Loyalty trends and best practices

Loyalty Analytics: 4 Steps to a Clearer Picture

by Kaleb Loosbrock
December 11, 2013

Kaleb Loosbrock
Tracking customer loyalty across web, CRM, and loyalty platforms can be like viewing a Georges Seurat painting one dot at a time. Continuing in that vein, I'll use a painting analogy to illustrate best practices around loyalty analytics.

  1. Prep the Canvas: Laying the groundwork for loyalty analytics
  2. Sketch:  Choosing what to track
  3. Paint: Distinguishing between variables and events
  4. Refine: Honing your results

Prep the Canvas: Laying the groundwork for loyalty analytics 

The Canvas

Optimize your website for mobile interaction
Many consumers have a work computer, a home computer, a mobile device, and a tablet — and many use all of them on a daily basis. With traditional loyalty analytics, each touch point looks like a different customer . Make sure you have a responsive, mobile-optimized website with tracking capabilities equal to those of your desktop site so you get a single customer view.

Deal with the Cookie Monster Much of your data collection — for loyalty and otherwise — probably relies on cookies, which means it’s vulnerable to the Cookie Monster. That's the phenomenon in which users clear their cookies or use privacy-enabled browsers. The effect for marketers is like shaking an etch-a-sketch — the customer view disappears. Luckily, quite a few people have started tackling this issue. Also, check out inventive storage solutions, such as ever cookie, that try and evade the Cookie Monster.

Encourage — and reward — login The cookie situation gets more complex on mobile devices, since cookies are generally turned off on many devices (iPod, iPhone, etc). To get around this, we have login! Start by encouraging members and non-members to login on various devices. For members, think about rewarding them — even with as few as 5 points (typically equivalent to a mere fraction of a penny) — each time they sign in.

Use universal identifiers Web analytics providers typically state in their terms and conditions that you’re not to send them personally identifiable data. To tell who's who, devise a universal unique identifier that doesn't disclose personally identifiable information. Most reward programs have a loyalty ID or account number that isn't an email address. Make sure this identifier is present across all systems.

Sketch: Choosing what to track

Figuring out what data you need  — and what data you don't — to measure loyalty can be tougher than it sounds. Here are two critical things to keep in mind:


Identify your key loyalty questions What questions do you want answered in order to evaluate program initiatives? For example, will you want to know how many loyalty program members are returning on a weekly, monthly, and daily basis? How about how their purchasing patterns for particular products differ from those of non-members? How does attribution differ across customers? Even if you initially track just a single customer action, having a sense of the questions you’d like to answer will help you establish a report structure that scales well with an increase in the number and types of reporting.

Sketch loyalty reports you want, and work backwards to implement them Knowing in advance how you’d like your reports to look will dictate the way you implement tagging and tracking. For example, if reports only need to show member versus nonmember activity, your tracking structure will be quite different than if you need other details, such as tier status, decile, conversions and AOV. If possible, mock up an example report with all the data that you’d like, and share it with your implementation specialists. They’ll be better equipped to meet or exceed your expectations if they know exactly what you want.

Paint: Distinguishing between variables and events

Events are key activities that your customers perform, such as adding a product to shopping cart, making a purchase, enrolling, and logging in. Variables are attributes that are associated with those events. For example, variables for a purchase might include the product purchased, time spent in checkout, the amount paid, etc.


Track login Login takes “guess who” out of the equation. Remember the pesky little problems we talked about earlier around cookies? With login tracking, you can see the customer across channels and devices.

Track member variables Deciding what to track at login depends very much on your business needs, your loyalty program, current data tracking practices, and your analytics platform. In general, though, there are a few key variables that you should track.

Variable 1: Unique identifier What’s the unique identifier that you us to track your customer today? Remember, don't use personally identifiable information (no email, address, name, etc.).

Variable 2: Member/non-member If using an e-commerce platform, record at login whether the user is a member or non-member of the loyalty program. Otherwise, set the member value when the user logs into the loyalty program.

Variable 3: Tiers/subscriptions If your loyalty program incorporates tiers, be sure to record the tier value so that you can track the performance of the tier.

Variable 4: Decile/Quartile If you divide your customers into deciles or quartiles, devise a naming convention for your customer splits, and at login place that customer’s group identifier into a variable. If you’re short on variables, you can look at appending your decile or quartile ID to any of the previous variables.

Track additional member events Once you're tracking login variables, move on to adding key customer events.

Event 1: Reward Redemption & Point Purchases

  • If your program doesn't allow for points at the time of purchase but has other reward-based redemption (e.g. free product or free shipping), capture redemptions in a variable so that you can report on them later.
  • If your program allows for points at the time of purchase, track the overall number of points redeemed with each transaction.

Event 2: Point searches Similarly, while point purchases will tell you the rewards redeemed, point searches will give you the demand side of the funnel.

Event 3: Unsubscribe If a member unsubscribes, hopefully that happens through your site. This way you can analyze campaign unsubscribes within certain segments.

Refine: Honing your results Your loyalty story is hidden in your data, and it’s your job to find it. Wait for results to accumulate until you have a clear picture (about 2-3 months), but then take action to optimize your program.


Start by analyzing conversion funnels and click-through streams of members and non-members:

  • Look at referral reports and keywords to see which sources are associated with your most valuable customers (e.g. Gold members, #1 decile) in order to decrease your CPA and ad spend.
  • Build segments based on the information and see how those segments differ from one another in conversion and behavior.
  • Define and create goals around conversion, email sign up, and other events.
  • Run A/B tests around your new dimensions and segments.

Armed with a broad palette of analytics tools from your various platforms—segmentation, A/B/multivariate testing, personalization, attribution modeling—you can tailor your loyalty program to meet your business needs. 

Want the full 500friends guide to Loyalty Analytics? Download it here.

Questions about loyalty analytics? Reach out to me at kaleb(at)


Enter your email address recieve notifications of new posts