How do I launch my startup?

By Bonaventure Ogeto, Founder and CEO, Mctaba Labs
Published

The short answer

Launch in stages, not in one big day. Open quietly to the ten or twenty people you already know want it, fix what breaks, then widen to their referrals and only then announce in public. On day one, three things must work: taking payment, a support line a person answers, and a way to hear complaints.

Picture the launch most founders plan. A date circled months ahead, a poster with a countdown, a room booked, a cake with the logo on it. At 10am the link goes out to everyone at once. By 10.20 the first person has tried to pay, the M-Pesa prompt has not arrived, and the founder is reading a row of messages that all say "it is not working".

A launch like that puts your largest audience in front of your least tested product. The safer way is duller and works better: launch three times, each time to a slightly bigger group, and fix what breaks in between.

The example here is a service that lets parents in Kisumu book a vetted home tutor and pay per lesson. The numbers used are round and invented.

Is a launch one day or a sequence?

Treat it as a sequence of three stages.

Stage Who you open to How many What you are trying to learn Move on when
1. Quiet People who already said they want it, by personal message 10 to 20 Does the whole journey work, from first message to payment to delivery? Several people have paid and used it without your help
2. Widening Their referrals, one or two WhatsApp groups, one partner 50 to 100 Do strangers understand it and trust it? What do they ask? New people sign up and pay without a call from you
3. Public Everyone you can reach As many as you can serve Which channel brings paying customers? You never stop this stage. It becomes ordinary marketing

The quiet stage may last a week or a month. There is no prize for speed. The only rule is that you do not widen while the same fault keeps appearing.

Two things make the quiet stage work. First, tell people it is early: "You are among the first twenty families. Things may go wrong, and I will fix them personally." Early users who know this become allies. Second, charge the real price, or a stated early price. A quiet launch with no payments tests nothing that matters.

What must work on day one?

Three things. If any one of them is missing, postpone.

1. Taking payment. A customer must be able to pay in the way they normally pay, which for most Kenyan customers means M-Pesa, and get a confirmation within seconds. Test it with real money on at least two different phones, including a cheap one on a slow network. Test the awkward cases as well: a wrong PIN, a cancelled prompt, a payment made twice. Know what you will do when a payment leaves the customer's phone and does not show in your system, because that will happen in the first month.

2. A support line that a person answers. One phone number or WhatsApp line, shown on every screen and in every message, with stated hours. During launch week the person answering should be you. No form, no "we will get back to you in 48 hours".

3. A way to hear complaints. An unhappy customer will often leave without telling you. Make it easy and ask directly. Message every new customer after their first use: "How did the first lesson go? What should we fix?" Keep every answer in one sheet.

Add a fourth for your own sake: a manual fallback. If booking fails, you can take the booking on WhatsApp and enter it yourself. If payment fails, you can accept it to a Till and mark it paid by hand. A launch with a fallback bends where one without it breaks.

Most other things can wait. A polished logo, an app in the app stores, a blog, a referral scheme and a dashboard for yourself are all pleasant and none of them stops a customer buying on day one.

What should I do in launch week?

Use this checklist for the quiet stage. Repeat a shorter version each time you widen.

The week before

  • Go through the full journey yourself as a new customer, on a phone, on mobile data.
  • Ask two people who have never seen the product to do the same while you watch in silence. Write down every place they hesitate.
  • Make a real payment and a real refund.
  • Write your list of the first 20 people, by name.
  • Write the announcement message and the answers to the five questions you expect.
  • Set up the support line and tell whoever is helping you when they are on duty.
  • Decide your three numbers for the week, for example 20 invited, 10 signed up, 5 paid.

Day 1

  • Send the first ten personal messages in the morning, not all twenty.
  • Keep the day clear. Answer every message within minutes.
  • Note every problem in the sheet, even small ones.

Days 2 and 3

  • Fix the top fault from day one before inviting anyone else.
  • Send the next ten messages.
  • Call the people who signed up and did not pay. Ask what stopped them.

Days 4 and 5

  • Message everyone who has used the product and ask how it went.
  • Ask each satisfied customer for two introductions.
  • Collect one or two short comments you have permission to share.

Days 6 and 7

  • Count your three numbers against the target.
  • List the faults by how often they occurred. Fix the most common.
  • Decide: widen next week, or stay quiet for another week.

How do I announce it when I widen?

Through people your customers already listen to, in words that are about the customer.

A weak announcement says "We are excited to announce the launch of our platform". A strong one says who it is for, what it does and what to do next. For a WhatsApp group, with the admin's permission:

Good evening parents. Many of us have struggled to find a reliable tutor for Grade 7 and 8 maths. I have started a small service in Kisumu that matches families with tutors whose certificates and references I have checked myself. You book a lesson on WhatsApp and pay per lesson on M-Pesa after the first session is confirmed. Twelve families are using it so far. If you would like a tutor this term, message me on 07XX XXX XXX and I will call you.

It names the reader, gives proof, removes the fear of paying before seeing anything, and has one action. Ask two of your early customers to reply under the post with their experience. A message from the founder is advertising. A reply from another parent is evidence.

Then use the same message, adjusted, in each place you widen: your Status, your early customers' Status, a partner such as a school or an estate committee, and a notice where your customers physically are.

What should I measure in the first 30 days?

A handful of counts, written down every week. They follow the customer's journey, so you can see exactly where people drop off.

What to count Example for the tutor service What a low number tells you
People who heard about it Messages sent, group members reached You need more reach or a better place to announce
People who enquired Parents who messaged or signed up The announcement is unclear or aimed at the wrong people
People who paid once Parents who paid for a first lesson Trust, price or the payment step is the problem
People who got the result First lessons that took place Something in your delivery is breaking
People who came back Parents who booked a second lesson The product is not good enough yet
Complaints, by type Tutor late, payment not reflected, and so on Your repair list, in order

Say that in the first 30 days 200 parents heard about the service, 40 enquired, 15 paid for a first lesson and 12 booked a second. The gap worth working on is between 40 and 15. Twenty-five parents were interested enough to ask and did not pay. Call ten of them and you will know what to change. The gap between 15 and 12 is small, which tells you the lessons themselves are good.

Record where every customer came from as well. By day 30 you should be able to say which one or two sources brought most of the paying customers.

What if launch day is quiet?

Then you have learned something cheaply, and nobody was watching.

A quiet first stage usually has one of three causes. The people you invited did not have the problem badly enough. The message did not make clear what they would get. Or a step in the journey, often payment or sign-up, lost them. Phone five people who did not respond and ask. Do not answer silence with a bigger announcement. A wider audience for something that is not working gives you the same result with more witnesses.

What does a staged launch change about the product you build?

It lets you build less before you open. For the quiet stage you need one job done end to end, payment that works, and a way to reach you. Bookings can be confirmed by hand. Reports can be a spreadsheet.

It also tells you what to build for stage two. The questions people ask in week one become the text on your sign-up screen. The tasks you repeat by hand become features. The payment cases that went wrong become rules in the system. By the time you announce in public, the product has been shaped by the people who paid for it.

Key takeaways

  • A launch is a sequence of three stages, each bigger than the last.
  • Go quiet first. Let people who already trust you find the faults.
  • Payment, a support line and a way to hear complaints must work on day one. Most other things can wait.
  • Test the full journey with real money on a real phone before anyone else does.
  • Announce through people and groups your customers already listen to.
  • For 30 days, count first payments, completed first uses, returns and complaints by type.

Frequently asked questions

Should I hold a launch event?
Not for a first product. An event costs money and effort, fills a room with friends and other founders, and rarely brings customers. Spend the same day visiting or calling twenty people who have the problem. Hold an event later, if ever, when you have customers to invite.
Do I need press coverage when I launch?
No. An article may bring a short burst of curious visitors, and few of them will be your customers. It helps later, when a buyer searches your name and wants proof that you are real. In the first month, a recommendation in the right WhatsApp group is worth more.
What is the best day to launch?
The day before your customers most need the product. A school fees tool should be live before term opens, and a payroll tool before month end. Avoid days when your own team cannot answer the phone, such as a public holiday or the day before you travel.
What if something breaks on launch day?
Something will. Tell affected customers quickly and plainly, give a time by which it will be fixed, and offer a manual way to finish what they came to do. People forgive a fault that is handled well far more easily than silence.
Can I launch before the product is finished?
Yes, and you should. Launch when one job works from beginning to end, including payment, even if the rest is missing. A finished product does not exist. Waiting for it only delays the day you learn what customers think.

Sources

  1. The Lean Startup: How Today's Entrepreneurs Use Continuous Innovation to Create Radically Successful Businesses, Eric Ries (book, 2011). Checked on 19 November 2026.
  2. Traction: How Any Startup Can Achieve Explosive Customer Growth, Gabriel Weinberg and Justin Mares (book, 2015). Checked on 19 November 2026.