A Google review costs your customer about two minutes.
A referral costs them their reputation. They are telling somebody they care about that you will do a good job, and if you don’t, they wear it at the next family dinner.
That one difference explains the referral program you set up in March with a generous reward that nobody has claimed. You have been solving an incentive problem. You have an asking problem, a timing problem, and a reward structure that quietly makes your best customers look like salespeople.
Three things to change. Ask later than feels natural, ask specifically rather than generally, and give the friend something rather than only the referrer. The reasons are more useful than the tactics, so let’s do the reasons.
The referral moment is not the review moment
I’ve written before that the time to ask for a review is immediately after the thing went well, while the customer can still feel it. That’s right for reviews. It is wrong for referrals, and I want to be clear that I’m contradicting the earlier advice on purpose.
A review is a report on something that already happened. A referral is a prediction that you’ll do it again, for somebody specific, whose opinion of the referrer is on the line.
Nobody wants to make that prediction on one data point.
So the moment of real referral willingness is later. After the second job. After the thing you fixed in January was still fixed in June. Above all, after they have watched you handle something going wrong, because that is when a customer learns what you’re actually like, and it is the only evidence that makes a person comfortable vouching for you.
Which gives you a way to build the list, and it takes twenty minutes in whatever system holds your jobs.
Pull the last twelve months. Find every customer with two or more completed jobs, plus every customer with one job and a follow-up problem you resolved well. That is your referral list. It is a lot shorter than your customer list, and it is the part that will actually refer.
Everybody else is a review request, not a referral request.
Why “do you know anyone?” gets you nothing
Here’s the mechanical reason the general ask fails, and it has nothing to do with willingness.
“Do you know anybody who needs us?” asks your customer to run a search across their entire social graph, with no index and no prompt. That is genuine cognitive work, performed on your behalf, for free, while they’re trying to get on with their day.
So they say they’ll keep you in mind.
“I’ll keep you in mind” is a no with better manners.
The specific ask works because you do the retrieval for them. You already know things about your customers. They mentioned a sister buying a place across town, or a neighbor who complained about the same problem, or that their office is moving in the spring.
Two sentences, in whatever channel you already use with that person, from the person who did the work:
“You mentioned your sister just bought the place on Oak Street. If she needs anyone for the same work, I’d be glad to look after her, and you can hand her my number or I can call her, whichever you prefer.”
That’s it. A named person or a named situation, and an explicit offer to take the awkward part off their hands. Giving them the choice between passing your number and you making the call matters more than it sounds, because it lets them control how much of their social capital they spend.
The reward, and who should get it
The consensus is that double-sided rewards, where the friend gets something too, beat rewarding only the referrer. I think the direction is right.
I’d be careful with the numbers, though. Every figure I can find comes from companies that sell referral software, and they disagree with each other by an order of magnitude. One says double-sided produces 2.3 times the shares, another claims a 300% to 500% lift, another reports completion rates of 52% against 29%. Those cannot all describe the same effect. Treat the direction as the finding and your own results as the evidence.
The mechanism is the part that survives scrutiny, and it’s the reason I’d design this way regardless of the figures.
A reward paid only to the referrer turns your customer into a salesperson who got paid to make an introduction. That is exactly the thing they are afraid of looking like, and it is why they hesitate even when they like you.
Give the friend something and the whole social transaction changes. Now your customer is arriving bearing a gift. They are doing their friend a favor rather than collecting a commission, and the thing they were protecting is no longer at risk.
Two practical notes. The two sides don’t have to match, and the friend’s side can be the richer one, because it doubles as a first-purchase promotion. And keep it simple: a clear amount or a clear free thing beats points, tiers and escalating bonuses, which all read as a scheme.
The tracking, or this dies without anybody noticing
Can you answer how many referrals you got last quarter, and from whom, by name?
If not, the program is already dead and nobody has told you, because referral programs don’t fail loudly. They just stop being mentioned.
The minimum is a field on the job record for referral source that holds an actual person’s name, not a dropdown with “word of mouth” in it. “Word of mouth” is where referral data goes to be useless. You cannot thank a category.
Then thank the referrer inside a week, whether or not the job closed. This is the part everybody skips and it’s the part that compounds. The reward buys the first referral. The acknowledgement is what produces the second and third, because people repeat behavior that gets noticed and quietly stop behavior that vanishes into a void.
Don’t bundle this with your reviews
One boundary worth respecting, because the two programs feel adjacent and are not.
The FTC’s rule on consumer reviews and testimonials does not ban offering incentives for reviews outright. What it restricts is conditioning the incentive on sentiment, whether that condition is express or implied, and it treats incentivized reviews as testimonials, which makes them advertising. Material connections between a reviewer and the business have to be disclosed.
So the practical line: don’t run “refer a friend and leave us a review” as a single incentivized ask, don’t pay more for a better review, and keep the two things in separate programs with separate records.
Whether a particular referral reward creates a connection you need to disclose depends on how you’ve built it, and that is a question for counsel rather than for me.
Before lunch
Twenty minutes, three steps.
Add the referral source field to your job record now, as a text field, and start filling it in from memory for the last few jobs while you still remember.
Pull your list of repeat and well-recovered customers.
Then pick the single one you’d most like another job like, and send those two sentences about a person you already know they know.
If you’d like help working out where your own referrals have actually been coming from, drop me a line on email, WhatsApp, phone or LinkedIn and we can have a quick chat, and my notes on lead generation cover the paid side once this is working. I’m taking on a small number of contracts at the moment, so I’ll tell you straight whether it’s something I could help with.