Telling people to spread the word produces nothing. Ask one happy customer to think of one person, in the week after delivery, and give them something to send.
Sailo team12 min read
You've written "please share with anyone who might like it" at the end of 11 posts now. Nobody has shared anything. The 2 referrals you've ever had both came from your sister, unprompted.
The problem isn't that your customers don't like you. It's that "share with anyone" is a request with no person in it, no deadline, and nothing to send.
A referral ask that works has three parts: it goes to one specific customer, it asks them to think of one specific person, and it arrives in the week after their thing turned up. One seller in Nairobi ran 27 of those asks over 8 weeks and got 4 sales at KSh 2,800 each out of it, from about two hours of typing. Ask the vague way instead and a hundred asks produce nothing, which is the rate most sellers are currently running at.
This is also the cheapest way to reach a stranger you'll ever have, which is why it belongs in the same week as getting your first ten orders rather than in some later growth phase. Referrals are how orders eight, nine and ten usually arrive.
Put yourself on the other side of it. You bought sandals, you like them, and the seller says "tell your friends!"
Now what? You'd have to scan your entire social circle for someone who might want sandals, work out whether recommending them makes you look odd, decide on a channel, and write something. That's four decisions and about ten minutes, in exchange for nothing, at a moment when you weren't planning to do any work at all.
So you say "will do!" and you don't.
Every part of the fix is about removing those decisions. Name the person for them, or narrow the category so far that a name pops into their head. Give them something to send so there's nothing to write. And ask at a moment when they're already thinking about you, which is a much shorter window than you'd guess.
Ask on the day it arrives, not the day they paid.
The gap between "it came and it's lovely" and "this is just a thing I own now" is about five days. Inside that window they're still slightly delighted and they've probably already mentioned it to someone. After it, they will not think about you again unless something breaks.
This is the single most common mistake and it's a scheduling error, not a wording one. Most sellers ask for referrals in a batch, months later, when they need sales. That's precisely the moment when the customer has the least to say about you.
So: message them when the delivery confirms, or when they send you the "it's here!" text. If they don't say anything, message on day three anyway to check it arrived, and put the ask in the reply to whatever they say back. Never in the same message as the check-in, because then the check-in reads as a pretext, and people can smell that.
Compare these three asks.
If you know anyone who'd like these, send them my way.
Nothing happens. No person, no verb, no timeframe.
Do you know anyone who's been complaining about not finding sandals in wide sizes?
Better. It's a specific problem, so a specific face might surface.
This is a weird question, but is there one person in your office who'd wear these? I'm trying to get a few more customers before I order the next batch of leather.
This one works. It asks for exactly one person, it narrows the search to a place they can picture, and it gives a reason that's true and mildly urgent without being pushy. "I'm trying to get to X before Y" is honest, it makes the ask feel like a favour rather than a marketing tactic, and people help small businesses they like when there's a concrete thing to help with.
The reason "one person" beats "anyone" is that the brain answers questions it can answer. "Anyone" is a search of everyone you know. "One person in your office" is a search of about nine people and it terminates.
After a delivery, in a chat thread:
So glad they fit. Can I ask you something cheeky? I'm trying to get to 40 customers before I order more leather, currently on 31. Is there one person you can think of who'd actually wear these? I'll send you the link and a photo so you don't have to write anything.
After someone leaves a compliment publicly:
Thank you! Would you mind if I sent you the link to forward to one person? No pressure at all if not.
After a repeat order, which is your strongest signal:
That's your third pair now, which is honestly the nicest thing that's happened this month. If there's someone you'd send these to, I'd rather one of your recommendations than any advert.
Notice what none of them do. None of them offer a discount as the opening move. None of them say "share on your stories". And none of them ask twice.
This is where most asks die even after a yes. They said they would, they meant it, and then they had to compose a message and it never happened.
Send them a ready-made forward. Three things in one block, so all they do is hold and forward:
Write it as though they wrote it. If it reads like an advert, they won't send it, because forwarding an advert makes them look like they're doing marketing for someone.
The link matters more than people think. One link that opens straight onto the product, with a price on the screen, is forwardable. A link that opens a list of buttons and makes the recipient hunt is not, and it's where forwarded traffic goes to die. Every extra tap costs you more on a forwarded link than on a normal one, because the person on the other end never chose to be there.
One email on pricing, photographs, delivery and getting paid. No pitch, no filler.
Wanjiru makes beaded leather sandals in Nairobi. A pair is KSh 2,800, delivery inside Nairobi is KSh 300 by boda, and most customers pay to her M-Pesa till, where the registered business name shows before they confirm. That name showing up is worth a surprising amount, because it's the moment a stranger stops worrying.
She'd had 31 customers in five months, almost all from Instagram, and had never had a referral she could identify. She started asking properly in March.
Her method: on the day a delivery confirmed, she messaged to check it fitted. When they replied, she asked the "one person in your office" version and offered to send a forward pack.
Over eight weeks: 27 asks, 19 replies, 11 people said yes and asked for the pack, 6 actually forwarded it that she knows of, and 4 sales came back. She only knows about the 6 and the 4 because she started asking every new customer "who sent you?" at the point of order, which is a question you have to ask deliberately, because on a manual payment rail nothing tells you.
Four sales at KSh 2,800 is KSh 11,200 from about two hours of messaging spread over two months. Her cost per referred customer was zero, and every one of the four converted from the first message, because a recommendation from someone they know does the work that ten posts don't.
What surprised her: three of the four referrals came from her cheapest customers, not her most enthusiastic ones. The woman who bought the most pairs referred nobody, and the woman who bought one pair on sale and had been slightly awkward about the delivery fee referred two people. Enthusiasm in the chat does not predict who talks about you in real life, so ask everybody rather than the people you think will say yes.
What didn't work: she tried offering KSh 300 off the next pair for a referral. Nobody took it, and two people said it made the ask feel like a transaction. She dropped it and the referrals went up.
Sometimes, and less often than you'd think. Work out the number before you decide.
Say your product is KSh 2,800 and it costs you KSh 1,500 to make and deliver. Your margin is KSh 1,300.
| Referral reward | You give | Margin left on that sale | Worth it? |
|---|---|---|---|
| Nothing | KSh 0 | KSh 1,300 | Usually yes |
| KSh 300 off referrer's next order | KSh 300 later | KSh 1,300 now, 1,000 later | Marginal |
| 10% off for the new buyer | KSh 280 | KSh 1,020 | Yes, if it closes them |
| 10% both sides | KSh 560 | KSh 740 | Only if repeat rate is high |
| Free product for 3 referrals | KSh 1,500 | Roughly one sale in three | Rarely |
Two rules come out of that table. Reward the new buyer rather than the referrer where you can, because a discount for the newcomer is a closing tool and a discount for the referrer is a bribe to a friend. And never give away more than about a quarter of your margin, because a referral programme that makes referred customers unprofitable is a very efficient way to grow yourself broke.
The full arithmetic on when a discount pays for itself and when it quietly doesn't is in how to run a discount without losing money.
There's a cultural point too. In a lot of markets, paying someone to recommend a friend reads badly, and Wanjiru's two customers who said so were being polite about it. Money changes a favour into a job. Test it on three people before you announce it to everyone.
Anyone whose order went wrong, even if you fixed it. They'll say yes and never do it, and you've spent goodwill you need for something else.
Anyone you've asked in the last three months. One ask per customer per quarter, maximum. Asking again turns a person who likes you into a person avoiding your messages, and how to follow up without being annoying is worth reading on where that line sits.
People who haven't received the thing yet. Obvious, and yet people do it in the order confirmation.
Everyone at once. A broadcast referral ask converts at approximately zero and signals that none of the individual messages you send are individual.
It will, eventually, and it costs more than a normal complaint because two relationships are on the line.
The referred customer had expectations set by a friend, which are higher than the ones your listing sets. If the parcel is late or the colour is off, they're disappointed and slightly embarrassed, and the person who referred them is more embarrassed still.
Handle it in that order: fix it for the new customer fast and without argument, then message the referrer before they hear about it. "Just so you know, Achieng's pair went out a size small, I've already sent the right one and she's fine, but I didn't want you finding out from her." That message costs you nothing and it's the difference between someone who never refers again and someone who refers twice as hard.
Sellers who skip that second message lose the referrer permanently and usually never work out why.
Ask "how did you hear about me?" at the point of order. Not in a form. In the conversation, as a normal question, because that's where the answer is honest.
Write it in your customer list next to the order. Over three months that column tells you which four customers are actually generating your growth, and those four are the people who get the first look at the next batch, the free samples, and the message when something new lands. Not because it's a loyalty scheme, but because they're doing your marketing and you should know their names.
This is where Sailo's limits are worth stating rather than glossing. There is an affiliate feature, but it lives on the Business plan at $49 a month, alongside broadcasts and memberships. Below that plan, referral tracking is you and a spreadsheet column, and honestly at 31 customers a spreadsheet column is more accurate anyway, because most of your referrals arrive as a forwarded WhatsApp message that no link tracker will ever see.
The other limit is worth repeating: on a manual rail, Sailo cannot tell you that money arrived and it cannot tell you where a buyer came from. Both of those are your job, and both are one question each.
Not a campaign. Three names.
Open your order list, find the three most recent people whose order arrived in the last week, and message each of them individually. Check the thing fitted or arrived properly first. When they reply, ask the one-person version, offer to send the forward pack, and send it within the hour while they're still in the thread.
Then write "who sent you?" into your order process so that the next time this works, you know it worked.
If the same customer keeps referring people, you've found something better than a marketing channel, and turning one customer into three is about what to do with them. And if the people you're asking haven't bought in months, start earlier: how to get repeat buyers comes before this, because a customer who's already come back twice refers at a completely different rate to one who bought once in April.
Written by
Sailo team
One link, your whole shop.
One email on pricing, photographs, delivery and getting paid. No pitch, no filler.
Sailo just gives it a front door — so people can browse, compare and see prices before they message you.
Get your linkFree while in beta · No card required