What is a Customer Engagement Score? Formula and Examples

Reading Time: 16 minutes

A customer engagement score is a single number that shows how actively a customer interacts with your brand. You calculate it by giving each customer action a weight and adding up how often it happens. If you’re running campaigns across email, push, SMS, and in-app, it turns scattered behavior into one number you can sort, segment, and act on.

Think of it like a video game. Every action in the game earns points, and beating a boss earns more than opening a menu. Your customer engagement score works the same way: points for what customers do, with bigger points for the actions that matter most to your business.

By the end of this blog post, you’ll know what a customer engagement score is, why you should use one, the formula, how to calculate it step by step, and a full worked example.

 

Key Takeaways

• A customer engagement score combines multiple customer behaviors into one number, so your team can prioritize audiences more effectively.
• The best scoring models give more weight to meaningful actions, account for recency, and are validated against outcomes such as repeat purchases, retention, and conversions.
• You can use customer engagement scores to personalize campaigns, identify disengaged customers, improve loyalty journeys, and decide when not to send a message.

 

 

What is a Customer Engagement Score?

A customer engagement score is a composite metric that measures how actively and meaningfully a customer interacts with your brand during a specific period.

The score combines different behavioral signals, assigns a value to each one, and produces a single number. Depending on your brand, those behavioral signals could include purchases, website visits, email engagement, app activity, product views, and customer service interactions.

In simple terms, a customer engagement score helps you answer this question: “How engaged is this customer with our brand right now?”

For example, a customer who purchased recently, clicked several campaigns, used a loyalty reward, and visited your app may have a high score. Another customer who hasn’t purchased recently and has stopped interacting with your messages may have a low customer engagement score.

The score can help you see these differences quickly. More importantly, it can help you decide what to do next.

Customer Engagement Score vs. Customer Engagement Metrics

Customer engagement metrics measure one type of behavior. A customer engagement score combines multiple metrics into one view.

This distinction is important. Your score won’t replace your individual metrics. Instead, it’ll help your team prioritize which customers or audiences deserve attention first.

For example, a customer may have a high number of website visits but no purchases. That customer may be interested, but they may also be struggling to find the right product or offer. Looking at the individual metrics behind the score will help you understand what’s really happening.

The table below outlines the differences between a customer engagement score and engagement metrics.

	Customer engagement metrics	Customer engagement score
What it is	Individual measurements (NPS, CSAT, open rate, retention rate)	One combined number per customer
Level	Often reported across a segment or whole program	Calculated for every individual customer
Job	Tells you what happened	Tells you who to act on and how
Example	“Email open rate was 13.2%.”	“Customer 4821 scores 73.”

Now that we’ve defined what the score is, let’s look at the types of customer behavior that can contribute to it.

 

What Does a Customer Engagement Score Measure?

A customer engagement score can measure customers’ interactions across marketing, commerce, loyalty, digital, and customer service touchpoints.

Engagement area	Example signals
Purchasing	Completed orders, repeat purchases, subscription renewals
Website activity	Product views, category visits, search activity
Email engagement	Clicks, conversions, replies, preference updates
SMS engagement	Link clicks, replies, coupon redemptions
Mobile app activity	App sessions, feature usage, push notification clicks
Shopping intent	Product saves, wish-list additions, cart additions
Loyalty	Points earned, rewards redeemed, tier progress
Content engagement	Video views, guides read, quiz completion
Customer service	Help-center visits, chat conversations, service requests
Advocacy	Reviews, referrals, social shares, user-generated content

You don’t need to include every signal in your model. In fact, you probably shouldn’t.

A good customer engagement score focuses on behaviors that indicate genuine interest, intent, or value. A completed purchase will usually matter more than a homepage visit. A product review may matter more than an email open. A loyalty reward redemption may be more meaningful than a push notification impression.

The signals you choose should also match your business model.

An Ecommerce brand may focus on product views, cart activity, purchases, and repeat orders. Or a streaming service may focus on content consumption, app visits, and subscription activity. On the other hand, a food delivery brand may prioritize order frequency, app engagement, and offer redemption.

But why should you even use the score in the first place?

 

Why Use a Customer Engagement Score?

Your team probably already has access to plenty of customer data. The problem is that the data can be difficult to prioritize.

A customer engagement score creates a single, behavior-based signal that your marketing team can use for segmentation and personalization, and lifecycle campaigns.

1. Identify Customers Who Are Becoming Inactive

A declining engagement score can signal that a customer is losing interest before they become completely inactive.

For example, imagine that a customer used to purchase every month and regularly clicked your Shopify automated emails. Over the last 30 days, they haven’t purchased, their website visits have dropped, and they haven’t clicked any messages.

Their engagement score may reveal that decline earlier than a traditional lapsed-customer segment would.

You could then test a relevant win-back campaign, replenishment reminder, personalized recommendation, or educational message.

The change in score may be more useful than the score itself. A customer with a score of 60 that’s increasing may be healthier than a customer with a score of 80 that’s dropped from 95.

2. Find Your Most Engaged Customers

Customers with consistently high engagement scores may be more receptive to new product launches, early-access offers, loyalty benefits, referral campaigns, product review requests, cross-sell recommendations, and VIP experiences.

High engagement doesn’t guarantee that a customer will purchase. It simply tells you that they’re actively interacting with your brand and may respond well to relevant messaging and personalized customer experiences.

3. Improve Audience Segmentation

A customer engagement score can add behavioral context to your existing segments.

For example, instead of creating one broad segment for “high-value customers,” you could identify:

• High-value customers with high and declining engagement
• New customers showing strong early engagement
• Recent purchasers with low post-purchase activity
• Lapsed customers showing renewed interest
• Customers who browse frequently but haven’t purchased

These customers may have similar demographic profiles or purchase values, but they’ll likely need different marketing experiences.

4. Personalize Customer Journeys

Your customer engagement score can help determine what message, channel, or offer a customer receives.

A highly engaged customer may be ready for early access to a new collection. A customer with a declining score may need product education or a win-back message. A customer who recently purchased may be better suited to a post-purchase customer experience than another promotional offer.

The score should support personalization, not replace it. You’ll still need to consider customer preferences, consent, purchase history, product interest, and channel engagement.

5. Measure Campaign Impact

You can compare customer engagement scores before and after a campaign to understand whether customer behavior changed.

For example, you could measure whether engagement scores increased after a welcome journey, an onboarding campaign, a personalized product recommendation, a loyalty promotion, a win-back campaign, or a Shopify post-purchase email series.

A score increase doesn’t automatically prove that your campaign caused the change. If you can, use a control group or holdout audience to measure incremental impact.

So, you know what the score measures and why it matters. Now, let’s look at how to build one.

 

How to Calculate Customer Engagement Score

Calculating a customer engagement score involves more than assigning random points to customer activity. You’ll need to define what engagement means for your brand, identify valuable behaviors, assign weights, set a time period, and validate the model against real customer outcomes.

Here’s a practical process you can follow.

1. Define the Purpose of the Customer Engagement Score

Start with the decision the score will support.

A score designed to identify customers at risk of becoming inactive may use recent purchases, website visits, email clicks, and app sessions. A score designed to identify loyalty opportunities may give more weight to repeat purchases, referrals, reviews, and reward activity.

One score can support several use cases, but defining a primary purpose makes the model easier to design and evaluate.

2. Identify Meaningful Customer Behaviors

Choose behaviors that indicate genuine interest, intent, or value.

For your B2C brand, these might include completing a purchase, making a repeat purchase, clicking an email or SMS link, visiting a product page, adding an item to a cart or wish list, using a mobile app, redeeming loyalty points, reviewing a product, referring a friend, contacting customer service, or renewing a subscription.

Don’t add every event just because you can track it. Too many weak signals can make your score difficult to understand and may reward activity that doesn’t lead to meaningful outcomes.

For example, an email delivery usually shouldn’t increase a customer’s score because it only shows that a message was sent. An email open may be a weaker signal than a click because privacy features and automated image loading can affect open tracking.

3. Choose a Measurement Period

Decide how long the scoring window should be, such as the previous 7, 30, 60, or 90 days.

The right period depends on how often customers normally interact with your brand.

A daily-use mobile app may need a seven-day score. A grocery or beauty brand may use a 30-day period. A furniture or luxury brand may need a 90-day window because customers don’t purchase as frequently.

You can also use multiple time periods. A 30-day score can show current engagement, while a 90-day score can provide broader context.

4. Assign Weights to Customer Actions

Assign higher weights to behaviors that represent stronger intent or deeper engagement.

While measuring the score for Ecommerce customer engagement, your brand can assign points to customer actions based on their expected value. The following example uses a low weight for browsing behavior and higher weights for actions that indicate purchase intent, loyalty, or advocacy.

Customer action	Example weight
Website visit	1
Email click	3
Product-page view	3
Wish-list addition	5
Cart addition	8
Loyalty reward redemption	8
Completed purchase	15
Repeat purchase	20
Product review	10
Customer referral	15

These weights are only starting points, by the way. They aren’t industry standards.

If you’re working with an Ecommerce brand, you should adjust the weights based on historical data. For example, if customers who redeem loyalty rewards are more likely to purchase again, reward redemption may deserve a higher weight. If product-page views rarely lead to conversion, you might need to reduce their weight.

Adapting the Weights by Industry

Customer engagement signals vary by industry because customers create value in different ways. The table below shows how B2C brands in fintech, media, travel, and food delivery might classify low-, mid-, and high-value actions when creating a customer engagement score.

Industry	Low-weight actions	Mid-weight actions	High-weight actions
Fintech	App open, balance check	Add a payee, set a savings goal, click on an offer	Complete KYC, first deposit, card activation or transaction
Media and OTT	App open, browse titles	Watch 10+ minutes, add to watchlist	Finish an episode, start a subscription, renew
Travel	Search, view listing	Save a trip, sign up for a price alert	Booking, review, loyalty signup
Food delivery	App open, view restaurant	Add to cart, apply promo	Order, reorder within 14 days, rating

The same action can have different meanings across industries. An app open may be a weak signal for a fintech brand, but a meaningful signal for a daily-use media or food delivery app. Your brand should validate your weights against outcomes such as activation, booking, order frequency, subscription renewal, or repeat usage.

Use historical data to improve your weights over time. If customers who redeem loyalty rewards are more likely to purchase again, that behavior may deserve a higher score. If product-page views don’t correlate with any meaningful outcome, you may want to reduce their weight or remove them.

5. Apply Frequency Caps

Customers can generate a large number of events without becoming proportionally more engaged.

For example, a customer might view the same product 50 times in one week. That doesn’t necessarily mean they’re 10 times more engaged than someone who viewed it 5 times.

Frequency caps can prevent repeated low-value actions from dominating the score.

You might count up to 5 website visits per week, up to 10 product views per week, and up to 3 email clicks per campaign. At the same time, you might count every completed purchase, referral, or loyalty reward redemption.

6. Account for Recency

Recent interactions usually provide a stronger indication of current engagement than older interactions. A recency multiplier gives more importance to recent customer activity than to older activity. This helps the score reflect current engagement rather than treating a customer interaction from several months ago as equally valuable to one that happened yesterday.

You can apply a recency multiplier to your formula in this way:

Time since action	Recency multiplier
0-7 days	1.0
8-30 days	0.7
31-90 days	0.3
More than 90 days	0 or excluded

The multipliers shown here are starting points. Brands with fast purchase cycles may need engagement scores that decay quickly, while brands with longer purchase cycles may need a slower decay rate.

Choose recency windows based on how often customers typically browse, purchase, renew, or use your product.

Once you’ve chosen your signals and assigned weights, you can bring everything together with a formula.

 

Customer Engagement Score Formula

The basic customer engagement score formula is:

Raw customer engagement score=(w₁​xn_1​)+(w₂​xn₂​)+…+(w_i​xn_i​)

Here’s what each variable means:

• w_i​ is the weight assigned to a customer action.
• n_i is the number of times the customer completed that action during the measurement period.
• i is the total number of tracked actions.

If you include recency, the formula becomes:

Raw score=∑(​​w_i​xn_ixr_i​)

In this formula, r_i​ is the recency multiplier for the action.

To convert the raw score to a 0-100 scale, use:

Customer engagement score = Customer’s raw score/Maximum expected score x 100

As in, divide the customer’s raw score by your maximum expected score, and multiply the result by 100 to get the final customer engagement score.

You can cap the final result at 100 if the customer exceeds your maximum expected score. The maximum expected score should be based on reasonable customer behavior during your selected period. Don’t base it on unlimited activity, or repeated low-value actions could distort the scale.

Let’s look at a customer engagement score example next.

 

Customer Engagement Score Example (Worked, Step by Step)

Suppose an Ecommerce brand calculates its score over 30 days. The scoring model includes the following actions. Each customer action is multiplied by its assigned weight, and the weighted values are then added together to find the raw score.

Customer action	Activity	Weight	Weighted value
Website/app sessions	4	1	4
Email clicks	3	2	6
Product-page views	5	0.5	2.5
Cart additions	1	5	5
Loyalty reward redemption	1	8	8
Completed purchase	1	20	20
Product review	1	10	10
Raw score			55.5

In this example, the customer receives a raw score of 55.5. The completed purchase contributes the most points because it represents a stronger commercial signal than a website visit or product-page view. The score can then be normalized against the brand’s maximum expected score to create a 0-100 customer engagement score.

If the maximum expected score for the 30-day period is 70, the calculation looks like this:

55.5/70×100=79.28

So, the customer engagement score is 79.28 out of 100.

At first glance, the score suggests meaningful engagement. But you’ll want to look at the behavior behind it.

The customer has purchased, clicked emails, and left a review. They may be a good candidate for a relevant cross-sell or loyalty message.

However, you probably shouldn’t send them another discount immediately after their purchase. A complementary product recommendation, care guide, replenishment reminder, or loyalty message may be more relevant.

For our next example, let’s compare two retail brand customers using the same scoring model as above and a 30-day measurement period.

Customer action	Weight	Melanie (count)	Melanie (weighted values)	Sam (count)	Sam (weighted values)
Website/app session	1	10	10	2	2
Email clicks	2	3	6	0	0
Product-page views	0.5	24	12	4	2
Cart additions	5	2	10	0	0
Completed purchase	20	1	20	0	0
Raw score			58		4

Melanie’s raw score is 58 because she visited the website or app, clicked emails, viewed products, added products to her cart, and completed a purchase. Sam’s raw score is only 4 because his activity is limited to two website or app sessions and 4 product-page views.

Say the brand’s maximum expected score is 80.

• Melanie: 58/80 x 100 = 73 (highly engaged)
• Sam: 4/80 x 100 = 5 (at risk)

This comparison shows why a weighted score is more useful than simply counting interactions. Melanie’s actions indicate stronger purchase intent, while Sam’s activity suggests that he may need a re-engagement or product-discovery journey.

What to Do With Each Customer Engagement Score

Customer engagement tiers translate numerical scores into practical marketing actions. The following ranges are an example of how a B2C brand might classify customers as at risk, moderately engaged, or highly engaged on a 0-100 scale.

Tier	Score	What it means	Suggested action
At risk	0-40	Low activity, likely to lapse	Win-back journey, onboarding nudge, incentive on their preferred channel
Moderately engaged	41-70	Steady but with room to grow	Cross-sell recommendations, feedback request, loyalty prompt
Highly engaged	71-100	Frequent, deep interaction	Early access, referral ask, upsell, brand advocacy

Based on the above table, Sam (with a score of 5) goes into a win-back journey, while Melanie (with a score of 73) gets a referral invite. Neither gets the generic blast.

 

How to Measure Customer Engagement Scores

Calculating the customer engagement score is only the first step. To measure it effectively, connect the score to customer behavior and marketing outcomes.

1. Monitor Score Changes Over Time

Monitor whether each customer’s engagement score is increasing, stable, or declining.

A customer whose score is rising may be responding well to onboarding or lifecycle messaging. A customer whose score is declining may need a different experience.

You’ll still need to consider the customer’s context. A customer who recently purchased may naturally have fewer pre-purchase interactions. A customer who hasn’t purchased in a year may need a more direct win-back strategy.

2. Compare Scores With Customer Outcomes

Test whether higher scores are associated with repeat purchases, subscription renewals, increased purchase frequency, higher customer lifetime value, loyalty participation, referral activity, reduced inactivity, or stronger campaign conversion.

If customers with high scores don’t behave differently from customers with low scores, revise the scoring model.

3. Create Category and Lifecycle Benchmarks

Customer engagement varies by product category and lifecycle stage. A new customer, a repeat buyer, and a lapsed customer should not necessarily be judged by the same benchmark.

You might create separate benchmarks for new customers, active repeat purchasers, subscribers, loyalty members, high-value customers, lapsed customers, and seasonal shoppers.

For example, a customer who purchases infrequently but makes high-value purchases may have a lower activity score than a frequent shopper.

That doesn’t necessarily mean they are less valuable or less engaged.

4. Use Test and Control Groups

When activating the score in campaigns, test your strategy against a holdout group where possible.

For example, compare customers who receive a win-back journey with similar customers who do not. Measure incremental purchases, revenue, unsubscribes, and long-term customer retention in addition to engagement score changes.

This will help you determine whether your scoring model is improving campaign decisions or simply describing customers who were already likely to engage.

After you’ve calculated your score and compared it with real customer outcomes, you can start creating practical engagement tiers. These ranges can help your marketing team decide which audiences need attention and what kind of action makes sense.

Once your scoring model is defined, your data team will need a practical way to calculate it at scale.

 

For Your Data Team: Spreadsheet and SQL Versions

Your analysts can build a view of your customer engagement scores in a spreadsheet or a data warehouse. A good way to go about it is to start with a spreadsheet while testing the model, then move the calculation into your data warehouse or customer engagement platform as your use case grows.

Spreadsheet Version

In a spreadsheet, put weights in one row and counts in another. Use =SUMPRODUCT(weights, counts) for the raw score and =MIN(100, raw/maximum expected score*100) to scale it.

SQL Version

In SQL, the same logic looks like this:

SELECT   customer_id,   LEAST(100, 100.0 * SUM(     CASE event_name       WHEN ‘session’ THEN 1       WHEN ‘product_view’ THEN 0.5       WHEN ‘message_click’ THEN 2       WHEN ‘add_to_cart’ THEN 5       WHEN ‘purchase’ THEN 20       ELSE 0     END   ) / 80) AS engagement_score FROM events WHERE event_time >= CURRENT_DATE – INTERVAL ’30 days’ GROUP BY customer_id;

A score sitting in a dashboard won’t improve customer engagement on its own. Its real value comes from how quickly your team can turn the score into a relevant customer experience across channels.

 

How B2C Marketing Teams Can Use Customer Engagement Scores

Here’s how your team can use a customer engagement score in your day-to-day marketing work:

• Build More Relevant Win-Back Campaigns: Use declining scores to identify customers who may be losing interest. Combine the score with purchase recency and expected purchase frequency before adding customers to a win-back journey. A customer who normally purchases every 30 days may need a different message from someone who purchases only twice a year.
• Improve Post-Purchase Journeys: A purchase may increase a customer’s score, but post-purchase engagement still matters. You can use the score to identify customers who may benefit from product education, setup instructions, care advice, complementary recommendations, review requests, or loyalty invitations.
• Prioritize Loyalty Experiences: Highly engaged loyalty members may be good candidates for early access, exclusive content, bonus-point events, or referral campaigns. However, high engagement doesn’t mean that customers want unlimited communication. You’ll still need to follow their channel preferences and frequency limits.
• Choose the Right Marketing Channel: Your score can be combined with channel-level engagement to help you choose where to communicate. A customer may have a high overall score but interact mostly through your mobile app rather than email. If they’ve opted in, a push notification or in-app message may be more effective than another email.
• Suppress Irrelevant Promotions: Engagement data can also help you decide when not to send a message. A customer who recently purchased may not need another discount immediately. Combining engagement data with purchase recency can help you reduce irrelevant promotions and message fatigue.

Customer behavior can be a powerful input for personalization, but it also needs to be handled responsibly.

 

Customer Engagement Score and Privacy Considerations

B2C marketing teams need to build customer engagement scores around appropriately governed customer data. You need to take the following things into account:

• Privacy: Before you use customer behavior in a scoring model or activate the score in a campaign, review your consent and communication permissions, state and national privacy requirements (such as the new TCPA rules and the California Consumer Privacy Act or CCPA), opt-out requests, sensitive data restrictions, retention policies, identity resolution practices, and vendor agreements.
• Apple Mail Privacy Protection: It inflates email open counts, so weight clicks and on-site or in-app actions above opens, or leave opens out of the score.
• Seasonality: For peaks like Black Friday or Cyber Monday, use a shorter recency window (say, 7 to 14 days) or score peak periods separately. That way, a holiday spike won’t make every shopper look highly engaged.

Your scoring model should generally rely on relevant first-party behavioral data rather than sensitive personal characteristics. Don’t use protected or sensitive information to unfairly determine which customers receive offers, services, or different treatment.

You’ll also need to respect a customer’s channel preference or opt-out request, even if their score is high. A high customer engagement score isn’t permission to contact someone through a channel they haven’t authorized.

But before you launch yours, take a moment to review the mistakes that can make customer engagement scores inaccurate, difficult to interpret, or risky to use.

 

Common Customer Engagement Score Mistakes to Avoid

Here are some mistakes to steer clear of before measuring your customer engagement scores:

• Counting every interaction equally: A product view, email click, and completed purchase represent different levels of intent. Giving each event the same weight can make the score less useful.
• Relying heavily on email opens: Email opens are an imperfect signal because privacy protections and technical changes can affect open tracking. Clicks, conversions, replies, purchases, and preference actions are generally stronger indicators of active engagement.
• Confusing engagement with customer value: A highly engaged customer is not always the highest-value customer. One customer may interact frequently but purchase inexpensive products, while another may purchase only a few times a year but have a high order value. Engagement score should be analyzed alongside revenue, margin, purchase frequency, retention, and customer lifetime value.
• Using a score without recency: A customer who engaged 6 months ago should not necessarily receive the same score as a customer who engaged yesterday. Include a time window or recency adjustment.
• Allowing event volume to distort the score: Repeated low-value actions can overwhelm more meaningful behaviors. Frequency caps and normalization help keep the score balanced.
• Applying a benchmark to every customer: Purchase cycles and engagement patterns vary across categories, customer groups, and lifecycle stages. Use benchmarks that reflect the context in which the customer operates.
• Sending more messages to highly engaged customers: High engagement does not mean that a customer wants unlimited communication. Use channel preferences, frequency controls, and campaign suppression rules to manage message fatigue.
• Treating the score as a prediction without validation: A customer engagement score may appear logical but fail to predict future behavior. Validate it against repeat purchases, renewals, inactivity, and other outcomes before using it to automate important decisions.

You now have the framework to define, calculate, validate, and activate a customer engagement score. As a final step, choosing a customer engagement platform can help your team turn the score into action.

Final Thoughts: Turn Your Customer Engagement Score Into Action With MoEngage

A customer engagement score can help your B2C marketing team turn fragmented customer activity into a practical signal.

The best scoring models don’t simply reward every click, open, or visit. They focus on meaningful customer behaviors, account for recency, cap repeated activity, and connect the score to real business outcomes.

But the real value comes when you validate the score against outcomes and connect it to campaigns. That’s where a customer engagement platform comes in.

With MoEngage’s agentic customer engagement platform, you can bring customer data, real-time behavioral insights, segmentation, and cross-channel engagement together in one place. Your team can identify changing customer intent, build relevant audiences, and activate personalized journeys without waiting on a separate team for every campaign update.

Want to make your customer engagement strategy more data-driven? Schedule a demo to see how MoEngage can help you turn customer behavior into better marketing decisions.