Referral Program Launch Checklist for DTC Brands

Most DTC referral programs launch with a button in the footer, a vague promise of a discount, and silence for the next six months. Then the brand concludes that referrals do not work for their category. The program did not fail because customers did not want to share. It failed because the incentive math was backwards, the placements were hidden, and nobody ever told existing buyers the program existed in the first place.

A referral program is a product. It has users (advocates and friends), a core loop (share, click, convert, reward), failure modes (fraud, confusion, low visibility), and metrics (share rate, click rate, conversion rate, revenue per advocate). Treat it like a product launch and you get a channel that compounds quietly for years. Treat it like a checkbox on a Shopify app list and you get a widget nobody uses.

This checklist walks the 20 decisions and setup steps that separate the programs that print revenue from the programs that gather dust. Use it end to end before you turn anything on. For the broader retention picture this program sits inside, see our retention marketing hub.

Incentive math

Incentive math is the first place most programs quietly break. A brand picks 20 percent off for the advocate and 20 percent off for the friend because it sounds fair, then discovers six months later that every referred order is losing money once you subtract COGS, shipping, payment processing, and the advocate credit. By then the program is baked into customer expectations and rolling it back feels like a tax hike.

Start with gross margin per order, not revenue. If your blended gross margin after shipping and processing is 55 percent and your AOV is 72, you have roughly 40 of margin to work with per order. A double-sided 15 percent discount costs you about 22 on a referred order (15 percent to the friend up front, then 15 percent credit to the advocate redeemed later). That leaves 18 of margin on the first order, plus whatever lifetime value the new customer brings. That is a workable trade. Flip those numbers (20 percent gross margin, 10 percent double-sided incentive) and you are essentially buying customers at breakeven hoping LTV rescues the math.

Fixed-dollar incentives versus percent incentives is the next fork. Fixed dollars feel concrete to the sharer ("give 10, get 10") and are easier to explain. Percentages scale with basket size, which protects margin on smaller orders but feels underwhelming on low-AOV catalogs. For stores with tight AOV bands (most of your orders cluster within 20 percent of the mean), fixed dollars usually win on clarity. For stores with wide AOV distributions, percentages protect margin.

Store credit versus discount code is the third decision and the one most brands ignore. Store credit for the advocate keeps the reward inside your ecosystem, which means the advocate has to come back to redeem. Discount codes are one-and-done. If your goal is to turn advocates into repeat buyers (and it should be), store credit is almost always the right choice for the advocate reward. For the friend, use a discount code because they are a first-time buyer and need the price reduction at checkout to pull the trigger.

Validate every incentive scenario against your real AOV distribution before you launch. Pull the last 90 days of orders, model the program on top of them, and ask what margin looks like if 5 percent of orders become referrals. Then ask what it looks like if 20 percent do. If either scenario breaks the unit economics, the incentive is wrong, not the program.

Platform setup

Platform choice matters less than most founders think and more than most agencies admit. Smile.io, Rivo, ReferralCandy, LoyaltyLion, Yotpo Loyalty, and Friendbuy all work. They differ on fee structure, placement flexibility, and how cleanly they integrate with Klaviyo and your theme. None of them will save a bad incentive structure or a program with no placements.

Pick based on three criteria. First, does it support the incentive type you need (store credit, tiered rewards, fraud rules)? Second, does it integrate with your email platform as a native block or through a webhook, so you can drop referral links into flows without manual copy-paste? Third, how customizable is the post-purchase placement, because that placement does more work than any other surface in the program?

Install the app on a staging theme or duplicate your live theme before the integration. Referral apps inject scripts and blocks that can shift page weight and change layout in subtle ways. Test checkout, thank-you page, and email render on mobile and desktop before you push live. Audit page speed with PageSpeed Insights or Lighthouse before and after the install. If the app adds more than 300ms to LCP on mobile, work with the vendor on a lighter implementation or move to one that loads async.

Configure your data flow. Your referral platform should be pushing events into Klaviyo (or your ESP) for advocate sign-up, friend click, friend purchase, and reward earned. Each of those events unlocks a flow you will build in the email section below. Without events flowing, you are flying blind and the program cannot talk to customers at the right moments.

Placements and copy

A referral program with no placements is a referral program with no referrals. The mistake is thinking the customer account page counts as a placement. It does not. Nobody visits their customer account page unless something has gone wrong with an order.

The six placements that actually drive share volume, in order of revenue contribution for most DTC brands:

Post-purchase thank-you page. This is the single highest-performing surface because the customer is in peak excitement, they just spent money on you, and they are not being asked to buy anything else. A clean "give 10, get 10" block with a pre-filled share link captures 15 to 25 percent of post-purchase traffic for well-designed programs. If you want the deeper playbook for this surface, read our breakdown of the post-purchase experience for repeat buyers.

Post-purchase email (order confirmation, shipping confirmation, delivered). The order confirmation is the most-opened email you will ever send. Put a referral block below the order summary and above the footer. Not in the footer. Above it.

Klaviyo flows and campaigns. Referral blocks should live in your welcome series (after the first purchase converts), your post-purchase flow, your win-back flow, and your VIP campaigns. See our guide to Klaviyo flows that move revenue for the flow structure these blocks belong inside.

Product detail page. A small, non-intrusive block below the add-to-cart button that says "already a customer? Share and earn" converts existing customers who land back on PDPs to browse. Keep it small. Do not let it compete with the main CTA.

SMS. If you have SMS consent, a single message 7 to 14 days post-purchase with the referral link and a short pre-filled share copy drives meaningful volume. Do not over-send. One referral SMS per customer per quarter is plenty.

Account and order history pages. The lowest-value surface but still worth populating. Customers who go looking for the program should find it immediately.

For every placement, pre-fill the share copy. Do not ask the advocate to write their own message. Draft three to five variants ("I love this stuff, here is 10 off your first order," "been using this for months, here is a discount to try it") and let the platform rotate them. The moment you ask a customer to compose their own share message, share rate drops by half or more.

Fraud and rules

Every referral program with real volume attracts fraud. Self-referrals (same person using a second email and shipping address), cookie-swapping (deleting cookies and clicking your own link), coupon stacking (using the referral code with other promos), and bot-driven sign-ups are the four most common vectors.

Configure fraud rules before launch, not after you notice a spike in suspicious orders. Most referral platforms let you block same-IP self-referrals, require a minimum order value for rewards to trigger, hold rewards in pending state until the return window passes, and flag patterns like 10 sign-ups from one device in 24 hours. Turn all of these on. The friction for legitimate customers is zero. The friction for fraud is meaningful.

Decide your stance on coupon stacking. Most programs disable stacking with other discount codes, which means the referral code does not combine with site-wide sales or newsletter-signup discounts. This protects margin but can confuse customers during promotional periods. Write a clear rule, explain it on the landing page, and train support to reinforce it.

Set a return-window hold on rewards. Advocate credit should not vest the moment the friend's order is placed. It should vest after your standard return window closes, usually 14 to 30 days. This prevents the pattern where an advocate refers a friend, the friend orders and returns, and the advocate still keeps the credit.

Cap rewards per advocate per period if your program is generous. A daily or weekly cap (say, five referred orders counted per advocate per week) prevents the edge case where one super-user gaming the system consumes outsized budget before you notice.

Measurement

A referral program you cannot measure is a referral program you cannot improve. The measurement stack is not complicated but almost nobody sets it up before launch and then debugging attribution six months in becomes a nightmare.

Set up GA4 events for four moments: advocate joins program, advocate shares link (by channel if possible), friend clicks link, friend completes purchase. Tag referral traffic with a consistent UTM structure (utm_source=referral, utm_medium=advocate, utm_campaign=program_name). Your referral platform should do most of this automatically, but audit the implementation. Pop open dev tools, click a test referral link, and confirm the UTMs and events fire as expected.

Decide attribution rules in advance. First-touch attribution credits the referral channel when a friend clicks a referral link, even if they come back a week later through a paid ad and convert. Last-touch credits whatever channel delivered the final click. Neither is wrong, but they produce very different ROI numbers. For referral specifically, a 30-day first-click attribution window is the most defensible. The advocate's referral opened the door, even if another channel closed it.

Track the right KPIs and track them weekly. Share rate (advocates who shared at least once / eligible customers). Click-through rate (friend clicks / shares). Friend conversion rate. Revenue per advocate. Cost per acquired customer via referral. Fraud rate. Post-launch, revisit all six at 30, 60, and 90 days. If share rate is below 5 percent, your placements are broken. If CTR is below 10 percent, your share copy is weak. If friend conversion is below your site-wide conversion rate, the landing experience is the problem.

Connect referral customers to lifetime value reporting. Referred customers typically outperform paid-acquisition customers on repeat purchase rate and LTV by 20 to 40 percent. Confirm this is true for your brand, because that multiplier is the real argument for investing more in the program. Our guide to ecommerce customer lifetime value walks the LTV math referral programs live inside.

Schedule a 30-day review with calendar invites before launch day. Put the review on the calendar now. If the program is underperforming after 30 days, the instinct will be to let it quietly die. The scheduled review forces a real decision: iterate, pivot, or kill.

-> Validate incentive math against your real AOV distribution and gross margin, not revenue.

-> Ship six placements at launch (post-purchase page, order confirmation email, PDP, Klaviyo flows, SMS, account page). Not just the footer.

-> Configure fraud rules, return-window holds, and attribution windows before day one, not after the first suspicious order lands.

-> Book the 30-day review on the calendar before you hit launch, so the program gets a real decision instead of a slow fade.