30-second answer: Marketing attribution credits revenue to the touches that produced it. The vocabulary divides into model classes (first-touch, last-touch, linear, time-decay, U-shaped, W-shaped, data-driven), measurement classes (sourced, influenced, lift), data terms (UTM, server-side tagging, identity graph, deterministic match), and operating terms (path length, attribution window, view-through, cookieless). This glossary defines 22 attribution terms B2B marketers need to read vendor documentation and design measurement frameworks.
See cookieless, account-level attribution running inside Abmatic AI, book a demo.
First-touch attribution credits 100 percent of revenue to the first marketing touch in an opportunity's journey. It is biased toward awareness channels (paid social, content syndication) and undercounts late-stage touches.
Last-touch attribution credits 100 percent of revenue to the final marketing touch before opportunity creation or closed-won. It overweights late-stage channels (branded search, direct, email) and ignores everything that came before.
Linear attribution distributes credit equally across every touch in the journey. It is simple to compute and interpret but fails to recognize that some touches matter more than others.
Time-decay attribution gives more credit to touches closer to opportunity creation, with the weight decreasing on a half-life. It is a sensible default when later touches are more predictive.
U-shaped attribution gives 40 percent to first touch, 40 percent to opportunity-creation touch, and distributes the remaining 20 percent across middle touches. It privileges the start and the conversion moment.
W-shaped attribution adds a third weighted touchpoint at MQA or MQL conversion in addition to first and opportunity-creation, splitting weight 30/30/30/10. It is the standard for B2B with a defined demand-stage funnel.
Data-driven attribution uses a model (often a Markov chain or Shapley value) to assign credit based on each touch's marginal contribution to conversion in the observed data. It removes hand-tuned weights but requires conversion volume to train reliably.
Sourced pipeline credits the program or channel that produced the first touch on a closed-won opportunity. It is the cleanest single-credit number but undercounts collaborative wins.
Influenced pipeline credits any program whose touch appears in an opportunity's journey, regardless of position. The same revenue can be influenced by multiple programs, so totals exceed real revenue. See multi-touch attribution for ABM.
Incrementality measures revenue that would not have happened in the absence of a marketing program, usually via geo-holdouts, randomized control groups, or causal inference. It is the gold-standard answer to did this campaign actually work.
Halo effect describes spillover from one channel to another that standard attribution models miss, such as paid social driving branded search. Incrementality testing is the right way to measure halo.
UTM parameters are query-string tags appended to URLs (utm_source, utm_medium, utm_campaign) that pass campaign metadata to analytics tools. Disciplined UTM hygiene is the foundation of any attribution stack.
Server-side tagging routes web events through a vendor's own server before forwarding to ad platforms and analytics tools. It improves data quality, sidesteps browser ad-blocking, and is the basis of cookieless attribution. See how to do cookieless attribution.
A conversion API is the vendor-side endpoint that ad platforms (Meta, LinkedIn, TikTok) expose for server-side conversion uploads. CAPI is the standard mechanism for cookieless conversion measurement.
An identity graph stitches device, cookie, email, and CRM identifiers into a single person-or-account record. It is the backbone of multi-channel attribution in cookieless environments.
Deterministic match resolves identity using exact identifiers (email hash, account ID, device ID). It is high-confidence, but volume-limited.
Probabilistic match resolves identity using behavioural signals (IP, user agent, geo, time pattern). It is wider in coverage but lower in confidence.
The attribution window is the maximum number of days between a touch and a conversion for the touch to receive credit. Standard B2B windows run 30 to 90 days, sometimes longer for enterprise deals.
Path length is the number of distinct marketing touches in an opportunity's journey before close. B2B path lengths typically run 8 to 25 touches; long paths are why first-touch and last-touch models are inadequate.
A view-through conversion credits a display or social ad that was viewed (impressed) but not clicked, when the same account converts within the attribution window. View-through credit is contentious but real for awareness channels.
A click-through conversion credits a marketing touch that was clicked. It is the cleanest signal but biased away from upper-funnel impression-only channels.
Cookieless attribution credits revenue to marketing without third-party cookies, using server-side tagging, identity graphs, deterministic CRM joins, and account-level resolution. See how to do cookieless attribution.
Account-level attribution credits revenue at the account rather than the contact level, summing touches across all known contacts at the account. It is the dominant attribution lens for B2B because the buying committee, not the individual, is the decision unit. See identity resolution and account graph.
Ready to see account-level attribution in production? Book a demo of Abmatic AI.
For most B2B teams the right answer is W-shaped (with explicit MQA/MQL stage credit) reported alongside data-driven attribution as a check. First-touch and last-touch are useful only as supporting cuts because both undercount the middle of the funnel.
Cookieless tracking shifts the foundation from browser cookies to server-side IDs and account-level joins. The model class (W-shaped, time-decay) is unchanged; the data plumbing changes substantially. See how to measure ABM ROI.
Both. Sourced answers which programs originate pipeline. Influenced answers which programs help close pipeline. ABM programs typically report influenced as the primary number because ABM is rarely a single-touch motion.
The window should match the actual sales cycle. Mid-market B2B usually settles at 60 to 90 days; enterprise often runs 180 days or more. Setting the window too short under-credits early-funnel programs.
Incrementality testing answers questions standard attribution cannot: did this campaign cause incremental revenue, or would the revenue have happened anyway. The right cadence is one or two designed lift tests per quarter on the largest investments.
Attribution is a measurement discipline, not a single number. The cleanest stacks combine a multi-touch model (W-shaped or data-driven) for ongoing reporting with periodic incrementality tests for big-investment validation. Use this glossary as a reference when reading attribution vendor documentation and designing reporting frameworks.