How to build customer engagement that works in Kenya

M-PESA launched in 2007 with one product, Send Money. Last financial year it handled 46.4 billion transactions. What keeps Kenyans coming back to a brand is rarely a clever campaign, and three brands every Kenyan knows show why.

At Tuwele we use a simple definition of customer engagement: the repeated contact a customer chooses to have with a brand over time, held together by a clear brand strategy. Likes and impressions are easy to count. One customer who returns without being asked is worth more than a thousand of them.

We looked at three brands every Kenyan knows: Safaricom, Java House and KFC Kenya. They engage customers in very different ways, and together they make one point clearly. In Kenya, engagement works when a brand proves it belongs here.

Kenya’s engagement channels are different

Kenya has 23.9 million internet users, about 40.6% of the population, but 85.5 million mobile connections, according to DataReportal’s Digital 2027 report. The Communications Authority counted 54.0 million active mobile money subscriptions in June 2026, more than one for every person in the country. WhatsApp is the everyday channel: 72.5% of internet users opened it in the past month.

X is small here, with an ad reach of about 2 million adults, yet it carries outsized weight. A study by FSD Kenya and Princeton found that Kenyans on Twitter (#KOT) used it to escalate complaints after hotlines had failed them. Most engagement advice is written for American and European markets and treats customer service as a separate subject. In Kenya, service is where a lot of engagement is won or lost, in public.

Safaricom built a habit out of small payments

Plenty of apps chase daily use with streaks and badges. M-PESA needs neither. In the year to March 2026 it processed 46.4 billion transactions worth KES 41.68 trillion, and 17.1 billion of them were kadogo payments, the small everyday amounts that pay for lunch, fare and airtime. Merchant acceptance grew to 3.1 million, including 2.1 million Pochi la Biashara tills for informal sellers.

Bonga Points, launched in January 2007, shows both sides of loyalty. During COVID, the Bonga for Good campaign raised the value of each point by 50% and let customers spend points at any Lipa na M-PESA merchant. In two months, more than 600,000 customers redeemed over a billion points, worth about KES 300 million, much of it on food. The other side came later. When a three-year expiry rule took effect in January 2023, the value of unredeemed points on Safaricom’s books fell 34.2%, from KES 4.51 billion to KES 2.97 billion. Billions of shillings in rewards had sat idle until the rule forced the question, which suggests the points were doing very little to bring people back.

Hakikisha: one screen, 8,000 fewer complaints a day

Before October 2015, requests to reverse money sent to the wrong number made up 60 to 70% of calls to Safaricom’s call centre, roughly 12,000 a day. Hakikisha showed the recipient’s name before a transfer went through, and later added a simple Yes or No prompt. Daily reversal requests fell from 12,000 to 4,000, a 67% drop. Name lookups are capped at five a day so nobody can use the feature to identify strangers.

This is our favourite fact in the set. A screen most customers barely notice took away the moment that made them angriest, and most of the calls went with it.

Safaricom’s newest layer is My OneApp, which merged the M-PESA and MySafaricom apps on 1 April 2026. It has 6.5 million 30-day users and more than 80 mini apps, and Safaricom’s Esther Waititu says “the more you use the app, the more it learns you.” Personalisation like that comes with duties. In May 2026 the High Court found Safaricom liable for a leak of data on about 11.5 million subscribers and awarded KES 900,000 to each of 11 petitioners.

Java House chose to become a Kenyan brand

Java opened in 1999 at Adam’s Arcade on Ngong Road, started by American former relief workers including Kevin Ashley and Jon Wagner. In the early years about two thirds of its customers were expatriates. Ashley later said the company set out on purpose to become a Kenyan brand, and that “it took us 10 years to grow to 10 branches.” In July 2026 it opened its 110th outlet, in Thindigua, Kiambu. Many of the newer sites are small Java Express counters at petrol stations, placed on the routes people already drive.

Java’s engagement is mostly physical: one consistent experience, placed where people already are. Its CEO Priscilla Gathungu, the company’s first Kenyan CEO, describes the cafés as places where people “connect, work, and unwind”. The brand has been through four ownership changes in 12 years, and it kept opening cafés through all of them.

Its app is modest: points per visit, buy one get one free deals, digital punch cards and a birthday treat. A punch card is the simplest loyalty tool a café can run, and it is where we usually tell small businesses to start. Java runs one across more than a hundred outlets. The more memorable engagement sits in the supply chain. The honey in Java’s Dawa comes from Sesia, a supplier since 2015 that moved from 5 litre jerricans to 50ml cups to serve Java. Every February, under “Java Love is Kenya Love”, Java profiles suppliers like Sesia, and more than 90% of what it sells is sourced locally. Since March 2024 customers have been able to donate in store to World Food Programme school meals.

KFC Kenya: the day the chips ran out

KFC opened at The Junction on 11 August 2011, the first American fast food chain in East Africa, and some customers queued for up to 90 minutes. A decade later it learned how quickly that goodwill can turn.

In January 2022 KFC Kenya tweeted that customers had loved its chips so much that it had run out, and offered swaps that put ugali on the KFC menu for the first time. Then Kenyans found out the fries were imported, partly from Egypt, while local potato farmers had more potatoes than they could sell. By 11am #BoycottKFC was the top trend in the country, and users joked that KFC should “import customers as well.” Burger King said it had enough fries for everyone. KFC’s regional CEO said local suppliers were welcome if they met its specifications, which he called “proprietary”.

Kenyans would probably have forgiven the shortage. The anger came from a cheeky tone on a bad day, and from the sense that the brand had looked past the farmers around it.

Kentucky Town 254 turned buyers into guests

KFC’s latest campaign reads like a brand that took the lesson. Kentucky Town 254 ran from 20 October to 15 November 2026 with seven limited items in mandazi, samosa, masala and choma flavours. A meal receipt was the ticket to four weekend pop-ups in malls, with music, games and merch, and there was a hand-printed apparel line made with the Nairobi label home_254. Kuku Foods’ CMO Dorothy Muthoni said the team didn’t want to simply put a Kenyan name on an imported idea.

Good campaigns turn an audience into participants, and a receipt that doubles as an invitation does exactly that. KFC now has 48 restaurants in Kenya, its second largest market in Africa.

The common thread: M-PESA’s kadogo economy, Java’s choice to become a Kenyan brand and KFC’s local makers all earned engagement by proving they belong here. The chips crisis and Safaricom’s data leak show how fast that is lost when a brand seems to ignore the place it trades in.

How to build engagement in Kenya

We plan engagement in a five step loop: set one goal, learn from customers, choose a mechanic, test it small, then measure. In Kenya it needs a few local adjustments.

  • Start with one goal, one mechanic and one number to track. Safaricom reports 30-day active users, point redemptions and call-centre volumes, and those are more useful than a generic satisfaction score.
  • Look for the moment that makes customers angry and design it away. Hakikisha cut two thirds of a call-centre problem with one extra screen.
  • Give rewards people will actually spend. A billion points moved in two months when they could buy food.
  • Plan for #KOT before you need to. Write down how the brand will speak on a bad day, because the tone that works on a good day can read as careless during a shortage.
  • Work offline too. About 60% of Kenyans aren’t online, and Bonga still signs people up through *126#.

If you are planning engagement for a Kenyan brand, start where your customers already pay, meet and complain, and make the brand useful in those places first.

Have a project in mind?

Tell us what your website needs to do and when you need it. We will come back with questions, a rough plan and the next step.


AI website builders or a custom build: which one does your business need?

An AI builder can put a homepage on your screen before your tea goes cold. Whether it should be your website depends on what the site has to do six months after launch.

Typing a description of your business and getting back a finished-looking site is ordinary now. Framer, Wix, Squarespace and Hostinger all have AI tools that draft the pages, the layout and the words. A newer group, including Lovable and Bolt, goes further and writes the code for an app from a chat.

Speed is no longer the hard part. What decides the choice is what the site has to carry once it’s live: payments, bookings, a team that updates it every week, a CRM, customer data. A builder handles some of these well and makes others expensive to add later.

What the builders do well

They are fast and cheap to start. For a launch page, an event, a pop-up shop or a first site while you test an idea, a builder gets you online in an afternoon, and you can change the words yourself.

They are also quicker than many people assume. In HTTP Archive’s July 2025 Core Web Vitals report, 73% of Wix sites and 69% of Squarespace sites passed Google’s speed and stability checks, against 44% of WordPress sites. A hosted builder handles the hosting, the images and the updates for you, and most small sites never outgrow that.

Where the cost moves

The first price you see is usually the cheapest you’ll pay. Hostinger’s builder, for example, is advertised from about $3 a month on a four-year contract and costs around $12 month to month. Add-ons for bookings, extra team members or a bigger shop are billed on top, every month, for as long as the site exists.

The bigger cost is leaving. A builder site is rented, and moving it to another platform can mean building it again. Wix can export code only for sites made in its newer Studio editor, as a one-time snapshot that is still in alpha. Webflow’s code export leaves out CMS content, user accounts and e-commerce, so a blog or shop comes out empty. Your words and records can usually be downloaded. The design and the way it all fits together stay behind.

The Kenyan checklist

Most builder reviews are written for American and European businesses, and they skip a few things that matter in Kenya.

  • Builders take M-PESA and other mobile money through payment partners. Wix, for instance, lists Paystack and Peach Payments for Kenya. Check the fees, how long payouts take, and what happens when a payment arrives late or fails.
  • Kenya’s Data Protection Act 2019 requires organisations that process personal data to register with the Data Protection Commissioner once they meet its thresholds. A contact form or booking form collects personal data, so know where it is stored and who can see it.
  • Kenya had 27.4 million internet users at the start of 2025, and most of them reach the web on a phone. Test every page, form and payment on a mid-range Android phone on mobile data as well as on a laptop.

AI that writes code is a different deal

Tools like Lovable and Bolt produce real code, which you can download and keep. That solves the ownership problem and brings a new one: the code is now yours to check. In 2025 a security researcher scanned 1,645 apps built with Lovable and found critical flaws in 170 of them, mostly databases left open because access rules were never switched on, even though the apps themselves ran fine.

Generated code is a fast first draft. If the site holds customer details or takes money, someone who can read the code should review it before launch and keep it updated after.

Before you pick a tool, write down what the site must still be doing in two years: who updates it, what it connects to and what data it keeps. Then ask each option how it handles those three things.

Custom work is getting faster too

The choice is less all-or-nothing than it was. Studios use AI to draft layouts, copy and code, which shortens the early stages of a custom project. Some businesses start on a builder for a launch page and plan a custom site once they know what customers use. If you start on a builder, plan the move to a custom site before you need it, while it is still cheap to make.

How to decide

Five questions will usually settle it:

  • Will more than one person or team update the site every week?
  • Does it take payments, bookings or orders?
  • Does it need to connect to other systems, such as M-PESA, a CRM or your stock records?
  • Do you need to own the code and the data, and be able to move them?
  • Will this still be your main website in two years?

If most of your answers are no, a builder is a sensible place to start. If two or more are yes, talk to a studio before you build, because those are the parts that cost the most to fix later.

Have a project in mind?

Tell us what your website needs to do and when you need it. We will come back with questions, a rough plan and the next step.


What does a website actually cost in Kenya?

Two quotes for “a website” can be a hundred times apart. Here’s where the money goes, and how to compare the quotes in front of you.

It’s the first question almost every client asks us. Prices published by Kenyan web hosts and agencies start at around KSh 10,000 for a one-page site and go past KSh 1 million for a large online shop, and both get called “a website”. Most of the difference sits in work you can’t see on the screen.

Take two safari lodges that each want a homepage, a rooms page, a gallery and a contact form. The first takes bookings by phone. The second wants guests to choose dates, see which tents are free and pay a deposit by M-PESA. The two sites can look almost the same, but the second needs a booking calendar, a payment connection, confirmation emails and far more testing, so it will cost several times as much.

What you’re paying for

A website is a set of decisions: how many pages, what each one has to do, how your team will update it and what it connects to. The more of those you’ve settled before you brief a studio, the faster and cheaper the project goes.

  • Planning and content, meaning who the site is for, what they should do next, and the words and photos that get them there.
  • Design, which can be a ready-made template set in your colours or a layout drawn for your brand. A template is fine for a first site.
  • The build. A few fixed pages are quick to make, while bookings, accounts, a shop or a members’ area each bring their own screens, rules and testing.
  • Connections to other systems, such as M-PESA, card payments, email, maps or your CRM. Each one is a separate piece of work.
  • Care after launch, which covers hosting, security updates, backups and small changes.

M-PESA is a good example of a connection that looks small. To take payments on a website through Safaricom’s Daraja API, a business needs its own Paybill or Till number, and the site has to wait for Safaricom to confirm each payment before it marks an order as paid. That confirmation arrives separately from the checkout, so the build has to cope with payments that come in late, get cancelled or never arrive.

Three common shapes of project

Most projects fall into one of three shapes, and each step up adds design, build and testing time.

A brochure site has a handful of pages that say who you are, what you offer and how to reach you. It suits a consultancy or a new business that mainly needs to look credible when someone searches for it.

A content-led site is bigger, with a blog or case studies and a content management system (CMS) your team uses to publish without calling a developer. Here the money goes into organising the content and setting up the CMS so it’s easy to use.

A platform is a product people log in to, to book, pay or manage something. Every feature needs its own screens and its own plan for when things go wrong, and that’s where most of the testing time goes.

Before you ask for quotes, write down three things: what the site must do, who it is for, and what a good result looks like in six months. Every studio can quote better from that page.

The costs people forget

Some costs come back every year. Domain names renew annually, and KENIC raised its .ke domain fees in 2025. Registrars often sell the first year for a few hundred shillings and charge KSh 1,000 or more to renew, so check the renewal price. Make sure the domain is registered in your business’s name, so you keep it if you change developers.

Hosting and business email are paid monthly or yearly as well, and email is usually priced per person. Photography and copywriting are often needed and rarely included. If a quote says “content supplied by client”, plan the time or money to produce it.

A site that nobody looks after also gets slower and less secure over time, as its software falls behind on updates. Budget for a small monthly care plan.

How to compare quotes

When two quotes are far apart, ask each studio the same questions:

  • Which pages and features are included, and how many rounds of changes?
  • Who writes the words and supplies the photos?
  • Who owns the domain, the hosting account and the code when the project ends?
  • What happens after launch, and what does it cost each month?
  • How will my team update the site, and will someone show them how?

Quotes that answer all five are much easier to put side by side, whichever studio you choose.

Have a project in mind?

Tell us what your website needs to do and when you need it. We will come back with questions, a rough plan and the next step.