How to Structure a B2B Referral Program That Actually Works.
Five real programs, their published terms side by side, and the one piece of rigorous evidence on whether referrals are worth paying for. None of the five publishes results, so the structure is what we can actually compare.
Most advice on referral programs is built on consumer examples: give a friend a free month, get a free month. B2B doesn't work like that. The person referring is a customer or a partner with a professional reputation, the person being referred is deciding a purchase with real money and real switching costs, and the reward has to be worth a referrer's time without being the reason anyone buys.
So instead of a list of tips, this issue reads the actual terms of five programs from Mercury, Gusto, HubSpot, Webflow and Shopify, pulled from each company's own pages in November 2026. Then it asks the question the programs themselves don't answer: how much can you afford to pay?
Five Programs, Side by Side
These fall into two camps. Mercury and Gusto pay a one-time bounty to a customer who refers another business. HubSpot, Webflow and Shopify pay a share of revenue to an affiliate or partner for a fixed period. The differences show up in what triggers the payment, how long it runs, and what is excluded.
| Program | Reward | Trigger and duration | Both sides? |
|---|---|---|---|
| Mercury (business banking) | $250 to the referrer | New business holds $10,000+ for a full business day within 90 days | Yes, $250 to the new business |
| Gusto (payroll) | $300 for the 1st referral, rising to $400, $500, $700, then $1,000 from the 5th | Referred business runs a paid payroll and keeps an active account | Yes, $100 |
| HubSpot affiliates | 30% of payments on core products | Up to 12 months while the customer stays active; last-touch attribution | No |
| Webflow affiliates | 50% of the first subscription's revenue | Up to 12 months on monthly plans; plus 10–15% on renewal at higher tiers | No |
| Shopify Partners | 20% of the monthly subscription fee plus 0.1% of eligible sales | 4 years; flat $2,500 for a Plus referral | No |
Three patterns stand out, and they are design choices rather than coincidences.
Pattern One: Pay for Value, Not for Sign-Ups
Not one of these programs pays when someone signs up. Mercury pays when the new business parks $10,000 in the account. Gusto pays when the referred business has run an actual payroll. Shopify's $500 POS Pro referral requires the merchant to transact on at least 10 days a month for two months and generate over $1,000 in monthly sales. HubSpot holds commission until the customer has stayed at least 30 days (monthly plans) or 60 days (annual plans).
That gate is the whole point. A referral reward paid at sign-up buys sign-ups, and sign-ups are cheap to fake. A reward paid at the moment the customer proves they're real costs you only when the acquisition worked. Pick the earliest event in your product that tells you the customer will stick, and attach the payment to that.
Pay at the moment the customer proves they're real, not at the moment they show up.
Pattern Two: Bounty or Revenue Share, and How Long
A flat bounty and a revenue share cost you very different things. A bounty is a known, one-time cost, easy to budget and easy to compare with your acquisition cost from Issue 44. A revenue share scales with the customer, which rewards referrers for sending you bigger accounts but makes your cost harder to predict.
To compare them fairly, put the same hypothetical customer through each. Suppose a customer pays $100 a month on the relevant plan. This is my arithmetic on the headline rates, not a statement of what any company actually pays on a given plan:
| Program | Rate | Paid over | Total to referrer |
|---|---|---|---|
| HubSpot | 30% | 12 months | $360 |
| Webflow | 50% | 12 months | $600 |
| Shopify (subscription only) | 20% | 48 months | $960 |
The ordering by rate (Webflow, HubSpot, Shopify) is the reverse of the ordering by duration. Shopify's rate is the lowest and its term the longest, so it pays the most in total on a loyal merchant. A high percentage over a short window and a low percentage over a long one are different bets on how long customers stay.
The right choice depends on your own retention. If customers stay for years, a long, low share rewards the referrers who send you durable accounts. If churn is high, a short window protects you.
Pattern Three: Rules That Stop the Program Being Gamed
Read the exclusions and you can see where each company has been burned, or expects to be. HubSpot won't pay on existing customers, direct client referrals, customers already assigned to a sales rep, or anyone already in an active sales process. Webflow excludes self-purchases and purchases by anyone at the affiliate's own organization, and routes agency client referrals to a separate partner program. Mercury won't pay you for referring your own business. Gusto's terms bar referrers from bidding on its trademarked terms in paid ads or posting links on coupon or cashback sites.
A referral program without exclusions quietly pays for customers you would have won anyway. The most important rule is the first one on HubSpot's list: no commission on someone already in your pipeline. Write yours before you launch, not after the first payout you regret.
Two smaller design choices deserve copying. Gusto's rewards escalate (from $300 to $1,000 as a referrer lands more customers), which turns a one-off favour into a reason to keep referring. And Mercury and Gusto reward the new customer as well, which gives the referrer something to offer instead of just something to gain. The affiliate programs skip that, because their referrers are mostly publishers and partners, not peers.
What Evidence Says Referrals Are Worth
None of the five programs publishes results: no conversion rates, referral volumes or payout totals that I could find. The best rigorous evidence on whether referred customers are worth paying for comes from outside B2B entirely. A 2011 study in the Journal of Marketing, by Schmitt, Skiera and Van den Bulte, followed about 9,500 customers of a German bank over 33 months.
Referred customers were more valuable. Their churn rate was about 18% lower after controlling for other factors, and their six-year lifetime value was about 16% higher. The bank paid referrers a 25-euro voucher, and the authors estimated a return of about 60% over six years. Two details are worth keeping. The margin advantage faded after roughly 29 months, while the retention advantage did not. And the effect didn't hold for low-margin customers.
Treat this carefully: it's one bank, one country, 2006 data, and a consumer product. It is evidence that referred customers can be better, not proof that your B2B referrals will be. Its most useful idea transfers anyway: the most you should pay for a referral is the extra value a referred customer brings over a normal one.
How Much You Can Afford to Pay
Revenue-share percentages are easy to copy and easy to misjudge, because the cost depends on your margin. A 30% share of revenue is a bigger bite out of a business with an 80% gross margin than one with 95%. The calculation is simple: share of revenue ÷ gross margin = share of gross profit you hand over.
An example, with made-up numbers. Say a customer pays $500 a month at an 80% gross margin, and you offer a 20% revenue share for 12 months. You pay $1,200 against $4,800 of gross profit over those months, which is 25%. A 50% share on the same product would hand over 62.5% of gross profit for the year. That can still be sensible if the alternative is paid acquisition at a worse payback, but you should know which you're choosing.
Then run it against payback, the number from Issue 44. If a referred customer repays their acquisition cost in a few months and a typical paid customer takes more than a year, you can afford to pay the referrer generously and still come out ahead.
Where These Numbers Come From
The program terms come from each company's own pages and were current when read in November 2026. They change often, so check before copying any of them. The study figures are from the published paper. Everything labelled hypothetical is my own arithmetic.
- Mercury: referral bonuses for business accounts
- Gusto: referral rewards terms (Oct 2025 promotion). Expired sign-up window; current terms may differ.
- HubSpot: affiliate program FAQ
- Webflow: affiliate program overview
- Shopify: how Partners earn. No effective date shown on the page.
- Schmitt, Skiera and Van den Bulte: Referral Programs and Customer Value (2011). Consumer banking, one German bank.
Write Your Referral Rules Before Your Reward
On one page, answer four questions. What is the earliest event that proves a new customer is real? What is the most you can pay before a referred customer stops being worth it, using your gross margin and payback? Who and what is excluded: existing customers, active deals, your own staff, self-referrals? And how long does the reward run? Only then pick a number. A program with the right rules and a modest reward beats a generous one that pays for customers you'd have won anyway.
