A customer rarely buys after a single advert. They might see a social post, read a blog, click a search ad weeks later, then finally convert after an email. So which of those deserves the credit? That question is the whole of marketing attribution — and the model you pick to answer it quietly shapes every budget decision that follows.
What attribution is
Attribution is how you assign credit for a conversion across the marketing touchpoints that led to it. A conversion might be a sale, a sign-up or a lead. The touchpoints are every interaction along the way — ads, emails, organic search, social, and so on. An attribution model is simply the rule you use to divide the credit among them.
Why it matters is straightforward: credit drives budget. If your model hands all the glory to the last click, you will pour money into closing channels and starve the ones that created awareness in the first place. Choose a fairer model and you fund the whole journey, not just its final step. (For the broader foundations, our guide to marketing attribution sets the scene.)
The five models, side by side
There are many models, but five cover the landscape. Here is how they share credit across a journey.
| Model | How credit is shared | Best at | Weakness |
|---|---|---|---|
| First-touch | 100% to the first touchpoint | Valuing awareness | Ignores everything after |
| Last-touch | 100% to the final touchpoint | Valuing conversion | Ignores everything before |
| Linear | Split equally across all touchpoints | Simplicity, fairness | Treats every touch as equal |
| Time-decay | More credit to touchpoints nearer the conversion | Longer sales cycles | Undervalues early awareness |
| Data-driven | Credit assigned from your own data | Realism | Needs volume and good tracking |
Each row is a different philosophy about what "deserves" credit. Let us walk through them.
Single-touch models: simple but blinkered
First-touch attribution gives all the credit to the very first interaction. Its logic: nothing happens without awareness, so reward whatever started the journey. The flaw is obvious — it pretends the dozen touchpoints that nurtured and closed the customer never mattered.
Last-touch attribution does the opposite, crediting only the final interaction before conversion. It is the most common model precisely because it is the easiest to track, and it flatters bottom-of-funnel channels. But it is arguably the most misleading: it rewards the channel that happened to be standing nearest the finish line, while ignoring everything that got the customer there.
Single-touch models are popular because they are simple. They are dangerous because real customer journeys almost never have a single touch.
Multi-touch models: sharing the credit
To reflect the journey more honestly, multi-touch models split credit across several touchpoints.
Linear attribution divides credit equally among every touchpoint. If a customer interacted five times, each gets a fifth. It is admirably fair and easy to understand. Its weakness is that it treats a casual glance and a decisive demo as equally important, which they rarely are.
Time-decay attribution weights touchpoints by recency: the closer an interaction was to the conversion, the more credit it earns. This suits longer, considered purchases where late-stage nudges matter — but it systematically undervalues the early awareness work that made the sale possible at all.
Data-driven attribution: letting the data decide
The most sophisticated approach abandons fixed rules entirely. Data-driven attribution uses your own conversion data — often with statistical or machine-learning methods — to estimate how much each touchpoint actually contributed, based on the patterns in journeys that converted versus those that did not.
In principle this is the most accurate model, because it reflects your real customers rather than a tidy assumption. In practice it has a catch: it needs volume and reliable tracking to work. Without enough conversions and clean data, the model has nothing trustworthy to learn from, and a fancy method on thin data is worse than an honest simple one.
This is the kind of measurement nuance that experienced practitioners stress. London consultancy CM Beyer, for example, published a detailed comparison of attribution models from first-touch to data-driven, making the point that the right model depends on the business, not on which one sounds most advanced.
How to choose
There is no universal "best" model. The sensible choice depends on three things:
- Your data. Limited conversions push you toward simpler models; rich data unlocks data-driven ones.
- Your sales cycle. Long, multi-touch journeys reward multi-touch and time-decay thinking; quick impulse purchases tolerate simpler models.
- The decision at hand. Measuring brand awareness calls for one lens; optimising a closing channel calls for another.
A pragmatic path: start with a multi-touch model rather than last-touch, so you stop ignoring most of the journey, and graduate to data-driven once you have the volume and tracking to support it. Whatever you choose, pair it with the wider discipline of measuring marketing ROI — attribution is one input into that picture, not the whole of it.
The bottom line
Attribution decides which touchpoints get credit for a conversion, and that decision quietly steers your budget. First-touch and last-touch are simple but blinkered, crediting only the start or the end of a journey. Linear and time-decay share credit more honestly across touchpoints. Data-driven attribution is the most realistic but the most demanding, needing volume and clean tracking to earn its keep. The goal is not to find a perfect model — none exists — but to choose one that reflects how your customers actually buy, and to keep its limits firmly in mind.