Referrals
How to ask for referrals: the moment, the question, the reward
Most satisfied customers never refer anyone, not out of reluctance but because nobody asks them at the moment they would have said yes. Ask within 24 hours of a visible result, and ask about a person rather than for a favour: "Can you think of anyone who needs the same thing right now?" Then ask for the introduction instead of the name, so your customer sends the first message. That single change lifts the response rate and removes the legal question of whether you were allowed to contact a stranger. Report back on what happened, and give something from your own work rather than cash.
The gap between willing and doing
Ask a founder where their best customers came from and the answer is almost always the same: someone told them. Ask the same founder what they do to make that happen and the answer is almost always silence. Referrals are treated as weather: pleasant when they arrive, nothing you can do about them.
The numbers say otherwise. A marketing survey by Texas Tech University found that 83 percent of satisfied customers say they are willing to refer, while only 29 percent ever do. The 54 points in between are not people who changed their mind. They are people who were never asked at a moment when they would have said yes.
Satisfied customers
Say they are willing to refer
83 %
Actually do refer someone
29 %
The 54 points in between are the cheapest growth available to you, and the only channel that costs a question instead of a budget.
Figures from a Texas Tech University marketing survey. They describe willingness and behaviour, not a promise about your own customers.
Why a referred customer is worth more
A referral does not just cost less to acquire. The customer who arrives through one behaves differently afterwards. Researchers at Wharton and Goethe University Frankfurt followed close to 10,000 new customers at a large German bank over twelve months and compared the referred ones with everybody else.
What the research found
The honest second half: the margin advantage levels off after roughly two and a half years, as the two groups converge. The loyalty advantage does not. What lasts is not the better deal, it is that the customer was matched to you by someone who knew both sides.
Referral study: Schmitt, Skiera and Van den Bulte. Trust figure: Nielsen's Global Trust in Advertising work, where recommendations from people one knows lead every wave. Averages from large samples, not a forecast for a single business.
The moment: a window of hours, not weeks
The same question produces a name or an empty nod depending on when it lands. Gratitude is at its peak right after the result and falls off fast, because the problem you solved stops being present the moment it stops hurting.
How far the window is open
Right after a visible result
Within 24 hours
The bike rolls again, the delivery arrived on time, the problem is gone. Relief is at its highest and the memory of why they hired you is still fresh. No other moment comes close.
After unprompted praise
Immediately, in the same conversation
When someone says they are happy without being asked, they have already made the recommendation in their head. The question only turns it into a name. Wait and follow up by email later and you start from zero.
After you fixed something that went wrong
One or two days later
The most underrated moment of all. A well-handled failure binds people more tightly than a smooth job, because the customer has seen how you behave when things get uncomfortable.
On the second or third job
At the close, not at the quote
Someone who comes back has chosen you twice. A good moment, but weaker than the first visible success, because by now the work feels ordinary to them.
In a round email to every customer
Whenever
The most common form and the weakest. It asks nobody in particular, names no occasion, and reaches most recipients at a moment when you are the last thing on their mind. It costs little and returns accordingly.
The bars illustrate how open the window is at each point, they are not measured data. The order matters more than the exact width.
What the question is worth in your case
The passive baseline is not zero, which is why founders believe referrals already work for them. It is also flat, which is why they never grow. Set the first slider to zero to see the baseline, then move it and watch what asking adds.
Try it yourself
Move the first slider from left to right. Nothing is stored or transmitted.
At zero you are looking at the passive baseline: what happens when nobody is asked.
Your own estimate. A vague "I'll think about it" does not count as a name.
Typically far higher than for people you approach cold, because someone vouched for you.
New customers from referrals
Nobody is being asked, so this is the baseline every business already has: 1.7 new customers. It costs nothing and it never grows.
The 29 percent baseline comes from the Texas Tech survey and is fixed. The three sliders are your own estimates, not study figures. The result is an order of magnitude to think with, not a forecast.
The question: ask about a person, not for a favour
Almost every weak version of this question shares one flaw: it asks the customer to do something later, in a place you cannot see. The strong versions all do the same thing instead, they ask a question the mind starts answering before it decides whether to reply.
"Feel free to recommend us."
"Can you think of anyone who needs the same thing right now?"
The first is a request with no task in it. The second asks a question the mind answers automatically, before anyone decides whether to reply.
"Do you know anyone who might be interested?"
"Who on your team has the same problem with their bike?"
"Anyone" is too large a space to search. The narrower the circle in the question, the more likely a name comes out of it.
"Would you recommend us?"
"Would you be willing to introduce me?"
The first question gets a yes and ends there. The second asks for an action, which is what makes it visible whether the yes was worth anything.
"Please share us with your network."
"May I send you two sentences you can forward?"
People who want to refer you usually stall on the wording, not on the willingness. Supplying the wording removes the actual obstacle.
Who sends the first message decides everything
This is the step most advice skips, and it is the one that decides both the response rate and, in Europe, whether the contact was lawful at all. A name handed to you does not make that person a contact of yours. Asking for the introduction instead of the name settles both questions at once.
Your customer introduces you
StrongestOne message to both of you, in which your customer says why they are naming you. The third person is replying to someone they know, not to a stranger, and the question of whether you were allowed to contact them never arises.
Your customer forwards your two sentences
Almost as strongYou write the text, your customer sends it under their own name. It removes the work from them while the recommendation still comes from them.
You write yourself, citing the customer
Weaker and riskierThe name in the subject line helps, but it does not turn the message into one that was asked for. A marketing email to a third person is still a marketing email, however warmly you were recommended.
You are given a name and go looking
WeakestWithout an introduction this is cold outreach with a keyword attached. The third person knows nothing of the recommendation and has no way to verify it.
General orientation, not legal advice. Where your recipient sits decides which rules apply to a first message.
The reward: the cheapest one works best
Every reward has the same catch: the more it looks like payment, the less the recommendation is worth to the person receiving it. Sorted from the one that carries furthest to the one that burns out fastest.
| What you give | Effect | The catch |
|---|---|---|
| Nothing, only thanks and a report back | Carries furthest, because the recommendation stays credible | It requires you to come back and say what became of the name. Forget that and you will not be asked a second time. |
| Something from your own work | Strong, because it ties the customer closer instead of paying them off | An extra service, an earlier slot, an extension. It costs you time rather than money and feels like a favour to both sides, not a trade. |
| A discount on the next job | It works, but it permanently lowers your price in their mind | Anyone who gets a referral discount twice expects one the third time. Better once, and with an end date on it. |
| Cash or a bounty | Fastest to work, fastest to wear out | It damages the very thing that made the recommendation valuable: that nobody was paid for it. It is also taxable for the recipient and has to be disclosed to the third person. |
One thing belongs in every row and is missing from most: come back and say what happened. Someone who gave you a name spent their own reputation on you and is now wondering whether that was a good idea. Two sentences settle it, and they are what makes a second name possible.
Five mistakes that keep the names away
None of them is a lack of goodwill. Every one of them is a step that was skipped.
Waiting for the right moment to arrive by itself. It does arrive, several times over, but it lasts hours. Postpone the question to a quieter week and you are asking through a closed door.
Asking for a referral but not for a name. "Feel free to recommend us" gets a nod and nothing else. Only the question about a specific person turns it into an action.
Contacting the third person yourself without an introduction. That is cold outreach with a familiar name in the subject line, and every rule for cold outreach applies to it.
Building a reward scheme before you have ten customers. The effort feels like progress and it replaces the one question that matters. Programmes pay off once the question already works, not before.
Never reporting back. Someone who gives you a name has spent their own reputation and wants to know how it went. A message with the outcome is the cheapest thing you can give in return, and the most commonly forgotten.
Frequently asked questions
When is the best time to ask for a referral?
Within 24 hours of a visible result, or immediately when someone praises you unprompted. Relief is highest and the reason is still present in those moments. The second-best time is one or two days after a mistake you handled well, because the customer has just seen how you behave under pressure.
How do I ask without being pushy?
By asking about a person rather than asking for a favour. "Can you think of anyone who needs the same thing right now?" is a question the mind answers before it judges it. What feels pushy is not the question itself, but repeating it without an occasion and sending it to everyone at once.
Why do satisfied customers so rarely refer on their own?
Because the willingness is there and the occasion is missing. A Texas Tech University survey found that 83 percent of satisfied customers say they are willing to refer, while only 29 percent actually do. The gap is not reluctance; it is that nobody asks and the right moment passes unused.
Can I just contact the person who was recommended to me?
Only within limits. A referral is not consent from the third person. A marketing email to them is still a marketing email, with every rule that applies to one, even though a mutual acquaintance named you. The cleanest and also the most effective route is to ask your customer for the introduction, so that they write, not you.
Should I pay a bounty for referrals?
Early on, probably not. Money works quickly, wears out, and damages the very thing that made the recommendation valuable: that nobody was paid for it. Add to that the tax treatment for the recipient and the duty to disclose the payment to the third person. At small numbers, something from your own work plus an honest report back is more reliable.
Read next
Who would recommend YOUR idea?
The value analyst and the strategist assess how strong the word-of-mouth effect of your specific idea is. 4 to 12 AI experts, in minutes, from € 5.99.
Assess my idea →