Almost every Kenyan gateway comparison concludes that the fees are "roughly comparable" and tells you to pick on support quality. That was lazy when it was written and it is wrong now: the published M-Pesa spread across the four gateways Kenyan merchants actually consider runs from 1.5% to 3.5%. That is a doubling, on every single sale, forever.
Every number below was read off the provider's own pricing page in August 2026, not off another blog. Where a figure comes from secondary reporting rather than a published page, it is labelled as such. Rates change — check before you commit, and if you find one of these stale, the fault is ours.
Skip to: The comparison table · The option nobody sells you · Each gateway in detail · What the spread costs · How to actually choose · Four traps
The comparison table
Published rates, August 2026. Pesapal does not publish a public rate card, so its figures are the ranges consistently reported by Kenyan merchants and industry write-ups — treat them as indicative rather than quoted.
| Provider | M-Pesa | Local card | International card | Settlement | Setup / monthly |
|---|---|---|---|---|---|
| Paystack | 1.5% | 2.9% | 3.8% | T+2 | None |
| Flutterwave | 2.9% | 3.2% | 4.8% | Next day (local) | None |
| IntaSend | 3% | 3.5% | 4.5% | Not published | None |
| Pesapal (reported) | 3–3.5% | 3.5–4.5% | — | 1–3 days M-Pesa, 3–5 cards | None |
| No gateway — Pay to Till | ~0.5%, capped KES 200 | n/a | n/a | Real time | None |
Two things jump out of that table, and neither is the one gateway marketing wants you to notice.
First, Paystack's M-Pesa rate is less than half of Pesapal's. The "fees are roughly comparable" line that circulates in Kenyan SEO content is simply not true on the payment method most Kenyan shops actually use.
Second, the cheapest row is the one with no gateway in it. Which brings us to the option almost nobody writing about this will mention, because almost everybody writing about this sells gateway integrations.
The option nobody sells you: no gateway at all
You do not need a payment gateway to accept M-Pesa on a website. You need a Safaricom Till or Paybill. The customer places the order, you send your Till number, they pay it in the M-Pesa app exactly as they would in a shop, and you confirm the order once you see the payment.
Safaricom's own Buy Goods merchant charge is roughly 0.5% of the transaction, capped at KES 200 — with a lower rate on small transactions. There is no gateway taking a second cut, and the money lands in your Till in real time rather than one to three working days later.
The catch is real and it is not technical: somebody has to read the confirmation and mark the order paid. At five orders a day that is two minutes of work. At fifty it is a job, and it is a job that does not get done at 2am when someone in Nakuru is trying to buy from you.
So the honest sequence for most Kenyan merchants is: start with no gateway, and add one when the manual confirmation starts costing more than the percentage would. That crossover usually sits somewhere around 10–15 orders a day, though it arrives sooner if you sell overnight or your margins are fat enough that the percentage barely registers.
This is exactly how we ship shops by default — see WhatsApp checkout for the no-gateway path, and M-Pesa integration for the automated one, including the case for not buying it yet.
Each gateway in detail
Paystack — cheapest on M-Pesa, slowest to settle
Published: M-Pesa 1.5%, local cards 2.9%, international cards 3.8%. No setup or monthly fees. Settlement T+2 to a bank account or M-Pesa wallet — money taken on Monday reaches you Wednesday. Volume discounts are available by arrangement.
Where it wins: the M-Pesa rate, and it is not close. If your shop is predominantly M-Pesa — which describes most Kenyan ecommerce — this is the cheapest of the four by a wide margin. Paystack also has the best developer documentation available in Kenya; it is the most Stripe-like option, which matters if you are paying someone by the hour to integrate it.
Where it loses: T+2 is the slowest published settlement here. For a business buying stock weekly on thin working capital, two days of float is a genuine constraint, and worth more than the percentage saved. Weekends stretch it further — Friday's takings may not land until Tuesday.
Flutterwave — fastest local settlement
Published: mobile money 2.9%, local cards 3.2%, international cards 4.8%. Transfers out cost KES 100 each, to bank or mobile money. Next-day settlement for local payments. Pricing excludes VAT. By default the customer bears the transaction charge, adjustable in the dashboard.
Where it wins: next-day local settlement, and the broadest pan-African footprint of the four — which matters if you sell into Uganda, Tanzania or Nigeria as well as Kenya, since one integration covers several markets.
Where it loses: nearly double Paystack on M-Pesa, and the highest international card rate in the table at 4.8%. The KES 100 per-transfer fee also punishes frequent small payouts — take money out weekly rather than daily.
Pesapal — the Kenyan incumbent
Reported, not published: M-Pesa 3–3.5%, cards 3.5–4.5%. Settlement typically 1–3 business days for M-Pesa and 3–5 for cards, into KCB, Equity, Co-op, NCBA or Absa accounts. Higher-volume merchants can negotiate faster settlement.
Where it wins: it is a Kenyan company with Kenyan support, a long track record, and settlement directly into the local banks businesses here actually use. For a merchant who wants someone to answer the phone in Nairobi hours, that is worth real money — and this is a case where paying more is a defensible choice rather than a mistake.
Where it loses: the most expensive M-Pesa rate here, and no published rate card, which means you negotiate without knowing the benchmark. Ask for the rate in writing before you integrate.
IntaSend — the developer-friendly local option
Published: M-Pesa 3%, local cards 3.5%, international cards 4.5%, CashApp 4.5%. M-Pesa payouts cost a flat KES 100. No setup or monthly fees. Settlement timing is not published — ask before committing.
Where it wins: a Kenyan company with genuinely good developer experience, transparent published pricing, and a wider payment-method spread than the others, including crypto and CashApp if that is relevant to you.
Where it loses: double Paystack on M-Pesa, and the unpublished settlement time is a real gap. Any provider that publishes its rates but not its settlement timing is telling you which number it would rather you compared.
What the spread actually costs
Percentages are abstract. Here is the same shop under each option, at an average order value of KES 3,000 — a realistic figure for Kenyan fashion, beauty and homeware.
| Orders/day | Annual revenue | Pay to Till (~0.5%) | Paystack (1.5%) | Flutterwave (2.9%) | Pesapal (3.5%) |
|---|---|---|---|---|---|
| 5 | KES 5.5M | KES 27,000 | KES 82,000 | KES 159,000 | KES 192,000 |
| 15 | KES 16.4M | KES 82,000 | KES 246,000 | KES 476,000 | KES 575,000 |
| 50 | KES 54.8M | KES 274,000 | KES 821,000 | KES 1,588,000 | KES 1,916,000 |
Rounded, and ignoring the KES 200 Buy Goods cap, which makes the Pay-to-Till column cheaper still on larger orders. The point is the shape rather than the decimals.
At five orders a day, the gap between the cheapest gateway and the most expensive is around KES 110,000 a year — real money, but not business-defining. At fifty orders a day it is over a million shillings annually, which is a salary. The percentage you shrug at when you are small is the one that funds a hire when you are not.
Note the first column. At five orders a day, staying off gateways entirely saves you more than choosing the cheapest gateway does. That is the decision most Kenyan merchants should be making first, and it is the one least written about.
How to actually choose
In order, because the first question makes the rest moot:
- Do you need a gateway yet? Under roughly 10–15 orders a day, probably not. Take payment to your Till, confirm manually, and revisit when confirmation becomes a chore rather than a moment.
- Is cash flow or margin your binding constraint? If you restock weekly on tight working capital, next-day settlement beats a lower percentage — take Flutterwave. If you have breathing room, take the lower rate.
- What share of your payments are actually cards? Most Kenyan shops are overwhelmingly M-Pesa. Optimise for the M-Pesa rate and let the card rate be whatever it is; obsessing over card fees on a 95%-M-Pesa business is optimising the wrong number.
- Do you need someone to answer the phone? If a payment problem at 4pm on a Friday needs a human in Nairobi, Pesapal's premium buys something real. Support is a feature, and it is priced like one.
- Are you selling beyond Kenya? Flutterwave's multi-market coverage means one integration across several African markets, which can be worth more than the per-transaction difference.
And a reassuring point: since none of these charge setup or monthly fees, choosing wrong costs you developer time to switch, not a forfeited subscription. Do not agonise. Pick the one that fits today's constraint and move on.
Four traps worth knowing about
1 — Payout fees hide the real cost. The headline percentage is not the whole bill. Flutterwave charges KES 100 per transfer out and IntaSend a flat KES 100 for M-Pesa payouts. Withdraw daily and that is KES 36,500 a year in fees nobody quoted you. Withdraw weekly instead.
2 — Settlement timing is a cash-flow decision, not a detail. T+2 means Friday's sales may not be spendable until Tuesday. If you buy stock with the proceeds of last week's sales, model this before you choose. A merchant who picks the cheapest rate and then cannot restock has optimised the wrong variable.
3 — VAT is often excluded from quoted rates. Flutterwave states this explicitly. Assume every quoted percentage is pre-VAT unless the provider says otherwise, and ask directly rather than discovering it on your first statement.
4 — Passing fees to customers costs more than it saves. It is legal, and Flutterwave even defaults to it. But Kenyan shoppers are conditioned to the Till price being the price, and a surprise surcharge at checkout is one of the more reliable ways to lose a completed cart. Build it into the product price. At 1.5–3% that is an adjustment, not a repricing.
The short version
- Under ~10-15 orders a day, use no gateway. Pay to Till is ~0.5% and settles in real time.
- Past that, Paystack has the lowest published M-Pesa rate at 1.5% — but the slowest settlement at T+2.
- Flutterwave at 2.9% buys you next-day settlement and multi-country coverage.
- Pesapal costs the most and does not publish rates, but sells local support and local bank settlement.
- IntaSend at 3% is transparent on price and silent on settlement — ask before you integrate.
- None charge setup or monthly fees, so a wrong choice is cheap to reverse.
If you are still deciding whether you need any of this yet, the full M-Pesa checkout guide walks through the three routes end to end. If you are earlier still and pricing a shop, start with what an ecommerce website actually costs in Kenya, or work through the 30-item launch checklist before you go live.
Rates verified against each provider's published pricing in August 2026. We will re-check them at each annual refresh, but providers change pricing without announcing it — confirm the current rate before you sign anything.