Sales

Free customer referral program designer and incentive calculator

Name: Referral Program Designer
Author: Charles Summers

By Charles Summers · Updated 8 August 2026 · Free, no signup

Short answer

This computes what you can actually afford to pay for a referral. Gross profit per customer and a payback window set the ceiling, then the incentive style applies a cost factor, because a dollar of account credit costs between the cost of goods and the full face value depending on whether it is consumed incrementally or displaces revenue you would have collected. You get projected referral volume as a chain of stated assumptions, effective CAC per referred customer, the payback in months, the asking moment, and the double-sided versus single-sided decision with the reasoning behind it.

Use the referral program designer

What does this tool actually do?

It runs entirely in your browser. Nothing you type is sent to a server, no account is required, and there is no usage limit, because there is no cost per run to control.

The incentive is customer acquisition cost, so margin sets the ceiling

A referral payout is not marketing spend in some soft sense, it is CAC with a different label, and it has to survive the same two tests: it must be less than what you pay elsewhere for the same customer, and it must be repaid out of gross profit inside a window you can fund. That makes the calculation short. Take annual revenue per customer, multiply by gross margin to get gross profit, divide by twelve, and multiply by the number of months of payback you are willing to accept. That figure is the ceiling for everything you hand out on both sides of the referral combined.

The most common way programs go wrong is paying a percentage of revenue rather than a share of profit. Twenty percent of order value sounds modest and is fatal at a thirty percent gross margin, because it is two thirds of the profit on that order, before any fulfilment, support or return costs. The same twenty percent on software at eighty percent margin is a quarter of the profit and comfortable. The percentage is identical; the businesses are not, which is why any advice framed as a percentage of revenue is advice from someone who did not ask about your margin.

The payback window is the second lever and it depends on how repeatable the purchase is. Software billed annually can reasonably wait six months of gross profit, because the customer is contracted and the second year has no acquisition cost attached. An ecommerce store with genuinely one-off purchases has to make the payout work against the profit on a single order, because assuming a second order is assuming the thing you are trying to build. Services sit in between, and the honest version of the model uses observed repeat rates rather than the ones in the business plan.

Then leave headroom. Spending the entire ceiling means the program breaks even at the payback horizon and returns nothing before it, which makes it indistinguishable from a slower version of doing nothing. Half the ceiling is a reasonable default: it leaves a real contribution per referred customer and it gives you room to raise the payout later, which is a far easier conversation than cutting one.

Cash, credit and recognition cost different amounts and buy different behaviour

The cost side first, because it is arithmetic and it is usually got wrong. A cash payout costs you its face value, full stop. Account credit costs somewhere between the cost of goods and the full face value, and which end you land on depends entirely on whether the credit is consumed as usage the customer would never have paid for, or displaces revenue you were going to collect anyway. A free month on a subscription is the second case and costs you the full price of that month. Credit against overage on a metered product is the first case and costs you only what it costs you to serve. At an eighty percent margin that difference is five to one, which is why software companies offer credit and why they should be honest that a free month is not credit at all.

A discount to the referred customer behaves the same way in reverse. Its true cost is the face value multiplied by the probability that person would have bought at full price anyway. For a genuinely incremental customer, the discount is simply a lower price on revenue you would not otherwise have had; for someone who was already on your pricing page, it is a straight transfer out of gross profit and the referrer just got paid for a sale you had already made. That second case is where a large share of referral budget quietly goes, and attribution windows are how you control it.

The behavioural side matters at least as much, and it does not follow the cost. Cash is unambiguous and it converts a favour into a transaction, which is exactly what you want in some contexts and precisely what you do not want in others. Ask someone to recommend you to a friend and then hand them money and you have reframed a social act as a commercial one, which changes who they are willing to ask and what they say. Credit avoids most of that because it reads as a discount on something they already chose, and it also keeps the money inside your own economy.

In business-to-business there is a further constraint that is not about psychology at all. Paying cash to an individual employed by another company can breach that company's policy on gifts and inducements, and in regulated sectors it can be considerably worse than a policy problem. That is the real reason charitable donations and account credit are common in enterprise programs: they are the two structures that let a referrer participate without having to disclose anything awkward. Status and recognition cost almost nothing and work only where an audience genuinely exists, such as a professional community, a partner directory or a user group; where there is no audience, recognition is a certificate nobody sees.

Participation is the number the whole model rests on, and it is skewed

Everything downstream depends on how many customers actually refer anyone, and that is the number programs report least honestly. Vendor case studies usually quote a share of invited or engaged customers rather than a share of the base, which can differ by an order of magnitude. Measured properly, as referrals made divided by active customers in a year, participation for most programs sits in low single-digit percentages, occasionally reaching into the teens for products with a genuinely social use case or a professional community around them.

More importantly, the distribution is not normal, it is heavily skewed. A small group of advocates produces most of the referrals, and many of them were already recommending you before any incentive existed. That has two consequences the average rate hides. First, a program mostly pays people for behaviour it did not cause, at least at the start, which is fine if the payout is affordable and disastrous if you sized it on the assumption of new behaviour. Second, the highest-return design work is aimed at the top few percent, making it easy for them to refer repeatedly, rather than at nudging the indifferent majority into a first referral they will never repeat.

This is also why a referral program cannot manufacture word of mouth that does not exist. It amplifies an existing rate. Before building anything, measure the organic version: add a free-text "how did you hear about us" field and count the answers that name a person. If almost nobody arrives that way today, the honest reading is that the product is not currently recommendable, and an incentive will buy a brief spike of low-quality signups followed by the same silence. Fix the reason first; the program will be worth more when it is built on something real.

The asking moment, and one-sided versus two-sided

Willingness to recommend is not a stable trait, it is a state, and it peaks immediately after a success the customer noticed. That is the moment worth engineering for. Asking at signup captures enthusiasm before evidence, and produces referrals that convert badly because the referrer has nothing concrete to say. Asking after a resolved support problem outperforms most people's expectations, because a failure handled well is more memorable than a service that simply worked. The general rule is to trigger on an event rather than on a date: first successful outcome, a milestone reached, a renewal completed, a five-star response to a survey.

The one-sided versus two-sided decision has a mechanism behind it, and it is about who carries the social risk. A two-sided incentive lets the referrer arrive with a gift rather than a solicitation, which is why it dominates in consumer products: giving your friend twenty pounds off is a different social act from telling them about a company that will pay you twenty pounds. Where the referrer's motive is professional standing rather than money, or where paying them creates a compliance problem, the reward should sit entirely with the referred customer, and the referrer's return is that they look good for the recommendation.

The split between the two sides is not usually fifty-fifty. Weight it towards the referee when the purchase is low consideration and the job is to remove friction from a decision made in seconds. Weight it towards the referrer when the ask is genuinely effortful, such as an introduction that requires a personal email to a peer, because that effort is the scarce input. And attach the payout to the outcome you actually want: paying on signup buys signups, paying on the referred customer's second month or first renewal buys customers, and the difference in what it costs you is very large.

Finally, keep the mechanics boring. A referrer who cannot see whether their referral was recorded stops referring, so status visibility is not a nice-to-have. Ask twice at most, with a long gap, because a third prompt reads as pressure applied to a relationship you have asked them to spend. And set an attribution window short enough that you are not paying for people who were already in your funnel, because that leak is invisible in the reporting and shows up only as a program that appears to work while nothing else grows.

Numbers worth knowing

Metric	Typical	What it means
Participation, measured against the whole base	usually low single-digit %	Referrals made divided by active customers in a year. Case studies quoting double digits are almost always dividing by invited or engaged customers, which is a different and much smaller denominator.
Value of a referred customer	one long-run study found 16 to 25% higher	A frequently cited multi-year analysis of a European bank's program also found referred customers churned less. Directionally repeated elsewhere, but the magnitude is specific to that setting.
Referred lead conversion versus cold	commonly quoted at 2x to 5x	Flattered by the comparison: the baseline is usually unqualified inbound. Real, but do not model volume on the top of that range without checking your own two cohorts.
Payout as a share of first-period gross profit	keep both sides under about half	This tool's rule rather than an industry statistic. Spending the full ceiling breaks even at the payback horizon and contributes nothing before it.
Distribution of referrals across advocates	heavily skewed, not normal	A small group produces most of them, and many were already recommending you unpaid. Design for repeat referrers rather than for converting the indifferent majority.

Mistakes that quietly cost you results

Setting the incentive as a percentage of revenue: Twenty percent of order value is a quarter of the profit at an 80 percent margin and two thirds of it at 30 percent. Work from gross profit and a payback window instead, then convert to a face value at the end.
Paying the referrer at signup rather than at a retained outcome: You will buy signups, because that is what you priced. Attach the payout to the referred customer's second month, first renewal or first repeat order, and the same budget buys customers instead of registrations.
Treating account credit as though it cost its face value: Credit consumed as incremental usage costs you cost of goods; credit that displaces revenue you would have collected costs the full amount. A free month is the second kind. Knowing which you are giving changes the affordable face value by several times.
Launching a program before measuring organic word of mouth: A program amplifies an existing referral rate rather than creating one. Count the free-text "how did you hear about us" answers that name a person first. If that number is near zero, the incentive buys a spike of low-quality signups and nothing after it.
Running a long attribution window with no exclusions: You end up paying for people who were already in the funnel, which is invisible in program reporting and shows up as a referral program that looks successful while total new business is flat. Exclude open opportunities and cap the window.

What does the output look like?

This is the exact output the tool produces from the example inputs. It is generated by the same code that runs when you click the button, so what you see here is what you get.

REFERRAL PROGRAM: B2B SaaS, annual contract | Incentive: Account credit INCENTIVE CEILING Gross profit per customer per year $1,872 ($2,400 x 78%) Gross profit per month $156 Payback window for B2B SaaS, annual contract: 6 months Maximum affordable cost per referred customer, both sides combined: $936 Recommended cash cost (half the ceiling): $475 Spending the full ceiling means the program breaks even exactly at the payback date and contributes nothing before it, which is why the recommendation is half. FACE VALUE AT THIS INCENTIVE STYLE (Account credit) Credit costs you cost of goods when it is consumed as usage the customer would never have bought, and full face value when it displaces revenue you were going to collect. A free month is the second kind, whatever it is called. Cost factor runs from 0.22 to 1.00, so $475 of budget buys between $475 and $2,159 of face value. The high end applies only if the reward is consumed as genuinely incremental usage. PROJECTED VOLUME (every assumption is stated, replace them with your own) Customer base 850 Participation rate assumed 4.0% (4% base rate for B2B SaaS, annual contract, x1.00 for account credit) Customers who refer 34 a year Referrals each 1.3 Referrals generated 44 Referral to customer conversion 30% New customers a year 13 (range 7 to 20 at half and one and a half times the participation rate) New revenue at plan $31,824 a year, $24,823 of it gross profit Program payout at plan $6,299 Participation is the least reliable number here and it is skewed: a small group of advocates will produce most of these, and some of them already refer you unpaid. EFFECTIVE CAC Cost per referred customer $475 Payback on that cost 3.0 months of gross profit Break-even against paid your blended paid CAC has to exceed $475 for this to be the cheaper channel. If it does not, the program is buying customers you were already acquiring more cheaply, and the case has to rest on retention instead. Excludes platform fees and administration. Below roughly a few referrals a month, a paid referral tool's monthly fee is the largest line in the program and the CAC above is optimistic by a wide margin. One frequently cited long-run study of a bank's referral program found referred customers were worth roughly a sixth to a quarter more and churned less. If that holds for you, the affordable payout rises; check your own two cohorts before spending against it. PROGRAM MECHANICS Trigger: the first measurable outcome inside the product, or the completion of a successful onboarding milestone, whichever the customer would describe as the point it started working. Payout condition: pay on the referred customer's retained outcome, not on signup. For this product type that means the second billing period completing. Paying at signup buys signups, which is a different and cheaper thing than a customer. Attribution: cap the window and exclude anyone already in an open opportunity or an active trial. Without that exclusion you pay for people who were arriving anyway, and it is invisible in program reporting. Asking cadence: at most twice, with a long gap. A third prompt spends relationship rather than budget. Visibility: the referrer must be able to see that their referral was recorded. Silence after an introduction is the single most common reason a repeat referrer stops. ONE SIDE OR TWO Double-sided, weighted to the referrer In a considered business purchase the scarce input is the referrer's effort, because a real introduction means a personal message to a peer and a small amount of their own credibility. That is what you are paying for, so it takes the larger share. Suggested split of the $475 budget: referrer $275, referred customer $200. Keeps the money inside your own economy and reads as a discount rather than a payment, which avoids most of the awkwardness of cash. It is worth nothing to a customer who is about to leave, so it under-rewards your least engaged users, which is usually acceptable. BEFORE YOU BUILD IT Count the organic referrals you already get. Add a free-text "how did you hear about us" field and count the answers naming a person. A program amplifies an existing rate; it does not create one, and if that count is near zero the incentive will buy a short spike of low-quality signups and then nothing.

Frequently asked questions

How much can I afford to pay for a referral?

Take annual revenue per customer, multiply by gross margin, divide by twelve, and multiply by the number of months of gross profit you are willing to spend acquiring a customer. That is the combined ceiling for both sides of the incentive. This tool sets the payback window from your product type, because an annual software contract can wait six months while a genuinely one-off ecommerce order has to work against the profit on that single order. It then recommends half the ceiling, so the program contributes something before the payback date rather than exactly at it.

Is cash or credit the better incentive?

Credit is cheaper per unit of perceived value whenever it is consumed as usage the customer would not otherwise have bought, because you are paying cost of goods rather than face value. It is not cheaper at all when it displaces revenue you were going to collect, which is what a free month is. Behaviourally, cash converts a recommendation into a transaction, which is fine for consumer products and a genuine problem in business-to-business, where paying an individual at another company can conflict with their employer's policy on inducements.

Should the incentive be double-sided?

It depends on who is carrying the social risk. A two-sided reward lets the referrer arrive with a gift rather than a solicitation, which is why it dominates in consumer products. Where the referrer's motivation is professional standing, or where paying them creates a compliance issue, put the whole reward on the referred customer and let the referrer's return be that they look good for the recommendation. The split is rarely even: weight it towards the referee for low-consideration purchases and towards the referrer when the ask requires real effort, such as a personal introduction to a peer.

When is the right moment to ask?

Immediately after a success the customer actually noticed, and triggered by that event rather than by a date. Asking at signup captures enthusiasm before evidence and produces referrals that convert badly, because the referrer has nothing specific to say. Asking after a support problem that was resolved well works better than most people expect, since a recovered failure is more memorable than a service that simply worked.

Why does the tool show a range for projected referrals?

Because participation is by far the most uncertain input and the model is a chain: base multiplied by participation, by invites per participant, by the rate at which a referred lead becomes a customer. An error in the first term propagates through all of it. The output prints every assumption so you can replace them with your own numbers, and it brackets participation at half and one and a half times the modelled rate, which is a fair reflection of how much this varies between businesses that otherwise look alike.

Related free tools

Sales Objection Handler Sales
Discovery Call Question Builder Sales
Sales Battlecard Builder Sales
Net Revenue Retention Calculator Analytics
Pipeline Coverage Calculator Analytics

Some links on this site are affiliate links, which means Hacking Demand may earn a commission if you buy through them at no extra cost to you. This does not influence which tools are listed. The tools on this page are free and have no affiliate relationship of any kind.