This question arrives often from people running a small business with happy customers: how do I build a customer referral programme? A customer brings me a customer, and I reward them for it.
The idea is sound. Most people implement it in a way that turns it into a cost rather than a channel. They set the reward by feel, announce it to every customer, and two months later discover they are paying for customers who were coming anyway.
The useful part is that this has actually been studied, and there are published numbers you can build on instead of guessing. We will use them, and then I will tell you exactly where they do not apply to you.
Asking for a referral is not the same as running a programme
These are two different things and each has its own rules.
The first is asking a satisfied customer to recommend you. It is free, it requires only that you ask, and most people never ask. I wrote about that in influencers or reviewers, and stronger than both, your own customers, where the conclusion was that an ordinary person who tried something and spoke about it outranks any paid advertising.
The second is a reward attached to the referral, announced and structured. That is a different animal, because money changes how people behave and it changes your own arithmetic too. This piece is about the second one.
If you have not done the first, start there. A paid programme built on top of zero asking is a cost, not a channel.
There is real research on this, and it is not based on impressions
The central study is in the Journal of Marketing, by Philipp Schmitt, Bernd Skiera and Christophe Van den Bulte, titled Referral Programs and Customer Value.
They tracked roughly ten thousand customers of a leading German bank for almost three years, comparing customers acquired through a referral programme against customers acquired by other means, holding demographics and acquisition timing constant.
What they found, in figures: the average value of a referred customer is at least 16% higher than that of a nonreferred customer with similar demographics and time of acquisition. The reward the bank paid was 25 euros as a voucher. The difference in customer value over six years was 40 euros. That works out to a return on the reward of roughly 60% over six years.
That last point is worth fixing in place. The programme paid. It did not pay spectacularly. Sixty per cent over six years, not over one.
The advantage changes shape over time, and that matters more than the headline
The study separated the advantage into two parts, and the difference between them is the practical bit.
First, contribution margin. Referred customers deliver a higher margin, but that difference erodes over time. The advantage is strongest in the first year and shrinks after it.
Second, retention. Referred customers stay longer, and that difference does not erode. The study estimated the reduction in churn at roughly 18%, reaching 20% in a robustness check.
Put those two sentences together and you get the practical consequence. If you judge the programme on the first sale alone, you are measuring the part that disappears and missing the part that lasts. The reverse also holds: if your business sells once with no repeat purchase, the lasting part does not exist for you, and the programme is weaker than you expect.
That connects to a broader question I have written about in whether to grow the business or look after the customers you have.
How to set the reward without losing money
This order is the whole thing, and most people start at the end of it.
Step one: know what a customer is worth to you over a realistic period. If a customer buys once a quarter and you know your average profit per order, multiply. Profit, not revenue.
Step two: work out the difference, not the value. You are not rewarding a customer. You are rewarding the increment a referred customer brings over an ordinary one. The study put that increment at 40 euros over six years. You have to put a number on yours, even an estimate.
Step three: set the reward clearly below that increment. In the study the reward was 25 euros against a 40 euro difference. If your reward approaches or exceeds your increment, you are buying growth at a loss. That can be a deliberate decision. It cannot be an accidental one.
Step four is the one people skip: do not pay for customers who were already coming. If somebody answers "how did you hear about us" with a name before you announce any reward, that is a free referral you are about to start paying for. Ask that question and record the answer for two months before the programme starts, so you know your baseline.
Who gets the reward, and should it be cash
Three options, each with a different effect.
Reward the referrer only. The simplest form, and the one the study tested. Its weakness is that the new customer has no reason to hurry.
Reward both sides. Stronger on conversion because the new customer has a motive, but your cost per acquired customer rises, so you need a larger value increment to absorb it.
Reward with something other than money. An added service, an upgrade, a free month. These usually cost you less than they are worth to the customer, and they avoid the feeling of buying someone’s words, which bothers people more than you would think in services.
My own view: in a small service business, non-cash is usually better. Not because cash is wrong, but because cash converts the relationship into a transaction, and the customer referring you is doing it because of the relationship.
One hard line though: do not pay for a positive review. That is completely different from rewarding an introduction. The first makes the words worthless, the second rewards genuine effort. I set that distinction out in what is the best thing your customers say about you.
When this programme is wrong for your business
The study itself says something important: the size of the value differential varies across customer segments, and firms should therefore use a selective approach for their referral programmes. A programme offered to every customer is not automatically right.
In practice there are three cases where I would not recommend one.
If your product is bought once, with no repeat and no subscription. The retention advantage, which is the part that does not erode, does not exist for you.
If your margin is thin. There is no room between the reward and the increment, and every referral eats the profit on the sale.
If the experience itself is not good yet. Referral magnifies what exists. A mediocre experience does not produce silence, it produces hesitant referrals, and you are paying for them.
That third case is the one I see most. Before any programme, the question worth answering is which channel has actually brought you customers so far, which is the subject of the one method that brought you the most customers.
The honest limits
Three, and the first is the big one.
This study is of a German bank. Banking is a high-involvement decision, people ask their friends before making it, and the relationship runs for years. Your small business in Egypt is not a German bank, and 16%, 40 euros and 60% are not your numbers. What transfers is the structure: keep the reward below the value increment, understand that part of the advantage erodes and part of it persists, and be selective rather than universal. The figures have to come from your own accounts.
Second, I do not have a study on the Egyptian market saying the same thing. If somebody quotes you a specific return for a referral programme in Egypt, ask where the number came from. The answer is usually nowhere.
Third, this is not a substitute for marketing. It multiplies trust that already exists, it works best when you genuinely have satisfied customers, and there is nothing in it that creates the first one. If you are still looking for that first customer, the order is different and I wrote it up in growing without ads, starting from your own network.
And if your version of this is B2B, where you are asking for an introduction to a person inside a company rather than a referral for a product, the structured form of that request is covered by BDG Labs in how to ask for a warm introduction and make it forwardable.
This belongs to tactics, the part of growth that actually gets executed. In my book Growth Engine, though, the tactic comes fourth rather than first, with one rule: one tactic at a time, with a number and a named owner, so that if it works you can repeat it. The detail is in Part 4, Tactics: how you execute. The book is free, and it is in Arabic.