A Referral Program Nobody Used, and Why We Should Have Seen It Coming

Eleven referrals in six months, from a user base of roughly 30,000 active accounts. The incentive was genuinely good — a month free for both parties — the UI was clean, and the launch email had a better-than-average open rate. By every launch-quality metric we track, this program should have worked. It didn't, and the reason was obvious in hindsight and completely invisible to us beforehand.

The product nobody wants to talk about using

The product was an accounting tool for freelancers. Referring a friend to your accounting software means telling that friend, implicitly, something about your business finances — that you're freelance, that you're at a certain stage, that you use this specific category of tool. We'd designed a referral program for a product category where the actual barrier wasn't incentive size or UI friction, it was that almost nobody talks about their accounting software socially, incentive or not.

A referral program assumes people are already willing to talk about your product. If that assumption is wrong, no incentive size fixes it.

What we should have checked before building it

• Whether users organically mentioned the product unprompted anywhere we could observe — support tickets, social mentions, NPS verbatims — before assuming they'd refer it with an incentive attached.
• Whether the product category itself carries any social visibility cost, which accounting and several other 'quiet' categories genuinely do.
• Interviewing a handful of power users specifically about whether they'd ever mentioned the tool to a peer, rather than assuming the incentive would create willingness that didn't already exist.
• Looking at comparable products in the same category for any evidence a referral mechanic had ever worked there, instead of assuming our execution would be the differentiator.

We eventually replaced the referral program with a case-study and testimonial program instead — asking happy customers to be featured, which carried positive social visibility rather than requiring them to vouch for a category they didn't want to be publicly associated with. That program outperformed the referral program by every measure, because it worked with the actual social dynamics of the category instead of against them.