Read the journey, not the channel

Most analytics setups are built to answer the wrong question. They ask "which channel drove this conversion?", and the tools happily answer, usually by crediting whatever the user touched last. Last-click attribution is the default in Google Analytics and most ad platforms for one reason: it is simple. It assigns 100% of the credit to the final interaction and fits neatly in a spreadsheet.

It is also misleading, in a way that quietly distorts decisions.

Picture a real path to purchase. Someone sees a display ad on Monday, reads a blog post on Wednesday, opens an email on Friday, then clicks a branded search ad on Saturday and converts. Last-click hands search all the credit. But search only worked because everything before it did the real work of building awareness and intent. That last click was the door; everything else built the house. Credit the door alone, and you will keep defunding the rooms.

That is the practical damage. When last-click over-rewards bottom-funnel channels, teams pour budget into the "closers" and starve the channels that created demand in the first place. Months later the pipeline dries up and nobody knows why. The measurement caused the problem.

The deeper issue is that a channel is not a journey. Channels are how we organise our own teams and budgets: paid, organic, email, social. Users do not experience any of that. They experience one continuous path, hopping between devices and sessions, and they do not care that your SEO and your paid search sit in different departments. When you analyse channel by channel, you are looking at fragments of a story and treating each as the whole.

The fragmentation is literal, too. Analytics tools default to treating each device as a separate user, so a journey that starts on a phone and finishes on a laptop gets split into two disconnected half-stories. For anything with a considered purchase, where a decision can span six to twelve touchpoints over weeks, the standard setup measures the final sprint and ignores the marathon.

So what do you do instead? You start from the person, not the source. The question shifts from "which channel converted?" to "what path did this person take, and where did it stall?" That reframe changes what you build. You begin caring about how touchpoints hand off to each other, not just how each performs in isolation. You look at the sequence, not the silo.

None of this means attribution models are useless or that you should throw them out. Every model is wrong in its own way: first-click over-credits awareness, linear flattens everything to sameness, time-decay assumes recent always means relevant. The mistake is believing any single model captures the truth of a messy human decision. They are lenses, not answers.

Reading the journey is less about a specific model and more about a posture. Before you optimise a channel, ask what role it plays in the wider path. A channel that looks weak on last-click might be the one quietly starting every journey that later converts somewhere else. Kill it because a spreadsheet told you to, and you will feel the loss three months later, in a number no dashboard warned you about.

Start from the journey. The channels only make sense once you can see the whole thing.