Position-based attribution is a rules-based method for dividing credit for a conversion among the ads a person interacted with beforehand. In its standard form, the first interaction receives 40% of the credit, the last interaction receives 40%, and the remaining 20% is split evenly among everything in between. The model exists to correct two opposite distortions. Last-click reporting credits only the ad that closed the sale, while first-click reporting credits only the ad that started the journey. Position-based attribution assumes both ends of the path matter most, and that the middle still matters a little. Because a chart of its weights dips in the middle and rises at either end, it is also widely called the U-shaped model.

How the credit is split

The model works on a conversion path: the ordered list of ad clicks, and in some tools impressions, that a platform can tie to one person before a conversion. Each path is scored with fixed percentages chosen in advance, not learned from data.

In Google's legacy Universal Analytics example, a customer arrives through paid search, a social network, an email and a direct visit. "The Paid Search and Direct channels would each receive 40% credit," while "the Social Network and Email channels would each receive 10% credit," according to Google.

The arithmetic scales with path length. On a $200 sale preceded by five interactions, the first and last each receive $80. The three middle interactions share the remaining $40, about $13.33 each. The longer the path, the thinner the middle slice becomes: a touchpoint in the middle of a ten-step journey earns 2.5% of the conversion, however influential it was.

Short paths need special rules. Adobe's Customer Journey Analytics documentation states that "for conversions with a single touch point, 100% credit is given" and "for conversions with two touch points, 50% credit is given to both." Google's Search Ads 360 help page does not say how it handles them.

Several asymmetric variants sit beside the U shape. Adobe's J curve gives 60% to the last interaction and 20% to the first, and its inverse J reverses those weights. Its custom model accepts any first, middle and last weights, which are "normalized to 100% even if the custom numbers entered do not add to 100."

Where it sits in the stack

Position-based attribution is applied after conversions are recorded and matched to prior interactions within a lookback window, the period during which an ad can still earn credit. Its influence comes in reporting and, where a platform allows it, in automated bidding. When Google added six attribution models to AdWords in May 2016, product managers Joan Arensman and Wilfred Yeung wrote that advertisers using automated bidding would have "bids optimized automatically to reflect your new model." A generic keyword that usually opens journeys reports more conversions under position-based rules than under last click, and a bidding algorithm reading those numbers spends more on it.

As of October 2026, the model remains available in Google's enterprise products even though it has left Google Ads and Google Analytics 4 (GA4). Search Ads 360, Google's search campaign management platform for agencies and large advertisers, lists position-based among nine basic models, assigning "40% of a conversion to the first click, 40% to the last click," according to its help centre. Campaign Manager 360 (CM360), Google's ad server, lets users choose position-based as a baseline in its attribution modeling tool and "set what percentage of credit goes to the first interaction and the last interaction." Display & Video 360 (DV360), Google's demand-side platform, offers position-based as one of six baseline models. Adobe offers it as U-shaped. The settings are operated on the buy side by analytics teams, agencies and search managers; sellers have no say in which model a buyer applies.

Origin and evolution

The model grew out of early-2010s ad server and analytics tools. Google added an Attribution Modeling Tool to Google Analytics Premium in 2012, and on October 29, 2012, announced "wider availability for the Attribution Modeling Tool through a public allowlist," according to a Google Analytics blog post by product manager Bill Kee. A month earlier, on September 27, 2012, Kee had announced attribution modeling for DoubleClick for Advertisers (DFA), the predecessor of CM360. One of its six standard models was "Position Based," giving 40% to the first interaction, 20% to the middle and 40% to the last.

The default weights became a point of debate almost immediately. Analytics author Avinash Kaushik, in a 2013 post on his Occam's Razor blog, rated position-based as "not great, but won't completely kill you" and called it "a dangerous first model to use." "Promise me you won't ever use the default one," he wrote, urging readers to build custom versions. Brian Clifton, an analytics author and consultant, replied on August 15, 2013, arguing readers should not "under-estimate the value of the default 40/20/40 split," particularly for companies with weak brand recognition, for whom the first touch carries more weight.

Search advertising adopted the model three years later. Google's May 10, 2016 announcement brought last click, first click, linear, time decay, position-based and data-driven models to AdWords conversion settings. In March 2020, Google revamped the Attribution section of Google Ads, adding a model comparison report across the six models.

The reversal came in 2023. On April 6 of that year, Google said first click, linear, time decay and position-based models would leave Google Ads and GA4. Fewer than 3% of Google Ads web conversions used them, according to Google's own data for February and March 2023. Google first planned to stop new selections in June 2023, then postponed to mid-July, according to Search Engine Journal. Conversion actions still on the four models were switched to data-driven attribution in September 2023. The dates for GA4 differ by source: Search Engine Land reported a mid-October 2023 retirement, while Google's GA4 documentation states the models were "no longer available as of November 2023." Universal Analytics, where position-based had lived in the Multi-Channel Funnels reports, was shut down for 360 customers on July 1, 2024.

Why it matters for marketers

Position-based rules make introducing and closing channels look strong at the expense of everything that nurtures a buyer in between. For advertisers whose video, display and social spend was invisible under last click, it was often the first step towards crediting those channels.

After 2023, advertisers on Google Ads and GA4 could choose between data-driven attribution and last click, with GA4 also offering Google paid channels last click, which gives 100% of credit to the last Google Ads channel where one exists. The transparent middle ground disappeared. Rémi Kerhoas, a digital marketing consultant, said Google retired the models "because they were unused (supposedly)" and argued that "Google wants advertisers to rely on DDA so they can maximize revenue."

The choice is not cosmetic. PPC Land's explainer on multi-touch attribution runs one $100 path through four rules and finds a display ad's credit ranging from zero to $40.

Limitations and criticisms

The central criticism is that the weights are arbitrary. Nothing in a customer's behaviour establishes that the first touch deserves exactly 40%. The same split applies to a $5 impulse purchase and a six-month enterprise software deal.

Second, the model ignores non-converting paths. It cannot tell whether a channel that often appears first also appears first in journeys that go nowhere. Algorithmic models compare the two populations; rules-based models cannot.

Third, the model is sensitive to what the platform can see. "First" means first observed, not first in reality. A podcast advertisement, a television spot or an untracked organic search may have started the journey, but if the earliest recorded event is a retargeting click, that click takes 40%. Consent gaps and shorter cookie lifetimes compress paths, collapsing the model towards a two-touch or single-touch split.

Finally, like all attribution, it measures association rather than causation. Experiments with unexposed control groups, covered in PPC Land's explainer on incrementality, address that gap; position-based rules do not attempt to.

Defenders counter that transparency has value. Any analyst can reproduce a position-based report by hand, which cannot be said of a data-driven model whose weights Google does not publish.

Not the same as

Data-driven attribution learns credit from converting and non-converting paths with a statistical model rather than fixed percentages. It replaced position-based for affected Google Ads conversion actions in 2023.

Linear attribution gives every interaction equal credit, regardless of position. On a four-step path, each touch gets 25%, against 40/10/10/40 under position-based rules.

Time decay attribution weights interactions by how close they fall to the conversion, using a half-life of seven days in Search Ads 360 and CM360. It favours the end of the path only, while position-based favours both ends.

Last-click attribution gives 100% to the final interaction. It is the model position-based rules were designed to soften.

In business-to-business marketing, the label differs. In Adobe's Marketo Measure, formerly Bizible, "U-shaped" gives 50% each to the first touch and the lead creation touch, with nothing to later interactions, according to Adobe.

Recent developments

Google has not restored position-based attribution to Google Ads or GA4, but it has built position-style views on top of data-driven credit. In January 2026 it introduced an attribution analysis report in GA4 that splits data-driven credit into early, mid and late stages, with early and late each defined as the first and last 25% of multi-touch paths.

TikTok's Attribution Portfolio, announced on May 13, 2026, reports first-touch and last-touch layers side by side. GA4 also made click-through lookback windows editable to any value from 1 to 90 days from August 11, 2026. As of October 2026, position-based attribution remains a selectable model in Search Ads 360, CM360, DV360 and Adobe's analytics products.

Timeline

  • 2012 - Google adds an Attribution Modeling Tool to Google Analytics Premium
  • September 27, 2012 - Attribution modeling arrives in DoubleClick for Advertisers with a "Position Based" 40/20/40 standard model
  • October 29, 2012 - Google announces wider availability of the Attribution Modeling Tool through a public allowlist
  • 2013 - Avinash Kaushik advises against using the default position-based weights
  • August 15, 2013 - Brian Clifton defends the default 40/20/40 split
  • May 10, 2016 - Google announces six attribution models for AdWords, including position-based, with bidding integration
  • March 2020 - Google revamps Attribution reporting in Google Ads with a model comparison report
  • April 6, 2023 - Google announces the removal of first click, linear, time decay and position-based models from Google Ads and GA4
  • June 21, 2023 - Search Engine Journal reports the Google Ads change postponed from June to mid-July
  • September 2023 - Remaining Google Ads conversion actions on the four models switch to data-driven attribution
  • Mid-October to November 2023 - The four models leave GA4 (Search Engine Land and Google documentation give different dates)
  • July 1, 2024 - Universal Analytics 360 is shut down
  • March 27, 2025 - Google adds last-touch comparison columns to Demand Gen
  • January 16, 2026 - Google announces the GA4 attribution analysis report with early, mid and late bands
  • May 13, 2026 - TikTok announces its Attribution Portfolio with first-touch and last-touch layers
  • August 11, 2026 - GA4 makes lookback windows editable

Summary

Who: Introduced in Google's DoubleClick and Google Analytics tools in 2012, used by advertisers, agencies and analytics teams, and offered today by Search Ads 360, Campaign Manager 360, Display & Video 360 and Adobe.

What: A rules-based attribution model that gives 40% of credit to the first interaction, 40% to the last and divides 20% among the middle, also known as U-shaped attribution.

When: Launched in 2012, added to AdWords in May 2016 and removed from Google Ads and GA4 between July and November 2023.

Where: In ad server, search management and analytics reporting, and in some automated bidding systems, applied to conversion paths within a lookback window.

Why: To credit both the channel that opens a journey and the one that closes it; critics say its fixed weights are arbitrary and blind to non-converting paths and untracked media.