Omnichannel attribution: real-world application

Let's be precise about what attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → actually means in practice. AttributionAttributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → is the process of assigning credit to marketing touchpoints in a customer journeycustomer journeyThe full sequence of touchpoints a customer has with your brand before, during and after purchase, spanning awareness, consideration, decision, retention and advocacy.View full definition →. A touchpoint is any interaction a customer has with your brand: a Google search ad, an Instagram story, a retargetingretargetingShowing ads to users who have previously visited your site or interacted with your brand, to bring them back and drive conversion.View full definition → banner, a YouTube pre-roll, an email, a podcast ad. The customer journeycustomer journeyThe full sequence of touchpoints a customer has with your brand before, during and after purchase, spanning awareness, consideration, decision, retention and advocacy.View full definition → is the full sequence of those interactions before a conversion happens. AttributionAttributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → models are the rules you use to split credit across those touchpoints. Last-touch attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → gives 100% of the credit to the final touchpoint. First-touch gives it all to the first. Linear splits it equally. Time-decay gives more credit to touchpoints closer to conversion. Data-drivenData-drivenAn approach where decisions are systematically informed by data analysis rather than intuition alone.View full definition → attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → uses machine learning to assign credit based on actual conversion patterns in your data.

The real-world challenge is that most companies are running 6 to 12 active channels simultaneously, customers are switching between devices, and the average B2C purchase journey involves 6 to 8 touchpoints over days or weeks. Your CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition → sees some of that. Your ad platforms see some of it. No single tool sees all of it. This is the core problem you are solving.

Four concepts drive real-world attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → decisions at the CMO level.

First: model selection is a strategic choice, not a technical one. The model you choose shapes which channels get budget. If you use last-touch, paid search and retargetingretargetingShowing ads to users who have previously visited your site or interacted with your brand, to bring them back and drive conversion.View full definition → always win because they sit closest to conversion. That is not because they are doing the most work, it is because of where they live in the funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition →. Awareness channels like display, podcast, and social will look useless under last-touch even if they are driving the initial demand that feeds your entire funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition →.

Second: incrementality testing is the only way to know what is actually causal versus correlational. Incrementality means: would this sale have happened anyway without this ad? Meta and Google both run geo-based lift tests. You split your markets into control groups that do not see certain ads and test groups that do, then measure the difference in conversions. This is not theoretical. It is the method Procter and Gamble used in 2021 when they discovered that a significant portion of their digital retargetingretargetingShowing ads to users who have previously visited your site or interacted with your brand, to bring them back and drive conversion.View full definition → spend was capturing customers who would have converted organically without the retargetingretargetingShowing ads to users who have previously visited your site or interacted with your brand, to bring them back and drive conversion.View full definition → ad. They shifted hundreds of millions of dollars as a result.

Third: cross-device identity resolution is non-negotiable for accuracy. A customer who sees your LinkedIn ad on a work laptop, clicks a retargetingretargetingShowing ads to users who have previously visited your site or interacted with your brand, to bring them back and drive conversion.View full definition → ad on their phone, and converts on a home computer looks like three separate users in your analytics. Without identity resolution, tying those touchpoints to a single person, you are counting phantom journeys. LiveRamp, Segment, and Salesforce CDPCDPA Customer Data Platform unifies customer data from all sources into persistent, actionable profiles that other systems can use.View full definition → all offer identity graph capabilities that stitch touchpoints across devices using deterministic signals like logged-in email addresses and probabilistic signals like IP and browser fingerprinting.

Fourth: attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → windows define what counts as a touchpoint. A 7-day click window means any ad click within 7 days of a purchase gets credited. A 1-day view window means any ad impressionad impressionThe total number of times an ad or piece of content is displayed, regardless of clicks. Each display counts as one impression, even to the same person.View full definition → viewed within 1 day gets credited. When Meta moved to a 7-day click, 1-day view default in 2021, reported conversions dropped across the industry overnight for advertisers still using the 28-day window. Nothing about actual performance changed. The window changed. Your window choice must match your actual sales cycle length.

Let's look at how three real companies have applied this.

Nordstrom rebuilt their attribution modelattribution modelA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → in 2019 after discovering that email was being credited with almost 40% of online conversions under last-touch. When they ran a holdout test, suppressing email to a control group, conversions in that group dropped by only 11%. That gap told them email was claiming credit for organic intent, not generating it. They shifted attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → to a time-decay model with incrementality holdouts, reallocated 22% of their email budget toward paid social prospecting, and saw a 14% lift in new customer acquisition within two quarters.

Rakuten Advertising published results from a multi-touch attributionmulti-touch attributionA method that distributes conversion credit across all marketing touchpoints in the customer journey, rather than crediting only the first or last interaction.View full definition → study with a major retail client showing that affiliate and influencer channels, which consistently appeared useless under last-touch, were actually initiating 31% of all journeys that eventually converted through paid search. The CMO had been considering cutting the affiliate program entirely. The attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → rebuild saved a channel that was generating millions in assisted revenue invisible to the old model.

HelloFresh, the meal kit company, uses a combination of marketing mix modeling (MMM) for strategic budget planning at the channel level and multi-touch attributionmulti-touch attributionA method that distributes conversion credit across all marketing touchpoints in the customer journey, rather than crediting only the first or last interaction.View full definition → for tactical optimization within digital channels. CMO Deise Palmeira has spoken publicly about using MMM to understand the contribution of TV and podcast advertising, which are impossible to track at the user level, while using MTAMTAA method that distributes conversion credit across all marketing touchpoints in the customer journey, rather than crediting only the first or last interaction.View full definition → for real-time bid optimization on paid search and social. This two-layer approach is now standard practice at sophisticated DTC brands running above $200M in revenue.

As a CMO, here is what you actually do with this knowledge:

• Audit your current attribution modelattribution modelA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → within 30 days. Ask your analytics team what model is being used in every platform: Google Ads, Meta, your BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.View full definition → tool, your CRMCRMCustomer Relationship Management: software and strategy to manage and analyse customer interactions throughout their lifecycle.View full definition →. They are almost certainly different, which means you have no single source of truth. Standardize on one model for cross-channel comparison, even if it is imperfect.
• Run at least one incrementality test per quarter on your highest-spend channel. This is the only way to know if the attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → credit a channel claims is real. Start with your retargetingretargetingShowing ads to users who have previously visited your site or interacted with your brand, to bring them back and drive conversion.View full definition → spend, which is almost always inflated under any click-based model.
• Invest in a CDPCDPA Customer Data Platform unifies customer data from all sources into persistent, actionable profiles that other systems can use.View full definition → or identity resolution layer before you invest in more ad spend. Without it, your attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → data is built on fragmented, device-specific journeys that do not reflect how real customers actually move through your funnelfunnelThe customer journey from awareness to purchase, typically Awareness, Interest, Consideration, Decision, Action, with prospects narrowing at each stage.View full definition →.

Common mistakes that kill results:

Using platform-reported attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → as your source of truth. Google credits Google. Meta credits Meta. Both use different windows and different counting methods. A customer who clicked a Google ad and a Meta ad before buying will appear as a full conversion in both platforms. Your actual revenue is being double or triple-counted across platforms, making total reported ROASROASReturn on Ad Spend (ROAS) measures the revenue generated for every unit of currency spent on advertising, calculated as revenue divided by ad cost.View full definition → look far better than it actually is.

Skipping the holdout test because it feels too risky. The most common objection is that suppressing ads to a control group will lose sales. That fear is exactly why most companies never discover how much of their spend is wasted. The holdout group size should be small, 5 to 10% of your audience, and the test duration should match your sales cycle. The cost of running the test is trivially small compared to the budget decisions it informs.

Treating attributionattributionA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → as a one-time project. Customer behavior shifts, channel mix changes, and new touchpoints are added constantly. An attribution modelattribution modelA framework for assigning credit to the touchpoints that contributed to a conversion, so you can measure which channels and interactions actually drive results.View full definition → built in 2022 does not reflect how your customers buy in 2025. Quarterly model reviews with your data and media teams are not optional at the CMO level, they are a core operating rhythm.