There’s a customer somewhere on your email list right now who loves what you sell.
You bought twice. You have left a review. They would love to tell a friend about you, and their friends probably look a lot like your ideal buyer.
And most likely, you never asked them to.
Most founders I talk to have a referral program, just like they have a gym membership. It technically exists. No one has used it since January. There’s a link somewhere in the footer, and that roughly shows the scope of the strategy.
That’s a shame, because calculating the recommendations is almost embarrassing. When a friend says, “You would like that,” the sale is already made. No advertising platform can buy that and no targeting algorithm can reproduce it.
Why do most referral programs produce almost nothing? Usually because they are created as a page and not a program. Let me show you the difference.
Do you have little time? Here are the key takeaways
- Referred customers are your best customers: They convert better, spend more, and stay longer because they arrive through trust, not an ad.
- Timing suggests incentive: The best time to ask is immediately after a positive experience and not always when there is a gap in your campaign calendar.
- Double-sided rewards eliminate the embarrassment: When both sides get something, no one feels like they’re selling it to their friends.
- Ask your happiest customers, not all of them: A recommendation request to your entire list fails. For your best segment, it works.
- Make sharing effortless: Unique link, pre-written message, one tap. Every additional step costs you referrals.
Why recommendations beat almost everything else
Think about the last thing a friend recommended to you. A restaurant, a show, a pair of shoes. Did you read reviews first? Compare alternatives?
Of course not. You just tried it because someone you trust has already done the verification.
This is a recommendation. Trust built over years, transferred to your brand in an SMS.
This is why referred customers consistently have higher conversion rates than any paid channel, cost next to nothing to acquire, and stay longer. People don’t recommend friends who aren’t right for them. Your customers target on your behalf better than any algorithm because they actually know the person on the other end.
The brands that do this right don’t view recommendations as a website feature. You think of it as an email strategy with a schedule, a trigger, and a specific audience. We’ve already written about why word of mouth never stops working, and a recommendation engine is simply that idea with the infrastructure behind it. That’s all the change, and the rest of this article is about how to create it.
Ask in the moment of joy
Here’s something almost everyone does wrong, and it’s not about the reward or the copy.
It’s the timing.
A recommendation request hidden in a random Tuesday newsletter is a background image. The same question, right after someone leaves a five-star review, hits completely differently. Same offer, same words, completely different result.
The moments to watch out for: a second or third purchase. A rave review. A reply to one of your emails saying they like the product. A support ticket that ended well. These are the windows when goodwill is at its peak, and “Do you know anyone who would like this too?” feels like a natural question rather than an extraction attempt.
Of course, you can’t manually monitor these moments across thousands of customers. This is exactly what job automation was designed for. If you purchase again, a referral invitation will be triggered a few days after delivery. A high review score triggers a thank you with a share link hidden within. You set it up once and the question lands at the perfect moment forever.
Reward both sides (or don’t bother each other)
I’ll be blunt: one-sided referral rewards don’t work, and I suspect most founders already know this when they set them up.
If only the referrer gets paid, they are monetizing their friendships, and that feels sleazy. If only the friend gets the discount, the recommender has no reason to lift a finger. Either way, you’ve created a program with a loophole.
Double-sided is the answer, and that’s not even close to being the case. “Give $10, get $10” turns the recommendation into a favor. The recommender sincerely helps his friend and receives thanks in return. The friend comes with a discount that makes the first purchase easier. All incentives point in the same direction, including yours.
The reward doesn’t have to be extravagant either. Free shipping works. A modest dollar amount is enough. Early Access works great for brands with any publishing cadence. What kills programs is not a small reward.
It’s next.
Friction kills recommendations
Every referral program leaks in the same two places: the moment someone has to think about what to write and the moment they have to figure out how to share it.
So repair both. Each customer receives a unique link that automatically handles tracking and assignment. The message is pre-written, something short and human that they can send as is or tweak. And the sharing buttons are where the thumbs are already: text, WhatsApp, email, one tap at a time.
My rule of thumb: If recommending a friend takes more than fifteen seconds after opening your email, it’s too difficult. Your super fans will still make it. But superfans make up maybe 2% of your list. The silently satisfied majority, the people who would happily recommend someone else if it were effortless, don’t struggle through a cumbersome process.
And that’s where the actual volume is.
One more thing, and this is the least glamorous advice in this article: remind people. A referral program once announced leads to a surge and then silence. The resulting programs are mentioned constantly, in post-purchase procedures, in newsletters and on receipts. Not loud. Just often enough that your customer remembers they have a link when a friend asks, “Where did you get that?”
Don’t ask everyone. Ask the right people.
Would you ask a stranger at a bus stop to recommend you to his friends?
This is basically what creates a referral explosion to your full list. The customer who made a purchase eight months ago and hasn’t opened an email since then isn’t going to recommend anyone. If you ask them anyway, your entire list will just learn to skim the program.
Instead, build the segment that deserves the demand. Repeat buyers. Reviewers. High commitment. Anyone who has ever responded to an email with something heartfelt. This group should hear about your referral program more often and more personally than anyone else, as they are likely the only ones who will act on it.
This is where Omnisend quietly does most of the work. Its segmentation isolates those exact customers and integrates with loyalty tools like Smile and Yotpo to manage points, tiers and referral tracking behind the scenes, triggering the right email when someone earns a reward or their friend makes a purchase. The sanitary facilities run by themselves. Your only job is to write a question worth answering “yes.”
Final thoughts
Here’s the part that I find really motivating about all of this.
Word of mouth around your brand is already happening, whether you have a program in place or not. People mention things they like. The only question is whether you’ve made it easy for mentions to travel and whether anyone will be thanked for it.
A recommendation engine is just an existing infrastructure for generosity.
Omnisend gives you this infrastructure: automation that asks at the right moment, segmentation that finds your happiest customers, and integrations that manage rewards end-to-end. Foundr readers also get 50% off the first three months, just use code FOUNDER50 when you log in and let your customers do what they’ve already done. Just louder.
P.S Are you switching to Omnisend from another platform? Your migration team will move any flow, list, and template for you in five days, free of charge. You just show up when it’s ready and pay up to 35% less with SMS starting at $0.007 per message.
