Member Scoring and Intelligence (MSI)

note

This feature may not be available in your version of Console depending on configuration and permissions.

Member Scoring and Intelligence (MSI) is an embedded analytics agent in ES Console that guides users through the creation and application of member scores, including RFM, Profitability, Engagement, and predictive scores such as Churn and Propensity to Buy.

Delivered through a conversational chat interface, MSI reduces the need for analyst support while enabling smarter targeting and loyalty decisions directly within your loyalty operations platform.

Key benefits include:

1. Elevating member engagement: Instantly create member scores to fuel personalization.
2. Reducing analyst costs: Eliminate the need for costly models and consultants.
3. Creating insights in seconds: No more analyst queues.
4. Empowering marketers: Democratize loyalty analytics.
5. Staying ahead of the competition: Make smarter loyalty decisions guided by AI.
6. Turning scores into sales: Activate scores across campaigns instantly.
7. Increasing retention: Identify and take measures to retain members in the loyalty program.

Types of Member Scores

There are two types of member scores:

1. Descriptive scores
2. Predictive scores

The last three in the list below (Churn, Propensity, and CLV) are predictive scores, which are aimed at forecasting future member behavior.

Scores are built from user-defined parameters and transaction history (descriptive) or by training models using historical data (predictive):

1. Recency: How recently members have engaged with the loyalty program.
2. Frequency: How often members interact with the loyalty program.
3. Monetary Value: The value the member creates through their spending habits.
4. Profitability: How much profit the member generates based on a revenue-to-cost analysis.
5. Offer Engagement: How well the member responds to promotional efforts.
6. Churn: Divides members into groups based on how likely they are to churn or no longer engage with the loyalty program.
7. Propensity to Buy: Identifies customers most likely to purchase, reducing wasted spend and maximizing conversion lift.
8. Customer Lifetime Value: Represents the predicted total revenue or value a member will generate throughout their entire relationship with the program. Used to identify high-value members and guide strategic decisions about resource allocation, retention efforts, and personalized engagement strategies.

Each score can include nuanced scoring segments that can be used in your marketing campaigns.

Predictive scores such as Churn, Propensity to Buy, and Customer Lifetime Value use machine learning models trained on user-defined parameters to forecast member behaviors. Generating predictive scores takes time to build the model, so you can leave the page while processing completes. A message appears in the Console with a link you can click to return to the completed predictive score.

Navigate to Member Scoring and Intelligence

To open the Member Scoring and Intelligence landing page (if selecting a business unit isn't required):

Select Members > Scoring and Intelligence.

In some cases, a member score is specific to a particular business unit (such as a division or subsidiary) of a larger business entity. If so, you must first select the business unit to access the correct MSI page. Business units aren't shown if your company doesn't use them.

To select a business unit, Select Members > Scoring and Intelligence > Business Unit - where Business Unit is replaced by the name of the relevant subsidiary or division.

In either case, the Member Scoring & Intelligence landing page opens. This page includes controls that allow you to:

• View a list of all available scores (shown across multiple pages if needed).
• Search for an existing score by member scoring name, or clear a search result to view all scores again.
• Create a score or learn more about scores by clicking the +Score button and interacting with the chatbot.
• Navigate to additional pages or select a different number of scores to show on each page using the controls at the bottom of the page.

Search for an Existing Score

To find an existing score, either scroll through the listed scores or use the search function on the Member Scoring & Intelligence page to search by name. The following steps show how to search for an existing score:

1. Enter the relevant information in the search field.
2. Click Search. All existing scores matching the search are displayed. If there are no matches, a message reads: "There are no search results that match your criteria."
3. To clear the current search term, click Clear Search. The page displays all scores again.

View, Edit, or Refresh a Score

On the Member Scoring & Intelligence page, all existing scores are listed (across additional pages if needed). You can scan the scores on the page and open an existing score to view or edit its settings. Note that not all fields can be edited — only those that aren't greyed out. The following steps show how to view existing scores:

1. Scroll down the scores list on the Member Scoring & Intelligence page. Use the page numbers or Previous/Next controls at the bottom of the page to view additional pages.
2. To show more or fewer scores on the page, select a different setting in the Show list.

Open a score for viewing or editing

1. Find the score you want to view or edit in the list, or use search to find it.

2. Click the name, or select View from the Action menu for the score. The score details open on the page, with the score ID and score name shown at the top.

3. In the Details section, information about the score is shown, including the type (see Types of Member Scores above) and additional information that depends on the type:

   • Base: Appears only for Frequency or Monetary member scores. Indicates whether segmentation is based on lifetime transactions or transactions within a selected timeframe (such as weekly, monthly, or yearly).
   • Distribution: Depends on the segmentation selected; can be quartiles, deciles, or buckets.
   • Buckets: Provides precise information about the parameters for specific buckets. Only appears if custom buckets (Distribution = buckets) are being used for segmentation.
   • AI Description: Summarizes the details of this member score.

4. The iterations or refreshes are shown in a table at the bottom of the page, with the current refresh at the top (labelled Current Execution). In the Manual Score Refresh section, you can click the Manual Refresh button or set up an Automated Refresh schedule.

5. To return to the Member Scoring & Intelligence page, click the browser back button () or the refresh button ().

Refresh a score manually

1. Find the score you want to refresh in the list, or use search to find it.
2. Click the name, or select View from the Action menu. The score details open on the page.
3. Scroll down to the Manage Score Refresh section.
4. In the Manual Refresh subsection, click Manual Refresh. Note that this button isn't enabled if the score status is Disabled or Deleted.
5. Click Accept in the confirmation window. A message confirms that a job has been scheduled to update the score based on current data.

Only the last completed iteration — labelled Current Execution — is used by the score. When the scheduled job completes, it becomes the current iteration.

Refresh a score automatically

1. Find the score you want to configure in the list, or use search to find it.
2. Click the name, or select View from the Action menu. The score details open on the page.
3. Scroll down to the Manage Score Refresh section.
4. In the Automated Refresh subsection, select a regular cadence from the Every dropdown list. Note that these controls aren't enabled if the score status is Disabled or Deleted.
5. To set a weekly refresh, select week, then choose the day and the hour and minute (24-hour clock in 15-minute increments) for the refresh. To set a monthly refresh, select month, then choose the date, hour, and minute. The refresh runs automatically on the schedule you set.

6. Click Save at the bottom of the page.
7. Click Accept in the confirmation window. A message confirms that a job has been scheduled to update the score based on your settings and the data current at that date and time.

Only the last completed iteration — labelled Current Execution — is used by the score. When the scheduled job completes, it becomes the current iteration.

Create a Score

When creating a score, you have the following options:

1. Go through the creation process from start to finish while accepting the recommendations along the way.
2. Go through the process while making changes — either by selecting options other than the recommendations or by entering different instructions.
3. Query information presented along the way to learn more about calculations or scores.

Using one or more of these options helps ensure that the resulting score meets your requirements.

The following steps show how to create a score:

1. Select Members > Scoring and Intelligence > Business Unit.
2. Click +Score.
3. An initial page showing the types of scores (each as a button) opens. Click the type of score you want, then make additional selections through the interactive interface to refine it. You can also enter questions or instructions along the way. Note: Intensive calculations occur in the back end of the application, so this and subsequent steps may each take a few seconds to resolve. Predictive scores (Churn, Propensity to Buy, and Customer Lifetime Value) may take significant time to train a model. For those scores, you can navigate away from the page and receive a completion message when training is done.
4. Once you've built your score, enter a name (up to 128 characters long) and click Send to save it. The name should include the score type (such as "RECENCY").
5. Return to the Member Scoring & Intelligence page. Your new score is listed on the page and can be managed from there.
6. Before using the score elsewhere, open it and ensure it's refreshed before first use. Click Manually Refresh or set up an Automated Refresh to refresh the score before first use.
7. Log out of the Console and back in to ensure the score is available when you want to use it in Offers, Audience Builder, or Audience Recommender.

Use a Score

To use a member score with an offer, Audience Builder, or Audience Recommender, follow the normal flow for creating an offer or audience. When you reach the targeting step, select one of the Member Scoring & Intelligence options. Each option represents one saved member score.

note

Refresh the member score first if using it with Audience Builder or Audience Recommender. Open the score from the Member Scoring & Intelligence page (select View in the Actions menu or click the name) and refresh it (either manually or via scheduling).

Once processing is complete and the status shows COMPLETED, the score marked Current Execution (at the top of the table) is ready for use. If the score is still being processed, don't use it yet, as not all members have been scored.

For more information, see Creating an audience targeting a member score segment in the Audience Builder topic.

Build Specific Types of Scores

You can build scores that are useful in a variety of ways. This section describes one type of score — the Recency score — as an example. Each type of score has its own process, but the process is self-directing and doesn't require additional description here.

At any point, you can enter a query to take a different path from the available options or to amend results. For example, if you selected deciles and prefer different decile names, you can send a query requesting a name change. You can also change the distribution of members between custom buckets or rename those buckets.

Example of Recency Score

The following steps show how to create a is an example of creating a Recency score example:

1. Select Members > Scoring and Intelligence > Business Unit.
2. Click +Score.
3. An initial page showing the types of scores opens, along with definitions for each type. Click Recency to select a metric that measures the time since the member's last purchase, considering the member's lifetime program purchases to date. Typically, returns aren't included in recency calculations, as the focus is on purchase activity. Note: Intensive calculations occur in the back end of the application, so this and subsequent steps may each take a few seconds to resolve.
4. The analytics agent asks you to select a method of grouping the data. Options include Quartiles (four groups each containing the same number of members), Deciles (ten groups), Percentile (100 groups), or Custom Buckets (defined by you). If you decide you want to create a different type of member score, click New Member Score at this point. In this example, click Deciles to identify top-performing members and potential churners.
5. The analytics agent explains the decile groups in text format and displays a graph of the deciles (calculated on active members). Hover over any element in the graph to see more details.

6. To make changes or ask questions about the results, enter your question or change in the Type your message... field and click Send. The analytics agent responds to your question or applies your change. You can do this multiple times to tailor your results.
7. Once you have the results you want, click Proceed. The agent displays a summary of your built metric. Alternatively, click Change Scoring Approach to change your segment, or New Member Score to start the scoring process again.
8. To view additional details about the score, click Explain Scoring Calculation to see how the scoring calculation was applied, or Provide Marketing Ideas to view suggestions for how the score might be used in marketing. Note: These options aren't available after you save your score.
9. If the summary is correct and you clicked Proceed, the agent prompts you to provide a score name that includes a suggested element (such as "RECENCY") to describe the segment. The score name can be up to 128 characters long.
10. Enter the name including the suggested element and click Send. The score is saved to the Member Scoring & Intelligence page.
11. To return to the Member Scoring & Intelligence page, click the browser back button ().
12. Open the score and refresh it by clicking Manually Refresh or by setting up an Automated Refresh that runs before first use.
13. To ensure the score is available in Offers, Audience Builder, or Audience Recommender, log out of the Console and back in.