Case study · Rwanda · UX, redesign

Kumwe Hub.A website that talked a lot and said very little.

Kumwe Hub sits in Rwanda and does something hard to explain in one line: it connects global impact investors with African businesses that need capital, and it runs its own lab, fund and incubator to get those businesses ready. The work is real. The website wasn’t carrying it.

They came to us because they felt the site wasn’t working. They couldn’t say exactly how. That’s usually the honest version of the problem, and it’s a good place to start.

We started with the numbers

And the numbers were lying.

The first thing we did was open their analytics. None of it held up. By then, mobile had been outperforming desktop worldwide for years, and a Rwanda-facing organisation drawing almost no mobile traffic isn’t a finding, it’s a broken measurement. Direct traffic that high means referrals aren’t being tracked. Which means nobody could tell which channels were bringing people in, or whether any of it was working.

So we told them that before we told them anything else. You can’t optimise what you can’t see, and the first fix wasn’t a design fix.

Direct traffic: 90%
Desktop: 90%
From the US: 57%

Then we read the site

The way a stranger would.

We went through every page and wrote down what a first-time visitor would run into.

The header took up most of the screen. The logo was the biggest thing on it.
The navigation broke onto two lines.
The language selector didn’t work, even though new content was being published in the local language.
The core message diagram on the homepage was covered by the cookie banner.
There was no footer at all.
The hero told a long story where it should have made a promise.
Open Calls sent people to a Google Doc.
Requirements were presented before benefits.
The Rwanda Ecosystem Map lived on another site, wasn’t explained, and fell apart on a phone.
The Partners page — the page that exists purely to build trust — wasn’t linked from anywhere.

None of this was carelessness. It’s what happens when a site grows one addition at a time and nobody gets to step back.

The competition had the same disease.

We pulled apart VisionFund, InkoMoko and Kua Ventures to see whether anyone in the space had solved this.

Nobody had. VisionFund gave visitors kilometres of text and almost nowhere to make contact, with the one real path buried in an FAQ that only appears in the footer. Kua Ventures pushed its most persuasive page, the portfolio, off the homepage entirely. Every one of them treated storytelling and conversion as separate things, and then never got around to the second one.

That was good news. The bar in this sector is low, and clearing it doesn’t require spectacle. It requires respect for the visitor’s time.

Two people

We stopped designing for “users”.

We built two personas and stayed with them for the rest of the project. We mapped how each of them would actually move through the site, from first search to the point where they’d reach out.

Sarah

Investment data analyst

She wants numbers she can trust, quickly, and she can smell filler from the first scroll. She browses on her phone between other obligations. Give her a wall of text and she’s gone.

John

Philanthropy programmes, bank

He needs to know the impact is real and the reporting will hold up when he takes it to a committee.

These people know what they’re looking for. Immediate hook, immediate contact. Be as transparent as possible.
Written on the board halfway through

What we rebuilt

Who each page is for.

We restructured the site around the visitor. The Lab, The Fund and The Incubator don’t speak to the same person, so they stopped looking identical and started reading like three distinct offers. Proof moved up. Obstacles moved down. Contact became available at every point where someone might be ready, instead of only at the end.

Then we wireframed it: two directions for the homepage, plus About us, the Lab landing and the rest of the structure, with the navigation redrawn so that nothing important is more than two clicks from anywhere.

How the work ran

Four phases, scoped in days.

01
Assessment
Review the project, dig through analytics, research the competitors, name the UX problems in terms of the business goals. Unglamorous on purpose.
02
User flow
Sarah and John, from first search to the point they’d reach out.
03
Visual design
Two homepage directions, then the rest of the structure.
04
Optional branding
Held until the questions we couldn’t answer ourselves had been asked.

We also left the client with the questions we couldn’t answer ourselves. Why the surge in US visitors. Whether a photo repository existed. What the ecosystem map was actually for. How prominent African heritage should be in the visual language. Some of those shaped the design more than anything we found on our own.