U-Shaped Attribution Model: How It Works in B2B
Library > U-Shaped Attribution Model
Written by Stephanie Dean
Published on 26 October 2026
What is U-shaped attribution?
U-shaped attribution is a position-based, multi-touch model that gives extra credit to the first and last eligible interactions in a measured journey. The remaining credit is shared among the interactions between them.
It helps B2B marketers compare the recorded introduction and endpoint with the supporting activity in the middle. The selected outcome determines where that measured journey ends.
In Dreamdata's U-shaped model, the first and last touch each receive 40%, while middle touches share 20%. Those percentages are allocation rules, not measured estimates of how much business each interaction caused.
You can analyze a journey ending at a lead event, opportunity milestone, or won deal, depending on your reporting setup. U-shaped is not inherently limited to lead creation or unable to include later sales activity.
Choose the outcome that fits your decision, then inspect the eligible interactions. The useful question is how emphasizing the two ends of that journey changes your interpretation of channel performance.
How the 40-20-40 allocation works
Consider the example of a $200,000 won-deal with six eligible interactions. U-shaped places substantial credit at each end, creating the shape that gives the model its name.
Keep the deal value separate from the model: the outcome remains $200,000 whichever allocation you choose.
The allocation in this six-touch example
First eligible interaction: 40%, or $80,000.
Last eligible interaction: 40%, or $80,000.
Four middle interactions: 20% combined, or $40,000; each receives $10,000.
The six allocations add up to $200,000. With more eligible middle interactions, each would receive a smaller share of the same 20% pool. Check the tool's rules for short journeys with fewer distinct interactions.
The selected outcome changes the measured journey
For opportunity-creation analysis, the endpoint is the final eligible interaction before that milestone. For won-deal analysis, additional later activity may be included, placing the final 40% on a different interaction. The endpoint is not automatically the lead-creation event.
What qualifies as an interaction?
Eligibility depends on your recorded sources, event grouping, time coverage, and exclusions. A website session, webinar attendance, or sales meeting may qualify under the defined scope. Missing activity cannot receive credit, and excluding an event can change which interaction is first or last.
Inspect a known journey and reproduce its allocation from the documented rules. That checks whether the reported model means what your team expects before you compare campaign totals.
Connect the B2B buying group to the outcome
The first and final interactions may come from different people. An account-level view brings relevant activity together around the company and opportunity being analyzed.
For example, a champion first visits through paid social, an analyst researches integrations, and a finance stakeholder later joins a pricing discussion. The account relationship lets you examine their activity as part of the same buying process.
If the chosen outcome is a won deal, eligible post-lead activity can appear in the journey. Its share depends on its position under the model, rather than being excluded simply because a lead already exists.
Make these definitions clear in the report:
First interaction: the earliest eligible recorded touch, which may come after untracked research or a recommendation.
Final interaction: the last eligible touch before the selected outcome, using the tool's exact event and timing rules.
Included activity: the sources, event types, grouping, and exclusions that determine the journey.
Account and opportunity relationships: how contacts and interactions connect to the relevant buying process.
Value: whether the allocated amount is pipeline or a defined won-deal measure, using consistent currency and dates.
Dreamdata's account-based data model connects contacts, activity, companies, and business outcomes across sources. That structure supports automated collection and attribution; the U-shaped model supplies the weighting rule.
Use U-shaped results to investigate campaign roles
Start with a decision, such as whether to keep funding a program that introduces relevant accounts but rarely captures the final interaction. Compare the U-shaped result with the underlying journeys and campaign cost.
Inspect both ends of the journey:
Introduction: which campaigns first appear in accounts that reach the chosen outcome, and do those accounts fit your target market?
Endpoint: which interactions appear last, and what happened earlier? A high final-position share can reflect returning buyers as well as a useful conversion path.
Review the middle, too. A webinar receiving a small share may have helped buyers evaluate, but U-shaped reserves only 20% for all middle activity. Its position explains the low allocation; it does not establish low importance.
In the six-touch example, each middle interaction receives $10,000 under U-shaped. Linear would allocate one-sixth of the $200,000 to each interaction, about $33,333.33. The difference comes from the rule, not a new business result.
Use the comparison to decide what to investigate or test. Check account fit, program cost, later progress, and buyer feedback before treating a larger credit share as a reason to increase spend.
The assumptions and limitations to understand
U-shaped deliberately emphasizes the endpoints. That can be a useful perspective, but it may give relatively little credit to substantive activity in the middle of an evaluation.
It can be applied to different outcomes when the reporting setup supports them. The key limitation is its fixed positional weighting, rather than an inherent inability to analyze pipeline or revenue.
Three common interpretation errors
Treating the first recorded touch as the buyer's first awareness. Earlier conversations or research may be absent from the data.
Treating the last eligible interaction as the cause of the conversion. It may reflect a decision already made.
Comparing different scopes. A change in outcome, exclusions, or account relationships can change the allocation even when the model name stays the same.
Compare another model when its assumptions matter
Linear provides an equal-weighting comparison. W-shaped emphasizes an additional point under its defined rules, while data-driven attribution calculates weights from patterns across journeys. None is an automatic upgrade; each offers a different perspective.
Keep the same accounts, outcome, dates, and exclusions when comparing models. Inspect individual customer journeys behind material differences and explain any change to your primary reporting method.
FAQs
Does U-shaped attribution credit lead creation or the last touch?
Check the implementation. Dreamdata's U-shaped model emphasizes the first and last touch in the measured journey. The selected outcome determines its endpoint, so the last eligible interaction is not always a lead-creation event. Other tools may define their anchors differently.
Should you customize the 40-20-40 split?
Only when you have a clear reason for different allocation rules and a tool that supports them. Document the assumptions and compare the effect using the same journeys. Changing weights to favor a campaign does not demonstrate that the campaign became more effective.