Research That Lands With Founders: Sharing UX Findings in Fast-Moving MENA Startups

You've run the interviews, tagged the notes, found the patterns and built a 40-slide deck. You present it on Tuesday. The founder nods, says "very useful", and on Wednesday the team ships exactly what they were going to ship anyway. The research was good. It just didn't land.

NN/g's recent video, Why Your Research Findings Aren't Landing, and Dovetail's piece asking why good research isn't changing anything both make the same point: findings that don't change a decision are wasted effort. Most of that advice assumes a large company with research ops, scheduled readouts and stakeholders who have time to read. Many of our clients are founder-led startups in the region, where decisions happen fast, sometimes in a WhatsApp thread or a voice note between meetings. Here's what works in that setting.

Why decks of findings die in founder-led teams

In a startup with a strong founder, decisions are fast, informal and often made by one or two people. There's no product council that meets on Thursdays. A pricing change might be decided over lunch. A feature might be killed in a voice note from the airport.

A long research deck doesn't fit that rhythm. It arrives after the decision window has closed, or it asks the founder to set aside an hour they don't have. Even when they do read it, the important part is on slide 23, after the methodology and the personas.

This isn't a criticism of founders. It's a mismatch between how research is packaged and how decisions are actually made.

Start from the decision, not the data

The single biggest change I've made is to ask, before any research starts: what decision will this inform, who will make it, and by when? If nobody can answer, the research probably shouldn't happen yet.

Then I frame the output around that decision:

• The decision: "Should we launch cash on delivery in Egypt before card payments?"
• What we found: three or four findings that bear directly on it.
• What we recommend: a clear answer, with the confidence level.
• What we don't know yet: the open questions and how to answer them.

Everything else, like methodology, full participant breakdown and secondary findings, goes in an appendix for the people who want it. NN/g's article on turning complex data into compelling stories makes a similar case for leading with the narrative rather than the data dump.

Formats that travel

Research in a startup needs to survive being forwarded, skimmed on a phone and discussed without you in the room. Three formats do that well.

The one-page decision memo

One page, structured as above: decision, findings, recommendation, open questions. It works as a PDF, a doc or a pinned message. If the founder reads only one thing, it should be this.

The 90-second clip reel

Nothing persuades like hearing a user say it. A short reel of four to six clips, each tied to a finding, does more than a page of quotes. Keep it under two minutes so it gets watched to the end. Subtitle every clip, because it will be watched on a phone with the sound off.

The chat-friendly summary

Decisions in many of our client teams happen in WhatsApp or Slack. So I write a summary that fits in one message: the decision, the top three findings in one line each, the recommendation, and links to the memo and the reel. It sounds trivial. It's often the thing that actually changes the decision.

[Waleed: add an anonymized example of a short summary message that changed a decision on a Breem project.]

Presenting bilingual evidence without losing nuance

Our research is often conducted in Arabic, sometimes in a specific dialect, while the stakeholders may include English-speaking investors, product managers or developers. Translation can quietly strip the meaning from a finding.

A few rules I follow:

• Keep the original quote. Show the Arabic, then the translation. Arabic-speaking stakeholders hear the tone, and others still get the meaning.
• Translate meaning, not words. A literal translation of an Egyptian expression can sound odd or even rude in English. Add a short note when the tone matters.
• Flag dialect and context. "Said by a Gulf user about paying in installments" carries more information than the quote alone.
• Subtitle clips in both languages when the audience is mixed, or produce two versions of the reel.
• Don't over-clean quotes. Hesitations and code-switching between Arabic and English are often part of the finding. A user switching to English for a payment term tells you something.

This also connects to how much you can trust synthetic research in the region. I've written about that in Synthetic Users Don't Speak Egyptian.

Following up: tying findings to the backlog

Research lands when it shows up in the work. After every readout, I do three things:

1. Turn each finding into a backlog item (a ticket, a task or a card) with a link to the evidence.
2. Give each item an owner on the client side, not just "the team".
3. Check back two weeks later. A short message, such as "Here's where the three findings from the payments research stand", keeps them visible without nagging.

Over time, this builds a track record. When the founder sees that a finding from last month prevented a costly mistake, the next round of research gets more attention.

Closing thoughts

Good research that doesn't change a decision is an expensive way to be right in private. In fast-moving startups, the packaging matters almost as much as the findings: start from the decision, keep it short, make it forwardable, respect the language the evidence came in, and follow it into the backlog. None of this needs a research ops team. It needs a researcher who designs the output as carefully as the study.

Sources: Why Your Research Findings Aren't Landing (NN/g video), Your research is good, so why isn't it changing anything? (Dovetail), Complex Data, Compelling Stories (NN/g).

FAQs