Two different numbers go by the name customer engagement score. One adds up what customers do, with a weight on each action. The other, built by Gallup, asks customers how attached they feel. Both try to spot who will stay and spend more. A loyalty program can run both on the same member.

What is a customer engagement score?

A customer engagement score is one number that sums up how actively a customer deals with a brand. Most scores give each action a weight, then add the weighted actions into a total (Salesforce). Software firms use it to spot trial users ready to buy and customers close to leaving (CFI).

Engagement covers more than buying. Jenny van Doorn and her co-authors define engagement behavior as what a customer does toward a brand beyond a purchase. Their list takes in word of mouth, tips to friends, helping other customers and writing reviews (van Doorn et al., 2010). A loyalty program logs many of these, from referrals to reviews to points spent.

Watch the short form. Teams write both this score and customer effort score as CES, and the two measure very different things.

How do you calculate a customer engagement score?

The general formula is the sum of each event’s weight times the number of times it happened (CFI). Pick the events that matter and give each a weight from 1 to 10. Then count them over a set window. Userpilot adds that recent events often carry more weight, since they say more about what the customer will do next (Userpilot). Scale the total so the top of the range reads 100.

The table shows a starting set of weights for a loyalty program. Set your own from the actions that came before higher spend in your data.

EventWeightWhy it earns that weight
Purchase10The behavior the program exists to grow
Referral that joins8Brings a new member at low cost
Points redeemed7Shows the member values the rewards
Review or survey answer5Adds knowledge the brand can act on
Offer activated3Shows intent to buy soon
App visit or balance check1A light signal, cheap to repeat

For illustration, two members each scored 40 last month. Member A bought four times and nothing else, so 4 times 10 gives 40. Member B bought twice, redeemed once, answered a survey and opened the app eight times. That is 20 plus 7 plus 5 plus 8, also 40. The same score hides two different members. A buys often but ignores the program. B buys less but uses every part of it, so B is the one a purchase offer should go to first.

What is Gallup’s CE11 engagement score?

Gallup took the survey route with CE11, a set of 11 questions. Three cover rational loyalty. They ask how satisfied the customer is and how likely she is to buy again and to recommend the brand. Eight cover the emotional bond. They come in pairs on four themes: confidence, integrity, pride and passion. One reads, “I can’t imagine a world without [brand]” (Gallup, 2001).

Gallup says the pairs build in order. Customers who agreed strongly with the first two statements were more likely to agree with the next two. Gallup also studied a large hotel chain’s affinity club. Strongly attached members gave the chain 32% more of their lodging spend than members with no bond (Gallup, 2001).

Gallup ranks customers from fully engaged at the top to actively disengaged at the bottom. By its count, the top group comes in 23% above the average customer. That holds for share of wallet, profit, revenue and growth. The bottom group comes in 13% below (Gallup, 2006). These are Gallup’s own figures from its client work. They come from before most loyalty apps, so test them on your own members.

What is customer engagement value?

Customer engagement value puts a price on engagement. V. Kumar and his co-authors split it into four parts. Lifetime value counts what the customer buys. Referral value counts new customers she brings in for a reward. Influencer value counts the word of mouth that sways what others buy. Knowledge value counts the feedback that helps the firm improve (Kumar et al., 2010).

Their point: judge customers on purchases alone, and some will look worth less than they are while others look worth more. A later study by Kumar and Anita Pansari drew on data from 120 firms at two points in time. Engaged customers and engaged staff both lifted results, and the customer side had the larger effect (Kumar and Pansari, 2016).

A loyalty program already sees most of these parts. Purchases feed lifetime value, and referral bonuses feed referral value. Reviews and survey answers feed knowledge value. Influence is the hardest to see, though a referral code shared online gives a partial read.

How do loyalty programs use engagement scores?

Use the score to rank members and to trigger the next move. A falling score can start a win-back offer. A rising score can open a tier challenge or a referral ask. Treat it as an input to decisions, member by member.

Keep it out of the headline. A composite hides which behavior moved, and no two teams build it the same way, as our guide to customer engagement argues. Report its parts instead: active member rate, events per active member, breadth of behavior and movement between value segments. Breadth deserves its own line, since members who do several things have more reasons to come back.

Pair behavior with attitude where the money is. Hilton says the share of travel spend it captures from Honors members rises as they move up its tiers (Hilton 10-K). Tier status is a behavioral reading. Gallup’s hotel club study adds the attitude side: attachment came with a larger share of lodging spend. A program that scores both on one record can tell a loyal regular from a habit that a rival could break.

Score new members against their peers. A member in her first 90 days has had little time to act, so a low score means less. Compare her with members who joined in the same month, and judge tenured members against their own past score.

Guard the weights against gaming. If app visits earn points and also lift the score, members will open the app for the points. Then the score rises while spend stays flat. Check each quarter that the score still tracks spend and retention by cohort.

How GRAVTY handles customer engagement scores

GRAVTY’s models learn from what members do. They predict how engaged each member will be and fit offers to each one (GRAVTY technology). Scores from GRAVTY or from outside tools sit on the member record, and offers can act on them (customer retention).

Events stream in as members act. Patented Visual Rules check each one against segments and offers. A patented method turns what members do into data the brand owns. It is called Multi-Dimensional Behavior Tracking. Deutsche Telekom’s Magenta Moments runs engagement this way on GRAVTY. It has 6M+ active members.

Frequently asked questions

Is customer engagement score the same as customer effort score?

No, they share the CES short form but measure different things. Customer engagement score adds up what customers do. Customer effort score is a survey question on how much effort it took to get an issue solved. It came out of research published in Harvard Business Review in 2010 (HBR, 2010).

What is a good customer engagement score?

A good customer engagement score is one that rises over time and tracks spend and retention in your own data. Each firm builds its own scale and weights, so scores do not compare across firms. Compare this year’s join cohort with last year’s at the same age instead.

What actions count as customer engagement?

Any action a customer takes toward the brand counts, buying included. Research lists word of mouth, tips to friends, helping other customers and writing reviews (van Doorn et al., 2010). In a loyalty program, add points spent, referrals, offer clicks and app visits.

How is engagement different from satisfaction?

Satisfaction looks back at past experience, while engagement covers the bond and what the customer does. Gallup argues that customers stay faithful to brands that earn both rational trust and deeply felt affection. So its CE11 keeps satisfaction as one of 11 questions. It adds eight more on the emotional bond (Gallup, 2001).

How often should an engagement score be updated?

Update it as events arrive when it drives triggers, and monthly when it feeds reports. Give recent events more weight or let old events fade, so last year’s burst does not hold up this month’s score. Review the weights against spend each quarter.