively delivering the same results as last-click attribution. No better insights. No better decisions. No added value.

My core issue with how MTA vendors talk about triangulation is combining MTA, MMM and incrementality testing. While I agree with some concept elements, the “M” and “T” are useless. Instead of layering complexity, brands should anchor their multipliers directly on simpler click-based or platform attribution.

The real path forward

We like attribution because it gives us a single number to grade our marketing efforts. However, there is no perfect way to measure marketing effectiveness, and there never will be. Instead, we need to use a series of tools to better understand what’s working and what’s not and how to adjust our strategy and budget to drive more growth or profit. 

While I’ve seen attribution (of any flavor) lead brands astray when used in a silo, brands leveraging consistent incrementality testing and MMM are making better decisions. Thankfully, there is no shortage of vendors at every price point, not to mention open-source tooling, that are making better measurement more accessible. 

The bottom line

Multi-touch attribution overpromises and underdelivers. The sooner marketing leaders accept that attribution cannot solve the fundamental question of incremental impact, the sooner they can invest in measurement approaches that deliver insights that, when acted on, drive better business outcomes and allow them to build marketing programs that are no longer cost centers, but profit centers.

Dig deeper: Why causal AI is the answer for smarter marketing