Last-click attribution is the attribution model that assigns 100% of conversion credit to the final marketing touchpoint — typically the last click before the conversion — within the platform’s lookback window. It is a single-touch rules-based model: weight w = 1 on the terminal click, w = 0 on every earlier touch. In the multi-touch taxonomy it is the boundary case — the single-touch baseline every multi-touch model is measured against.
How last-click assigns credit
The mechanic is a one-line rule: weight w = 1 on the terminal click inside the platform’s attribution window, w = 0 on everything earlier. The other rules-based models redistribute credit across the path — first-click loads the opening touch, linear splits evenly, time-decay tilts toward touches near the conversion, position-based loads first and last. Data-driven attribution fits weights to observed paths rather than declaring them.
A four-touchpoint path
A customer sees a TikTok prospecting ad Monday, clicks a Meta retargeting ad Wednesday, searches the brand name Friday and clicks the Google result, then checks out. Last-click assigns 100% to Google branded search; the TikTok impression and Meta click read as zero.
Why platforms disagree
Each ad platform credits conversions inside its own tracking and lookback, with no cross-platform deduplication. Two distinct things hide behind the same labels.
Channel-level reporting in the ad managers is one. Meta Ads Manager applies a last-touch attribution default inside its own window; the specific click and view windows have shifted with platform changes (most recently in early 2026). TikTok Ads Manager applies a similar last-touch default inside TikTok’s window. Attribution-model defaults inside Google’s stack are the other: in 2023 GA4 changed its default reporting model to data-driven attribution, and Google Ads moved most conversion actions to data-driven attribution as the recommended default. Last-click remains selectable in both, but is no longer the headline default. Shopify’s standard channel reports sit separately, crediting the final UTM-tagged session before checkout.
Apply that to the path above. Meta credits the Wednesday click; Google Ads, if it served the Friday search, credits its click; Shopify reads Google as the final UTM. The same order is credited in full to multiple platforms simultaneously — a property of single-platform reporting, not a tracking bug.
Where the bias lands
Last-click credits the channel that closed the order, not the channels that built the demand. Routine beneficiaries are bottom-funnel: branded search, retargeting, direct visits, lifecycle email. Routine losers are upper-funnel: prospecting paid social, podcast, OOH, influencer seeding, top-funnel video. In the path above, the TikTok ad that created the demand reads as zero-ROAS.
Why operators still use it
The trade is well understood. Last-click is deterministic — same path, same result — explainable to a non-technical stakeholder, requires no modeling assumptions, and reproducible across tools. The known-bad measurement beats the opaque-better one in many operating contexts. The move is to pair it with incrementality for the budget question and MER for the topline efficiency read.