.jpg)
Referral programs often get designed like tiny social campaigns. The icons sit there waiting for a customer to broadcast the brand to a feed: Facebook, X, maybe Threads if everyone is feeling current. The mental picture is public. A customer loves a product, posts about it, and the crowd follows.
The actual sharing behavior is more private than that. The public behavior is more common in affiliate programs, where public personalities (influencers, content creators, etc) share a brand with their audience.
We selected a sample of one thousand live Shopify merchants with tracked referral share-button activity on ReferralCandy to answer the question of whether word of mouth is a public or a private channel.
Chat apps and email combined accounted for 69.2% of raw share-button clicks over the twelve months from June 1, 2025 to May 31, 2026.
On the other hand, public social accounted for 30.8%.
But when we dig deeper into the data, we see that chat apps (such as WhatsApp, Telegram or Messenger) alone drove 61.6% of tracked share-button clicks. Email only added another 7.6%.
Social buttons still mattered, but they were not the main way customers chose to pass a referral along.
Once we created the merchant segment for the study, we then focused on the merchants where customers clicked at least one tracked referral share button during the measurement window.
We grouped the raw button clicks into three channels:
Note: We counted only clicks on specific share buttons, not broader “email” activity fields. That matters because clicking from an email does not always mean the customer chose to share by email; they may have simply opened a referral email or used it to navigate elsewhere. So we used the narrower definition: only explicit email-share button clicks counted as email sharing.
The aggregate was not carried by a handful of unusually large programs. Among merchants with any tracked share-button click, 93.7% saw chat and email clicks outnumber public social clicks.
The pattern also held every month. From June 2025 through May 2026, direct channels beat social in all twelve months. The lowest month still had chat and email at 60.7% of tracked share-button clicks; the highest reached 80.2%.
Customers do not usually refer like broadcasters. They refer like people sending something to one person who might actually care.
That is the useful marketing truth underneath the numbers. A referral is a social action, but "social" does not necessarily mean a public feed. Most of the time it looks more like a message: "You should try this," sent to a friend, a sibling, a group chat, or someone who had already been asking for a recommendation.
The biggest surfaces in the data were still email surfaces. Reminder emails generated the most tracked share-button clicks, followed by the original referral email. But when customers clicked from those emails, the next action often moved into chat.
On reminder emails, chat and email share buttons made up 60.5% of tracked share-button clicks. On original referral emails, they made up 63.8%. On sharing pages, post-purchase widgets, embedded sharing pages, and extension referral links, direct channels were even more dominant, generally landing in the mid-80% range.
That makes the referral email less like the whole referral journey and more like the prompt. It reminds the advocate that the program exists. The actual pass-along often happens somewhere more conversational.
Do not remove social buttons. Nearly a third of tracked share-button clicks still went to public social channels, and for some brands those clicks will be meaningful. But do not design the referral experience as if the public post is the default behavior.
Make the direct paths obvious. Put chat actions where customers naturally decide to share: the referral email, the reminder email, the sharing page, and the post-purchase moment. If WhatsApp or Messenger is common for your audience, those buttons deserve real placement, not leftover space after the social row.
Write the prefilled message like one person sending it to another. A feed post can tolerate a little polish. A direct message cannot sound like ad copy in a costume. The best referral message should feel easy to send without editing, because it sounds like something a customer could plausibly say.
And treat copy-link behavior as part of the same instinct, even though it is outside the channel split above. Customers often want control over where the link goes. A clean link, a clear reward explanation, and a message that survives being pasted into any app will usually do more for sharing than another public icon.
Referral programs work best when they match the shape of real recommendations. A customer rarely wants to announce every good product to everyone they know. They want to send the right thing to the right person.
That is why referrals convert differently from ads. Trust is the channel, moving through whatever conversation feels natural: Facebook, WhatsApp, email, a copied link, or whatever comes next. Most of the time, that conversation is direct.
Raúl Galera is the Growth Lead at ReferralCandy, where they’ve helped 30,000+ eCommerce brands drive sales through referrals and word-of-mouth marketing. Over the past 8+ years, Raúl has worked hands-on with DTC merchants of all sizes (from scrappy Shopify startups to household names) helping them turn happy customers into revenue-driving advocates. Raúl’s been featured on dozens of top eCommerce podcasts, contributed to leading industry publications, and regularly speaks about customer acquisition, retention, and brand growth at industry events.
Grow your sales at a ridiculously
lower CAC.
