Latest / MBA Training Marketing / Attribution in 2026: why last-click is still killing your budget decisions
Transcript
- 0:00Welcome to the Leaders Insights Podcast.
- 0:07Today's episode, Attribution in 2026, why your last click data is lying to you.
- 0:12Last click attribution is still the default setting for most marketing stacks in 2026.
- 0:17Why hasn't it died yet?
- 0:19Because it's easy to defend in a budget meeting.
- 0:21You can point to a click, point to a sale, and draw a straight line.
- 0:25The problem is that line is fiction.
- 0:28And it's an expensive fiction.
- 0:30I've seen companies allocate 40% of their paid media budget to branded search terms that
- 0:35were simply collecting credit for purchases that were already decided.
- 0:39So the budget misallocation is happening at scale.
- 0:42Can you put a number on what that actually costs a business?
- 0:45Forrester published figures earlier this year, and I'll flag this is vendor-adjacent research,
- 0:51suggesting that companies running last-click models are misattributing somewhere between
- 0:5625% and 35% of conversion value.
- 1:00For a mid-size e-commerce brand spending $50 million a year on performance marketing,
- 1:05that's potentially $15 million flowing to the wrong channels, consistently, year after year.
- 1:11Where does that money typically end up?
- 1:13Which channel is the biggest beneficiary of the distortion?
- 1:17Good paid search and retargeting, almost universally.
- 1:21Those touchpoints sit at the bottom of the funnel by design.
- 1:24They intercept someone who already decided to buy, stamp their pixel on it, and claim
- 1:29the sale.
- 1:30Meanwhile, the podcast ad, the YouTube pre-roll, the newsletter sponsorship that actually created
- 1:36the intent, those get zero credit.
- 1:39Unilever and L'Oreal have been fairly public about overhauling their measurement frameworks.
- 1:44What did they actually change?
- 1:46Both move toward a combination of media mix modeling and incrementality testing rather
- 1:51than relying on pixel-based attribution alone.
- 1:54The shift is philosophical before it's technical.
- 1:58You stop asking, which ad did this customer click last?
- 2:02And start asking, what would have happened if we hadn't run this campaign at all?
- 2:07L'Oreal ran holdout experiments across several European markets in early 2026 and reportedly
- 2:14reallocated roughly 20% of their digital budget toward upper funnel placements after seeing
- 2:19that their retargeting spend had near zero incremental lift.
- 2:24Incrementality testing sounds rigorous but also slow and expensive.
- 2:28Is it realistic for a company that isn't L'Oreal?
- 2:31It's become much more accessible.
- 2:33Platforms like Northbeam and Rockerbox have built geo-based holdout testing into their
- 2:38standard offering at price points that a $10 million revenue brand can absorb.
- 2:43The honest answer is that even a rough incrementality test run quarterly is more accurate than perfect
- 2:50last-click data running daily.
- 2:53Precision is less valuable than directional truth.
- 2:56If a CMO is walking into a board meeting next week and needs to defend a measurement overhaul,
- 3:01what's the single argument that lands?
- 3:03Show them what branded search is claiming versus what organic would have converted anyway.
- 3:09Pull the overlap.
- 3:10In most businesses, that number alone is enough to make the case that the current model is
- 3:15actively misleading capital allocation decisions.
- 3:18And that's a risk conversation, not a marketing one.
- 3:22Clear and actionable.
- 3:23Thanks for cutting through it.
- 3:24To go further, Leaders Insights offers strategic training programs, diagnostic tools, and
- 3:29an executive self-assessment designed for CMOs, CDOs, CFOs, and leaders aiming for
- 3:36board-level performance.
- 3:38Or everything at MBA-training.com.
- 3:41Thank you for listening to the Leaders Insights podcast.
- 3:44See you tomorrow for another episode.