Last-Click Lies.
If you judge your marketing by last-click attribution, you are systematically rewarding the channels that happen to be last in line and punishing the ones that actually created the demand. That leads to cutting exactly the wrong things.
Last-click attribution gives 100% of the credit to the final touchpoint before conversion โ usually branded search or direct, which are just people who already decided. It under-credits the awareness and consideration channels that created the demand. Better approaches: multi-touch attribution, incrementality testing (turn a channel off and measure the impact), and tracking blended metrics like total CAC and marketing-influenced pipeline rather than channel-level last-click ROAS.
01The trap in one example
Someone sees your Instagram ad, does not click. A week later they read your blog post from a Google search. A few days after that they search your brand name directly and convert. Last-click attribution gives 100% of the credit to branded search and 0% to the Instagram ad and the blog that actually did the work.
Cut the Instagram ad and the blog because they "do not convert," and a few months later your branded search volume mysteriously drops. You cut the top of the funnel and wondered why the bottom dried up.
02Why it persists
Last-click is the default in most analytics tools because it is simple and unambiguous. There is no debate about which touch gets credit โ the last one does. That simplicity is exactly why it is dangerous: it gives a confident, precise, and wrong answer.
03What to measure instead
- Multi-touch attribution โ distribute credit across the touchpoints in the journey, not just the last one
- Incrementality testing โ the gold standard. Turn a channel off (or run a geo holdout) and measure what actually changes. This tells you what a channel truly contributes.
- Blended CAC โ total sales and marketing cost divided by new customers, across all channels. Harder to game than channel-level numbers.
- Marketing-influenced pipeline โ for longer sales cycles, track every deal any marketing touch contributed to
04The practical move
You do not need a data-science team. Start by switching your reporting away from last-click as the sole truth. Layer in a data-driven or position-based attribution model, run one incrementality test on your biggest channel, and watch blended CAC over time. The goal is to stop making channel decisions based on a metric that structurally misleads you.
Get your attribution set up right
23+ years, performance-obsessed, AEO/GEO specialists. Marketing that actually moves numbers.
Get in Touch → 615-274-9555