Last-Click Attribution

Like every model, last-click takes a person's recorded touchpoints, a conversion and an attribution window. Then one rule applies.

1. Take every touchpoint recorded for the person inside the window.
2. Find the latest one before the conversion. That is the last click, or the last touch where impressions count too.
3. Give it 100% of the conversion, and every earlier touchpoint 0%.
4. Sum per channel, campaign or ad.

Because the conversion and the credited touch sit close together in time, last-click needs the least data of any model. A tag that can only see the session in which the form was submitted can still produce a last-click report, which is why it became the default everywhere. The label hides several variants, and two reports called 'last click' can disagree widely because of which one they use.

• Last click: the plain rule, including direct visits.
• Last non-direct click: skips a direct visit and credits the last known channel before it. The old Google Analytics default.
• Last paid click: credits the last paid interaction and ignores organic and email in between. What each ad platform effectively reports about itself.
• Last touch: the same rule with impressions counted as touchpoints, so an ad that was seen but not clicked can hold the credit.

The same journey used across these glossary 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: €0.
• Day 8: Google Search adClicks a generic search ad on a category keyword and reads the pricing page. Credit: €0.
• Day 15: newsletter clickOpens the nurture email and clicks through to a case study. Credit: €0.
• Day 22: brand search ad, form fillSearches the company name, clicks the brand ad and books a demo. Credit: €12,000.

The brand search campaign reports €12,000 of revenue on a click that cost a euro or two. The Meta campaign that introduced the company reports nothing. Both statements are true under the rule, and both would lead a budget meeting astray.

• It is trivial to compute

  One touch, no weights, no half-life. Two people with the same data get the same answer, and a spreadsheet can reproduce it.

• It matches the CRM's moment

  The CRM creates the lead when the form is submitted, and the session around that form is exactly what last-click credits. The lead source field is, in most CRMs, a last-click value.

• It was the platform default

  Google Ads used last-click as its default until data-driven attribution took over for new conversion actions in late 2021, and Universal Analytics reported on last non-direct click for years. A generation of marketers learned to read reports through it.

• Brand search and retargeting harvest the creditPeople who already know you search your name or click the retargeting ad right before they convert. Those campaigns did not create the demand; they collected it. Last-click cannot tell the difference.
• Awareness looks worthlessProspecting, content, podcasts, events and organic social almost never hold the last click, so their reported revenue is close to zero. The model turns 'we cannot see it' into 'it did nothing'.
• Direct swallows returning visitorsA person who comes back a week later by typing the URL converts as 'Direct'. Universal Analytics invented 'last non-direct click' to skip over that, which shows how much traffic the plain model was misfiling.
• Each platform runs its own last clickMeta credits the last Meta interaction inside its attribution setting, Google the last Google click inside its window. Add the two together and the total is more than your CRM holds, because the same lead is the last click on both.

0:39

Last-click is a reasonable model when the journey really is short: a single-channel business, a same-day purchase, a local trade whose customers search and call within the hour. It also stays useful as one column in a longer table, because 'what closed this' is a real question. It stops being fine the moment it is the only column and awareness budgets are decided on it.

First-click attribution is the counterweight, and the models compared under attribution model are what a longer B2B journey deserves. Multi-touch attribution is what keeps the middle touches from disappearing entirely, and marketing attribution software is what records enough of the journey for the comparison to be possible.

Last-click attribution tells you what closed the lead, which is worth knowing, and hides everything that made the lead want to close, which is most of the marketing. Keep it as one view, never as the only one, and pay attention to which variant the report is using. The CRM's lead source field is usually a last-click value too, which is why the same distortion appears in revenue reports that never mention attribution at all.