What if you could pinpoint your most engaged users, understand what makes them tick, and tailor your strategies to boost conversions and retention? With Adobe Analytics (AA) and Adobe Customer Journey Analytics (CJA), you already have a wealth of data on user behaviour. The real challenge lies in distilling that data into clear, actionable insights.
An Engagement Scoring Model helps you do just that. By assigning scores based on user interactions, you’ll quickly identify your top-performing audience segments, understand which behaviours drive the most value, and prioritize efforts to maximize business outcomes.
In this guide, Jude Felix, Senior Consultant at Accrease will show you how to build an Engagement Scoring Model step by step. You’ll learn how to transform raw analytics data into a powerful tool for targeting, personalization, and optimization, giving you a clearer view of what drives success—and how to replicate it.
Jude Felix, Senior Consultant
What is an Engagement Scoring Model?
An Engagement Scoring Model is a framework that assigns a numerical score to users based on their interactions with your digital properties. By scoring user engagement, you can:
- Identify which user behaviors or interactions are most valuable.
- Segment users based on their engagement levels.
- Make data-driven decisions for targeting, personalization, and optimization.
The process for creating this model combines both gut-based intuition and data-driven analysis to determine which metrics matter most for your business goals. Below are the steps to make the model.
Steps to create the Engagement Scoring Model
Creating an effective engagement scoring model is a powerful way to understand and quantify user behavior. It helps businesses tailor strategies and drive more meaningful customer interactions. Following a structured approach, you can identify key engagement metrics, analyze their impact, and assign appropriate weights to create a robust scoring system.
In this process, we will walk through 7 key steps, starting with defining the data timeframe, selecting relevant metrics, and conducting correlation analysis. We’ll then benchmark metrics, assign weights, and create an engagement score formula. Finally, we validate the model to ensure it accurately reflects user engagement and drives real business results. This systematic approach helps you gauge engagement effectively and provides actionable insights to refine customer experience and strategy for better outcomes.
1. Decide the data time frame
The first step is to decide how much data you want to include in your model. The last 30 days is typically a good starting point, as they capture recent user behaviour without including too much stale data. However, depending on your business or campaign needs, you may choose a different timeframe (e.g., last 7 days or last 90 days).
2. Select key metrics
Selecting the right metrics is the most important step in building the engagement model. There are two main approaches to choosing these metrics:
The intuition-based approach: Based on your experience and knowledge of your customers, you can choose metrics that you believe are important for driving engagement. Some of the metrics might include:
- Page views
- Time spent on site
- Number of sessions
- Product views
- Purchases
- Revenue
The data-driven approach: If you want a more empirical approach, you can identify the metrics most correlated with achieving a specific goal, such as purchasing or signing up for a newsletter. You can do this through statistical correlation analysis, which we will discuss in the next step.
3. Perform Correlation Analysis
To validate and refine the chosen metrics, you can use statistical methods to find out which ones are most closely associated with user success or your defined goals. The goal here is to understand which metrics have the greatest impact on user engagement.
How do you conduct the correlation analysis in Excel?
Gather your data:
- Export relevant data from AA or CJA for your selected metrics over your chosen timeframe (e.g., the last 30 days).
Enable Analysis Toolpak in Excel:
- Go to File > Options > Add-ins.
- In the Manage box, select Excel Add-ins, and click Go.
- Check the box for Analysis Toolpak, and click OK.
Run the correlation analysis:
- Organize your data in columns, with each column representing a metric.
- Navigate to Data > Data Analysis and choose Correlation.
- Select the range of data and choose Output Range to display the correlation matrix.
Analyze the results:
- The correlation matrix will show you the correlation coefficients between each metric and the goal metric (e.g., purchases, signups).
- A correlation coefficient ranges from -1 to 1, with 1 indicating a perfect positive correlation, -1 indicating a perfect negative correlation, and 0 indicating no correlation.
Figure 1 below shows the results of a correlation calculation in Excel based on a few selected metrics.
Figure 1 Correlation calculation in Excel.
4. Benchmark a Score Using the Correlation Coefficients
Once you have your correlation coefficients, the next step is to decide which metrics to include in your engagement model. To do this, you can benchmark a threshold score by calculating the average correlation coefficient across all metrics.
- Metrics with correlation coefficients equal to or above the average should be included in the engagement model.
- Metrics below the average threshold may not significantly contribute to engagement and can be excluded.
Based on the coefficients of correlation from the Figure 1 we can create a benchmark chart like in Figure 2. The metrics number 3, 5, 6, 7 and 8 are he metrics that is highly correlated.
Figure 2 Benchmark chart
5. Assign weights to the metrics
Now that you have selected the most relevant metrics, you need to assign them weights. These weights determine how much influence each metric will have on the overall engagement score.
Normalizing Correlation Coefficients to Assign Weights:
Take the correlation coefficients from the previous step and normalize them so that they add up to 100%. This can be done by dividing each metric’s correlation coefficient by the total sum of all selected metrics’ correlation coefficients and multiplying by 100.
For example, if you have three metrics with correlation coefficients of 0.8, 0.5, and 0.3, you can calculate the normalized weights as follows:
- Total correlation = 0.8 + 0.5 + 0.3 = 1.6
- Normalized weight for metric 1 = (0.8 / 1.6) * 100 = 50%
- Normalized weight for metric 2 = (0.5 / 1.6) * 100 = 31.25%
- Normalized weight for metric 3 = (0.3 / 1.6) * 100 = 18.75%
These normalized values become the weights for each metric, ensuring that the metrics most closely associated with user engagement have the greatest impact on the score.
6. Create the Engagement Score Formula
Now that you have both the metrics and their respective weights, you can calculate the engagement score for each user. The formula will look something like this:
𝐸𝑛𝑔𝑎𝑔𝑒𝑚𝑒𝑛𝑡𝑆𝑐𝑜𝑟𝑒 = (𝑀𝑒𝑡𝑟𝑖𝑐𝑠1 × 𝑛𝑜𝑟𝑚𝑎𝑙𝑖𝑠𝑒𝑑𝑊𝑒𝑖𝑔ℎ𝑡1) + (𝑀𝑒𝑡𝑟𝑖𝑐𝑠2 × 𝑛𝑜𝑟𝑚𝑎𝑙𝑖𝑠𝑒𝑑𝑊𝑒𝑖𝑔ℎ𝑡2) + (𝑀𝑒𝑡𝑟𝑖𝑐𝑠3 × 𝑛𝑜𝑟𝑚𝑎𝑙𝑖𝑠𝑒𝑑𝑊𝑒𝑖𝑔ℎ𝑡3) + …
For example, if you score based on Page Views, Time Spent on Site, and Purchases, figure 3 shows how this can be created in AA or CJA.
Figure 3 Calculation of engagement score
7. Validate the Engagement Model
After building the engagement scoring model, it’s crucial to validate it by applying the model to a sample of users and comparing the scores with actual business outcomes. As shown in figure 3, we can Do higher engagement scores correlate with higher conversions or more purchases?
If the scores don’t align with real-world outcomes, you may need to revisit the metrics or re-run the correlation analysis to fine-tune the model.
If the model accurately reflects user engagement and correlates with success, it’s ready to be deployed.
Why use an Engagement Scoring Model?
An engagement scoring model provides businesses a tangible way to measure and track user engagement over time. Here’s why it matters:
Identify high-value users: By scoring users, you can quickly identify your most engaged users, allowing you to focus your marketing efforts where they’ll have the greatest impact.
Target personalization: Use engagement scores to deliver personalized experiences based on the level of user engagement, creating a more tailored and relevant experience for your audience.
Optimize user journeys: By understanding which behaviors drive engagement, you can optimize your website, app, or marketing efforts to encourage those behaviours, improving overall customer satisfaction and conversion rates.
Conclusion
Building an engagement scoring model using Adobe Analytics or Customer Journey Analytics data is a powerful way to turn raw user interaction data into actionable insights. By following a data-driven approach backed by statistical analysis, you can create a model that accurately reflects user engagement and drives better business outcomes.
This method combines both intuition and rigorous analysis, ensuring that the metrics you choose are grounded in actual data and that the engagement scores you generate align with your business goals. Whether you’re focusing on retention, acquisition, or conversion optimization, an engagement scoring model provides the foundation for smarter, more effective marketing and product strategies.
Start building your model today to maximize your Adobe Analytics data.
Want to dive deeper? Join Jude Felix for a 20-minute Mini Masterclass: 'How to Activate Engagement Scoring to Drive Business Impact.' In this session, Jude will walk you through building and activating engagement scoring models using Adobe Analytics and Customer Journey Analytics (CJA). Don't miss out, sign up here: Mini Masterclass