{"id":21233,"date":"2020-01-30T15:57:56","date_gmt":"2020-01-30T15:57:56","guid":{"rendered":"https:\/\/www.arimetrics.com\/glosario-digital\/attribution-model"},"modified":"2026-10-02T17:19:49","modified_gmt":"2026-10-02T17:19:49","slug":"attribution-model","status":"publish","type":"encyclopedia","link":"https:\/\/www.arimetrics.com\/en\/digital-glossary\/attribution-model","title":{"rendered":"Attribution Model"},"content":{"rendered":"<p><img decoding=\"async\" class=\"boxpad alignright wp-image-22897 size-full\" src=\"https:\/\/www.arimetrics.com\/wp-content\/uploads\/2021\/11\/attribution-model.jpg\" alt=\"Attribution model\" width=\"300\" height=\"300\" srcset=\"https:\/\/www.arimetrics.com\/wp-content\/uploads\/2021\/11\/attribution-model.jpg 300w, https:\/\/www.arimetrics.com\/wp-content\/uploads\/2021\/11\/attribution-model-150x150.jpg 150w\" sizes=\"(max-width: 300px) 100vw, 300px\" \/><strong>Definition:<\/strong><\/p>\n<p>An <strong>attribution model<\/strong> is a rule, set of rules, or algorithmic method that determines how credit for a conversion is distributed among the recorded touchpoints in a person&#8217;s journey.<\/p>\n<p>The <strong>credit allocation<\/strong> changes how channel, campaign, and advertisement performance is reported, but it does not change the observed conversion. The same purchase may be assigned entirely to one interaction or distributed among several interactions depending on the model.<\/p>\n<p>Attribution works with the <strong>recorded data<\/strong> available to a platform within a defined scope and period. It therefore represents the path that the system can reconstruct, not necessarily every influence involved in the decision.<\/p>\n\n<h2>Attribution process<\/h2>\n<p>The process begins with a <strong>measurable action<\/strong>, such as a purchase, enquiry, or subscription. The platform orders the identified touchpoints that preceded the <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/conversion\">conversion<\/a> and forms a path within the analysed <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/funnel\">funnel<\/a>.<\/p>\n<p>The model then assigns a <strong>contribution value<\/strong> to each eligible interaction. This allocation feeds campaign and channel metrics, so it can change how <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/roi\">ROI<\/a>, <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/cpa\">CPA<\/a>, and other relationships between investment and outcomes are interpreted.<\/p>\n<p>The <strong>attributed unit<\/strong> may be a channel, campaign, advertisement, keyword, or another touchpoint. The result also depends on the time window, integrated sources, cross-device identification, and the rules each platform uses to admit or exclude interactions.<\/p>\n<h2>Model types<\/h2>\n<p>Models include <strong>predefined rules<\/strong>, which apply a known allocation, and data-driven methods, which estimate contribution from observed journeys. The most common categories include the following:<\/p>\n<ul>\n<li><strong>Last click:<\/strong> assigns all credit to the final eligible interaction before the conversion.<\/li>\n<li><strong>First click:<\/strong> attributes the entire outcome to the first identified touchpoint in the path.<\/li>\n<li><strong>Linear model:<\/strong> distributes credit equally among all included interactions.<\/li>\n<li><strong>Time decay:<\/strong> gives more weight to contacts that occur closer to the conversion.<\/li>\n<li><strong>Position based:<\/strong> reserves more credit for particular places in the journey, usually its beginning and end.<\/li>\n<li><strong>Data driven:<\/strong> uses information from converting and non-converting paths to estimate a different contribution for each interaction.<\/li>\n<\/ul>\n<p><strong>Actual availability<\/strong> varies between tools. Google Analytics 4 currently offers data-driven attribution, paid and organic last click, and Google paid channels last click; <a href=\"https:\/\/support.google.com\/analytics\/answer\/10596866?hl=en\" target=\"_blank\" rel=\"noopener\">Google Analytics<\/a> removed first-click, linear, time-decay, and position-based models from its reports. For compatible conversions, <a href=\"https:\/\/support.google.com\/google-ads\/answer\/6259715?hl=en\" target=\"_blank\" rel=\"noopener\">Google Ads<\/a> retains last-click and data-driven attribution.<\/p>\n<h2>Data scope<\/h2>\n<p>The <strong>observable path<\/strong> is built from events, identifiers, campaign parameters, and connections between platforms. An incomplete <a href=\"https:\/\/www.arimetrics.com\/en\/digital-glossary\/web-analytics\">web analytics<\/a> implementation may omit interactions or assign them to the wrong channel before the model performs any calculation.<\/p>\n<p>The <strong>attribution window<\/strong> defines how much time may pass between a contact and the conversion for that contact to receive credit. Channel definitions, treatment of direct traffic, eligible views, and the specific scope of each measured action also affect the outcome.<\/p>\n<p>Measurement often has <strong>incomplete coverage<\/strong> because of consent restrictions, blockers, device changes, offline interactions, and data retained inside closed platforms. Two systems may reconstruct different paths and produce different allocations for the same commercial activity.<\/p>\n<h2>Analysis limits<\/h2>\n<p>An attribution model <strong>does not prove<\/strong> causality merely by assigning credit. It describes or estimates how observed contacts relate to a conversion according to the available rules and data.<\/p>\n<p>Data-driven attribution provides a <strong>statistical estimate<\/strong> specific to the analysed action and dataset. Its result depends on data quality, volume, representativeness, and stability, as well as the methodology used by the platform.<\/p>\n<p>Personal recommendations, prior brand awareness, price changes, seasonality, and other <strong>external factors<\/strong> may influence the decision without appearing in the path. Experiments, control groups, and incrementality tests answer causal questions that a descriptive allocation cannot resolve by itself.<\/p>\n<h2>Result comparison<\/h2>\n<p>The selection should match the <strong>analytical decision<\/strong> it is intended to support and the scope of the available data. Interpreting differences between models requires reviewing these elements:<\/p>\n<ul>\n<li><strong>Measured action:<\/strong> the conversion should represent the outcome being valued and remain consistently configured.<\/li>\n<li><strong>Coverage:<\/strong> the compared channels and touchpoints should be included in the data source.<\/li>\n<li><strong>Window:<\/strong> the attribution period should reflect a reasonable duration for the purchase journey.<\/li>\n<li><strong>Comparability:<\/strong> the model, configuration, and period should be documented when interpreting historical changes.<\/li>\n<li><strong>Causal validation:<\/strong> investment decisions may be tested through experiments when measuring genuine incremental impact is necessary.<\/li>\n<\/ul>\n<p>A <strong>controlled comparison<\/strong> holds the action, period, and population constant while changing only the model. This shows which channels gain or lose credit without confusing the effect of allocation with changes in demand, measurement, or budget.<\/p>\n","protected":false},"excerpt":{"rendered":"<p>An attribution model allocates conversion credit. Learn its main types, data scope, limitations, and difference from causal measurement.<\/p>\n","protected":false},"author":7,"featured_media":81099,"template":"","encyclopedia-tag":[1424,1300,1471],"class_list":["post-21233","encyclopedia","type-encyclopedia","status-publish","has-post-thumbnail","hentry","encyclopedia-tag-ad-tracking","encyclopedia-tag-attribution-models","encyclopedia-tag-meta-advertising"],"_links":{"self":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia\/21233","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia"}],"about":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/types\/encyclopedia"}],"author":[{"embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/users\/7"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/media\/81099"}],"wp:attachment":[{"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/media?parent=21233"}],"wp:term":[{"taxonomy":"encyclopedia-tag","embeddable":true,"href":"https:\/\/www.arimetrics.com\/en\/wp-json\/wp\/v2\/encyclopedia-tag?post=21233"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}