Referral programs have a measurement problem that is easy to miss. Word of mouth was already sending you customers before you installed ReferralCandy or Friendbuy. Once the program is live, some of those same customers now arrive through a referral link with a discount attached. The referral dashboard counts every one of them. So the real question is not how much revenue the program reports. It is how many of those customers you would not have gotten otherwise, and what you paid in rewards to turn existing word of mouth into tracked orders.
What the referral dashboard counts
Referral apps credit an order when the buyer arrives through a referral link or uses a referral code. That is precise about the last step and silent about everything before it. A friend who was told about your brand at dinner, then saw two Meta ads, then got a referral link from that same friend shows up as a referral order in the app and a Meta conversion in Ads Manager. Both tools will happily count it.
Codes add their own noise. Referral codes get posted to coupon sites and deal forums, and customers sometimes refer themselves with a second email address. Any order that used a referral code without a referral link click deserves a second look before you count it as a new customer from a friend. The same issues come up with influencer codes, which I covered in attribution tools for discount and influencer codes.
Use your survey as the baseline
The best test I know for a referral program uses a post-purchase survey you were already running. Look at the share of first-time buyers who answered “friend or family” in the months before the program launched. Then look at the same share after launch. If the program is creating new word of mouth, that share should rise. If it stays flat while referral orders climb in the app, the program is mostly relabeling customers who were coming anyway, and paying them a reward for it.
If you don’t have a survey yet, start one before launch so you have something to compare against. Guidance on question design is in running post-purchase surveys that tell you something.
Measure the cost per referred customer
Referral programs usually pay twice. The new customer gets a discount and the referrer gets a credit or reward. Add both to the app fee, divide by the number of new customers the program brought in, and you have a cost per referred customer you can put next to your paid CAC. Most programs look cheap on this measure, which is fine, as long as you are honest about how many of those customers were incremental.
Then follow those customers. Referred customers often have strong repeat rates, because a friend’s recommendation carries more trust than an ad. If your referred cohort repeats at a higher rate than paid social customers, that supports giving the program more room, such as bigger rewards or more prominent placement in post-purchase emails. If they behave like discount hunters and never come back, the rewards are probably too generous.
Tag referral links properly
Make sure referral links carry UTM parameters so they land as their own source in every tool, not as direct traffic. Check this in your referral app’s link settings and test a link yourself. Untagged referral visits get absorbed into direct, which makes direct look healthier and hides the program’s real contribution. The bigger picture on direct traffic is in attributing direct traffic sales.
Where ThoughtMetric fits
ThoughtMetric combines multi-touch attribution from its own pixel with post-purchase survey answers, so you can see a referred buyer’s earlier ad and email touches alongside their survey response instead of handing the whole order to one channel. (Disclosure: ThoughtMetric sponsors this site.) Its survey also gives you the “friend or family” baseline described above without a separate app. Pricing is based on monthly pageviews, starting at $99 a month for 50,000 pageviews, with every feature included on every plan and a 14-day free trial. Your referral app is still the place to manage rewards and fraud rules.
A quick example
Say “friend or family” was 14 percent of first-order survey answers before launch. Six months later the referral app reports 300 new customers a month, which is 20 percent of new customers, and the survey share is 17 percent. The survey suggests word of mouth grew by about three points, not twenty. Some of the program’s 300 customers are genuinely new, and a larger share are people who would have found you through a friend anyway. That is still a useful program, but it is a much smaller one than the dashboard claims.
Picking an approach
- Running a survey already? Track the “friend or family” share before and after launch.
- Not running one? Start it before you launch the referral program so you have a baseline.
- Seeing code use without link clicks? Check for coupon site leakage and self-referrals.
- Deciding on reward size? Compare referred customer repeat rates to your paid channels.
- Want referral touches, ad touches and survey answers in one report? ThoughtMetric (our sponsor) covers that.
Leave a Reply