Introduction to Attribution AI

A high-level overview of how marketers and analysts can use Attribution AI to understand the impact of their marketing channels and campaigns. For more information, please visit the Attribution AI documentation.

Hi, I’m Desmond Chan, Senior Product Manager. In this video you’ll learn how marketers can leverage Attribution AI to understand the impact of their marketing channels and campaigns. We’ll cover what Attribution AI is the use cases for Attribution AI the different types of Attribution AI scores and the basic architecture. Attribution AI quantifies the marketing impact of each individual marketing touch points across customer journeys at every level of granularity. As a customer interacts with your brand they may encounter many of your marketing and touch points such as Display Ad Impressions, Email sent Email opens, Paid Search Clicks et cetera. Attribution AI determines how credit for conversions is assigned to each touch point using our advanced AI ML algorithm. Marketers can use this insight to optimize marketing budgets and campaign tactics.
Attribution AI can support the following use cases. Executive reporting, allow executives to understand the true incremental impact of marketing both as a whole and by channel region SKU et cetera. Budget allocation, inform budget allocation decisions across marketing channels. Campaign optimization, within each channel understand what campaigns, creatives, keywords et cetera are working better for what SKUs and GOs so that each channel marketing team can optimize their tactics. Full-funnel attribution, understand marketing’s impact across the entire customer journey from free account signup to paid conversion and beyond. Planner evaluations, evaluate effectiveness of marketing agencies or partners based on attribution results. Attribution AIs advanced AI ML based attribution algorithm provides marketers the most accurate attribution results at the individual touch point level. Our machine learning attribution algorithm generates two distinct outputs for every individual touch point on each customer path incremental score and influenced score. Influenced score is the fraction of the conversion that each marketing touch point is responsible for. Incremental score is the amount of marginal impact directly caused by the marketing touch point. The main difference between the Incremental score and the Influenced score is that the Incremental score takes the baseline fact into account and thus does not assume that a conversion is caused purely by the proceeding by marketing touch points. To determine the incremental score Attribution AI first determines the portion of the conversion that is marginal as the result of marketing touch points. The model then algorithmically divides the credits of the marginal portion across the proceeding touch points. On the other hand the influenced score algorithmically divides all the credits of the entire conversion across the proceeding touch points. Let’s look at an example of influenced and incremental scores on a sample conversion path. In this example a customer engaged with Luma an athletic apparel online retailer across multiple touch points before purchasing a watch for $500. For influenced scores, Attribution AI assigns fractional portions of the conversion value of $500 across the touch points using our machine learning based algorithm. This way you know how much of the conversion is attributed to a certain touch point all the fractional portions sum up to the total conversion value of $500. In case of incremental scores attribution AI it measures the marginal increase in conversions solely due to the marketing touch points considered after taking out the baseline effect. Baseline on effect is the portion of conversion attained without any marketing activities for example, brand value or other associated factors. In the case shown, the total incremental impact across the touch point considered is $250 and this impact is distributed across the touch points considered using our advanced algorithm. While we recommend customers to utilize Attribution AIs algorithmic attribution for the most accurate result Attribution AI also provides rule-based attribution models including First touch, Last touch Linear, Time decay and U-shaped. Now let’s take a look at the Attribution AI Workflow and how it integrates with Adobe Experience Cloud and Experience Platform. First, with the help of professional services the data from various sources such as Adobe Analytics is ingested, mapped and transformed into XDM and stitched in experience platform with the appropriate data governance in place. The marketing analyst would now be able to easily configure their desired attribution AI instance for any specific business objective in mind. Then after training and scoring powered by Attribution AI the scores are written back to experience platform and snowflake data warehouse for marketing analyst to operationalize. There are two main ways of operationalizing the attribution AI insights. First, you can consume insights to a dashboard provider in the intelligence services interface and second, you download scores for custom visualization and analysis using third-party tools like Microsoft’s Power BI. Attribution AI predictions can be configured using the Attribution AI, UI or API. Once configuration is done Attribution AI will start the model training process using the customers data as specified in the configuration. After the models are trained, Attribution AI will perform scoring according to the specified cadence. Customers can utilize attribution scores in the following ways. They can consume the Attribution Scores and Insights through Attribution AI insights page as shown here. In addition, customers can also download the Attribution AI output to do custom analysis and dashboarding. That’s a quick introduction to Attribution AI you should now know what Attribution AI is the use cases for Attribution AI the different types of Attribution AI scores and a basic workflow. -