Flavia Mbuki
Flavia MbukiVenture Builder · Africa
← Back to portfolio
Case study · Sourcing → Partner → Decision gate
Creative Creator monetization OTT / streaming 🇰🇪 Kenya Live · 2026

The partner scored highest on the one thing that never worked

A creator-monetization venture needed local payments to work on day one. We ran a build-buy-partner analysis, scored four OTT vendors, and partnered for the proof of concept. The vendor won mostly because M-Pesa came "already integrated." Two months and four thousand dollars later, it still did not take a single shilling.

This is the honest version of a partnership: how we chose, what we were sold, what actually showed up, and the decision gate we built with a few founders who have been burned before. Personal details and the vendor's brand are redacted. Everything else is straight from the record.

The question

Give creators a way to sell content in Kenyan shillings through M-Pesa.

My responsibility

Compared technology partners, ran the engagement and challenged delivery against the agreed requirements.

The outcome

The payment integration remained blocked at the July 2026 review. The case records the recovery options, not a completed launch.

My role
Founder / venture lead
Sector
Creative · creator economy
Path chosen
Partner for POC, build later
Spend to date
~$4,340 over ~8 weeks
Status
At the decision gate
4OTT vendors scored on weighted criteria
#1ranked vendor, chosen for the POC
8 wksof tickets with M-Pesa still not working
0shillings collected through the platform
15%discount later won back, tied to milestones
01 · Context

Creators can make the work. Getting paid for it is the hard part.

In the creative sector, the talent problem is mostly solved. People shoot, edit and publish from a phone. What stays broken is the money. Ad-share on the big platforms pays cents, takes a cut, and owns the audience. The creator never holds the customer relationship, and almost never gets paid in the currency their fans actually spend.

For African audiences that gap is sharper. The wallet that matters is mobile money, M-Pesa first, and none of the off-the-shelf creator tools speak it well. So a filmmaker with a following and a library of work still has no clean way to charge 200 shillings for a film and have it land, tonight, from a phone in Nairobi.

The problem
Creator monetization plans today are not built to last, and for local audiences paying in local money, a working option barely exists.
02 · The bet & my role

A creator-owned storefront, paid in the money fans already use.

The venture thesis was simple to say and hard to ship: give a creator their own branded streaming storefront, on web and in the app stores, where fans subscribe or pay per view, in Kenyan shillings, over M-Pesa. Own the audience, own the pricing, keep the margin.

Before writing a business plan, I wanted a proof of concept in front of real fans in weeks, not quarters. As founder I owned the whole call: the sourcing, the money, the integration work, and the decision about whether the first partner was worth keeping. What follows is that decision, documented.

The fastest way to learn whether fans will pay is to let them try, not to spend six months building the till first.

The reasoning behind partnering for the POC
03 · Build, buy or partner

Rent the platform to learn fast, build it once the demand is real.

Standing up an OTT platform from scratch, apps, transcoding, DRM, billing, a payment stack, is a year of work and a serious burn before a single fan is charged. Three paths went on the table, scored against one question: what gets a paying fan on the platform quickest, without betting the runway?

Build

Our own streaming stack

  • Full control of the M-Pesa checkout and the creator experience
  • Months of engineering before the first fan can pay
  • Wrong order: building the till before proving anyone lines up at it
Later, once demand is proven
Buy outright

License and self-host a platform

  • One-time control, no monthly platform lock-in
  • Heavy setup, hosting and maintenance land on us from day one
  • Too much commitment for something still being tested
Too heavy for a POC
Partner ✓

White-label OTT SaaS for the POC

  • Branded apps and a storefront live in days, not months
  • Monthly cost, walk away if fans do not bite
  • Lets us test desirability before committing to a build
Chosen: fastest, reversible, runway-safe

Partner for the proof of concept, build later once the numbers earn it. The strategy was right, and I would make the same call again. The risk was never the path. It was picking the right partner to rent it from.

04 · Vendor scorecard

Four platforms, one deal-breaker: does it take M-Pesa?

I ran a weighted vendor scorecard, four OTT platforms across functional fit, ease of use, technical alignment, vendor viability and total cost. Two capabilities were non-negotiable: M-Pesa and pricing in Kenyan shillings. Without those, nothing else on the sheet mattered.

[E1] Sourcing[ Vendor scorecard ]

Vendor
Scorecard

We evaluated white-label OTT platforms on the capabilities a creator storefront actually needs, then weighted them against price. Vendors are anonymized. One capability decided it: only one platform claimed M-Pesa was already integrated.

04platforms
evaluated across 5 weighted criteria, 2 people scoring each
Vendor Aselected for POC28
Vendor Brunner-up22
Vendor Cnot scored fully—
Vendor Dnot scored fully—
CapabilityABCDPriority
M-Pesa payment Can't live without
Local currency (KES) pricing Can't live without
Subscriptions & pay-per-view Core
Apps on TV, mobile & web Core
Cancel / change plan anytime Core
Modern sign-in (Google / social) Core
Customization & templates Core
Local (Kenya) market experience Nice to have
On-premise deployment option Nice to have
Total cost of ownership 30% of weight, never scored
Supported Partial Not supported Coral = the selected vendor
Vendor A won on paper, 28 to 22. It was the only platform claiming M-Pesa was already integrated, it priced in shillings, it carried apps everywhere, and it had Kenyan customers to point to. Note the last row: total cost of ownership carried 30% of the weight and was left blank for every vendor. The sheet decided on features it could see, and the money it could not.

The scorecard did its job. Given what each vendor claimed, Vendor A was the right pick, and M-Pesa was the right thing to weight above everything. The flaw was quieter: every green mark in that column was a sales claim, not a tested feature.

05 · What we bought

One click to launch, the pitch said.

The vendor's own marketing set the expectation: launch an OTT platform instantly, a thousand-plus enterprise features, native apps on every screen, and the payment rails our market runs on, ready to go. We took it at its word and paid to start.

↳ What the marketing sold
  • "Launch your OTT platform in 1-click"
  • 1000+ enterprise-grade features
  • M-Pesa already integrated for subscriptions and pay-per-view
  • Native apps across TV, mobile and web
  • Support SLA measured in minutes
↳ What week one delivered
  • A subscription that needed back-end activation to even appear
  • Payments and integrations greyed out in the console
  • M-Pesa "set up," but the button did nothing on click
  • App submissions stalled on repeated access mix-ups
  • Weeks of tickets, and credentials re-sent five times over

None of this was catastrophic on its own. Software is fiddly. What made it a pattern was the one column that mattered most on the scorecard, M-Pesa, being the one that would not move.

06 · What actually happened

Eight weeks, one payment button, and a timeline that kept slipping.

Here is the engagement from first inquiry to the latest invoice, drawn straight from the email and ticket record. Dates are exact. The vendor's brand and every person's contact detail are removed.

Sourcing & sign-up
MAY 4, 2026
Requested a 3-month trial quote for Android, iPhone and Android TV apps.
MAY 7, 2026
Sales call. Stated the need plainly: monetize video on demand, multi-platform, local payments.
MAY 13, 2026
Subscribed to the player SDK and asked for setup help to go live "in a few days." Chased support: "reaching out for weeks with no success."
MAY 21-22, 2026
Told the SDK was not enough. Upsold to the "Professional" tier on a call. Custom invoice raised, $2,198.
Setup & integration
MAY 25, 2026
Paid the $2,198 upgrade. The prepaid card drained to $3.84. Dashboard did not update; the plan needed manual back-end activation. delay
MAY 27 - JUN 1, 2026
iOS app request stalled on developer-account access and an admin-invite mix-up over an email's capital letters. "Is this a show stopper?" stalled
JUN 4-11, 2026
M-Pesa ticket opens. Credentials shared, then re-requested, then shared again. PayPal integration greyed out. Sandbox keys go back and forth; I point them to Safaricom's own API docs. churn
JUN 16-19, 2026
Escalation call. A named timeline at last: M-Pesa "tentative ETA July 7."
The deadline that slipped
JUN 30 - JUL 1, 2026
Cannot create subscriptions or pay-per-view. Currency is locked to USD, Kenyan shillings greyed out. Asked again: are we on track for July 7? blocked
JUL 7, 2026
The ETA day. "M-Pesa integration done, add it to your CMS." I try it: the gateway does not respond on click, no redirect, no prompt. Support asks me to move the conversation into individual tickets. missed
JUL 8-9, 2026
"M-Pesa set up is done" again. I try to pay: the page stalls, no STK push, no PIN prompt, nothing reaches the phone. Meanwhile Google Play admin access has expired again. "We are way over the initial timeline." still broken
JUL 11, 2026
Monthly invoice arrives, $2,122.45, "15% discount included," the discount I had negotiated and tied to delivery. Billed on time; the core feature still did not work. billed anyway
The scoreboard, 8 weeks in
Spend to date~$4,340
M-Pesa payments taken0
Days past the M-Pesa ETA4+ and counting
Times credentials re-sent5+
What still did not work
M-Pesa STK pushNot firing
Pricing in shillingsLocked to USD
iOS app liveNot submitted
Named, kept deadlineNone yet
07 · The receipts

Straight from the ticket thread, redacted where it counts.

A few exchanges tell the whole story. Names, emails and every credential are blacked out. The words are unedited.

M-Pesa payment gateway Ticket #480██
Vendor support
"We have done the M-Pesa integration; you can now add it to your CMS."
Me, same day
"The M-Pesa gateway is not responding on click."
Me, again
"Please see M-Pesa details reshared below: ████████████████ ████████"
M-Pesa payment gateway Ticket #480██
Vendor support
"M-Pesa set up is done."
Me
"I'm trying to pay via M-Pesa and the page doesn't move beyond here. What we expected is an STK push and users input their M-Pesa PIN. Please provide a timeline as we are way over the initial timeline that was set for this integration."
Payment integration error Ticket #480██
Me
"I cannot add local currency (Kenya Shilling) in any of the PPVs or subscriptions. The only currency is USD and the option to specify other currencies is greyed out."
Vendor, escalation
"Could you follow up directly within the respective support tickets instead of this email thread?"
Invoice, June subscription Bill #178██████
Vendor billing
"We have raised an invoice of $2,122.45 for June Subscription for iOS, Android Mobile & Android TV, 15% discount included."
The reality
Billed in full and on time, for apps not yet live and a payment rail that still would not take a shilling.

Public reviews told the same story before we signed. A reference check against the vendor's Trustpilot page would have surfaced the exact pattern, slow delivery, ticket ping-pong, features that work in the demo and not in production. That check cost nothing and I skipped it. It is the first item in the gate below.

08 · The decision gate

Three doors, and the playbook I built to choose.

Sunk cost whispers "just wait one more week." To answer it with something better than hope, I sat with a handful of founders who have run SaaS partners into the ground and back, and turned their advice into a gate. First the three doors, then the rules for walking through them.

Door 1

Wait and keep paying

Trust the next "it's done," keep the monthly running, hope M-Pesa fires this week. Costs runway, keeps no leverage, and rewards the behavior that got us here.

Rejected
Door 2 · recommended

Renegotiate on milestones, prep the exit

Restructure to milestone-gated payments with UAT sign-off, hold the 15% discount against delivery, set one hard M-Pesa deadline, and start building the M-Pesa piece in parallel so we are one decision from independent.

Recommended
Door 3

Walk now, switch vendor

Cut losses, move to the runner-up or self-host. Clean, but throws away eight weeks of setup and restarts the same integration risk with a new party. Held in reserve if the deadline slips again.

In reserve

The playbook underneath door two, distilled from people who have paid this tuition already:

01Milestones with sign-off. Break the work into phases with a UAT document signed off at each one. "Done" means a feature I tested, not a note on a ticket.
02Prioritize the can't-live-without. M-Pesa is the deal-breaker. Everything else is negotiable and sequenced behind it.
03Tie money to delivery. The 15% discount and each payment attach to a milestone, not a calendar date. Pay for what shipped.
04Get the full cost picture. Every add-on, infra and platform fee on the table now, so it can go into the unit economics instead of surprising them later.
05Check references first. Public reviews and past clients on cost, delivery and experience. Cheap insurance I will never skip again.
06Protect the runway. The goal is to reach fans before the money runs out. Discounts and milestones exist to keep the venture alive that long.
07Hold the poker face. No lack of faith signaled upfront. It keeps the partner from hedging and keeps our leverage intact.
08Write it down, record the calls. A script before each call, a record after. This case study is that habit paying off.
09 · What I'd carry forward

Lessons.

The one that cost the most

Test the deal-breaker before you sign, not after you pay

M-Pesa being "already integrated" was the single reason this vendor topped the scorecard, and it was the one feature that never worked. Every green mark in that column was a claim from a sales deck, scored as if it were a fact. A free trial and thirty minutes in a sandbox, pushing one real test payment to one real phone, would have turned the deciding capability from a promise into proof, before any money moved.

◐ What we scored
  • A feature marked "supported" on a sales call
  • A checkbox on a comparison sheet
  • A demo environment, not our market's rails
● What we should have scored
  • One real M-Pesa STK push, to one real phone
  • A price set in shillings, checked in the console
  • The capability working in a trial before signing

Weight the deal-breaker heaviest, that part was right. Then make the vendor prove it before you weight it at all.

01
Partnering for the POC was still the right call
The build-buy-partner logic held. Renting the platform to test demand beat spending a year building the till. The mistake was the vendor, not the strategy, and reversible beats perfect.
02
Pay for milestones, never for calendars
Paying the upgrade upfront handed over the only leverage that makes a SaaS partner move. Money tied to a signed-off feature would have changed every conversation after.
03
Score the cost you cannot see
Total cost of ownership carried 30% of the weight and sat blank for every vendor. Infra and add-on fees only appeared as invoices. Price the whole relationship, not the sticker.
04
A reference check is the cheapest diligence there is
The public reviews described this exact experience before I signed. Fifteen minutes reading them would have priced the risk honestly. I never skip it again.
05
A thousand features lose to one working payment
Global feature breadth is worth little if the local rail is dead. In this market, one M-Pesa checkout that fires outranks a page of enterprise capabilities that do not.
06
Keep the exit warm
Because M-Pesa is the deal-breaker, building just that one piece keeps the venture a single decision away from independence. The best leverage in any partnership is not needing it.

Choosing or salvaging a SaaS partner? I've kept the receipts.

Weighted vendor scorecards, build-buy-partner calls, and the unglamorous work of holding a partner to delivery without burning the runway. I can help you evaluate a partner, define delivery milestones or plan a recovery.