Varun Mundra Get Another Perspective

Topic · Measurement

Attribution models: useful, and never quite true.

The main attribution models, what each gets wrong, and why chasing perfect attribution leads you astray.

The short answer

Attribution models assign credit for a conversion across the touchpoints that led to it — first-click, last-click, linear, time-decay, data-driven. Each tells a different, partial story. The essential thing to understand is that none is ‘correct’ — attribution is a useful lens, not truth, and treating any model as reality leads to systematically bad decisions.

The main models

Last-click (all credit to the final touch), first-click (all to the first), linear (equal across all), time-decay (more to recent touches), and data-driven (algorithmic). Each emphasises a different part of the journey and hides the rest.

Why none is right

Attribution models try to assign clean credit to something inherently messy and overlapping. Last-click over-credits the bottom of the funnel and ignores the brand-building that made the click likely; first-click does the reverse. Every model has a systematic blind spot.

Use them as a lens, not truth

The mature approach: use attribution to inform, not dictate. Combine it with brand tracking, incrementality tests and judgement. Chasing a single ‘perfect’ attribution number usually leads to over-investing in whatever’s easiest to attribute — typically at brand’s expense.

Related questions

Q

Which attribution model is best?

None universally — each has blind spots. Data-driven is often most balanced, but all should inform rather than dictate decisions.

Q

Why is attribution so hard?

Because customer journeys are messy, multi-touch and partly invisible — and brand effects happen before any trackable click.

Q

Should I trust my attribution data?

Trust it as one input, not the truth. Combine it with brand measurement and testing, and stay skeptical of precise-looking numbers.

Chat with me on WhatsApp
Get Another Perspective