Linear Attribution

Linear takes the person's recorded touchpoints inside the attribution window and applies one rule: count them, and divide the conversion by the count.

1. Collect every touchpoint recorded for the person before the conversion.
2. Count them. Each touch gets 1 divided by that count.
3. Multiply by the conversion's value, if you attribute revenue rather than lead counts.
4. Sum per channel, campaign or ad.

Because a single touch's credit shrinks as the journey grows, the model rewards channels that appear in many journeys rather than channels that appear in long ones. That is a mild bias in the right direction for most lead gen businesses, and a distortion in a few, covered below.

The same journey used across these entries: a €12,000 deal, four recorded touches over three weeks, the deal closing five weeks after the form.

• Day 1: Meta ad clickDownloads a guide from a Meta lead campaign. Credit: €3,000.
• Day 8: Google Search adClicks a generic search ad on a category keyword and reads the pricing page. Credit: €3,000.
• Day 15: newsletter clickOpens the nurture email and clicks through to a case study. Credit: €3,000.
• Day 22: brand search ad, form fillSearches the company name, clicks the brand ad and books a demo. Credit: €3,000.

Every channel that was in the room gets the same share. Compared with the €12,000 the brand search campaign showed under last-click, its number falls by three quarters, and the Meta and newsletter lines appear for the first time. Whether €3,000 is fair to the newsletter is exactly the question linear refuses to ask.

• An honest first multi-touch step

  It removes the single biggest distortion, all credit to one touch, without requiring anyone to agree on weights. That makes it easy to adopt.

• Nothing disappears

  Every channel that appears in the journey shows a number. Awareness, nurture and closing spend can all be seen on one table, which is the precondition for any better model.

• Easy to explain

  A client or a CEO understands 'split evenly' in one sentence. A seven-day half-life takes a paragraph, and a data-driven model takes a leap of faith.

• Intent is invisibleA pricing-page visit from a search ad and an accidental click on a retargeting banner get the same credit. The model has no way to say one of those was the decision.
• Touch count can be inflatedChannels that generate many small touches, retargeting above all, collect a share for each one. A journey with one prospecting click and five retargeting clicks gives retargeting 83% under linear.
• It inherits the storage problemLike every model, linear can only split what was recorded. If the first two touches are outside your cookie window or on another device, the split starts from the third touch and the opener still gets nothing.
• Openers and closers look alikeFirst-click and last-click each answer a real question. Linear answers neither; it reports participation, which is useful for budgets and useless for deciding what to scale.

Position-based attribution is linear with an opinion: 40% to the first touch, 40% to the last, the rest shared in the middle. Time-decay is linear tilted toward the finish, with credit halving for every week you go back. Data-driven models estimate the weights from converting and non-converting journeys. The comparison of all of them, on the same example, is under attribution model.

For a lead generation business with a sales cycle of weeks, linear is a good default report and a poor decision model. It works best as the middle column between a first-click and a last-click view, and it needs multi-touch attribution data that reaches back further than a browser cookie to be worth reading at all. Customer journey tracking is what makes the touch list complete enough to split.

Linear attribution is the fairest model and the least informative. Fair, because nothing recorded is ignored; uninformative, because it cannot tell an opener from a closer or a decision from an accident. Use it to see the whole journey on one page, then move to a weighted model for anything that decides budget.