Skip to content
Resources / The Question / Issue 4

Arash Farhani: the first data he trusts is a testimonial

Arash Farhani. Serial entrepreneur and AI-native builder, founder of Harmonea. Issue 4 of The Question, published by A Founder's Friend on 9 August 2026. By Nicolas Dolenc, author of A Founder's Friend. Skene adapted this page from his published issue.

A missed day in Harmonea's 21-day course does not turn a chart red. It raises a yellow flag in the CRM, which the series editor reads as a question rather than a verdict. Someone follows up to find out what got in the way. For Arash Farhani, the flag is the cue to get in touch: it tells the team who to contact, not why that person stopped.

The short answer: Before a few hundred users, intuition and testimonials lead, and a number earns authority only once direct contact with clients confirms it.

What Arash Farhani trusts, and for what

EvidenceTrusted forNot trusted for
Research, lived experience, intuitionFirst decisions on a service with no comparable productBeing the only input past two or three hundred active users
Testimonials and direct feedbackThe first data that earns trustScale
Day-by-day course steps in the CRMSpotting who fell behind, and whenWhy they did
Web analytics: clicks and landing pagesSupporting detailLeading a decision

Knowing when the numbers have not earned it

Harmonea is a new kind of service, so there was no comparable product to study. Arash says that before two or three hundred active users, a new service has to run on research and intuition. The series editor named the discipline: recognising that a number has not earned trust yet.

Testimonials are data

When real users arrived, the team did not respond by stacking up features. It listened harder. Direct testimonials and feedback became the signal Arash trusts, and he is firm that this is still a data-driven approach, with the trustworthy data defined narrowly. Interview sessions and weekly catch-ups with people taking the course sit in front of the analytics, which exist but come second.

A missed step is a question

The course is tracked step by step in a CRM. A completed step logs a check mark and fires a trigger. A missed step raises a yellow flag that routes a follow-up through Telegram.

The flag does not conclude anything about the person. It tells the team who to ask.

When shipping outruns the checks

Arash spent two months building a marketing agent, then moved the team's whole development process into Telegram. A technical agent acts as the CTO: describe a feature in chat, and it builds it, opens the pull request and reports back, often within the hour, that it is live on staging. With building that cheap, Harmonea prefers to build its own tools rather than buy them.

The series editor named the open risk that follows from that speed. When a feature reaches staging in under an hour, shipping can outpace the checks that keep its tracking honest.

What to do with this

A yellow flag is a good model for any metric that moves: treat it as a prompt to ask someone, not as the answer. And if an agent ships your features, give it a check for the tracking each feature needs, so speed does not delete the evidence along the way.

For seven checks drawn from across the series, see the questions to ask before trusting a number.

Frameworks this interview tests

  • Jobs to be done (JTBD): Clayton Christensen's jobs to be done says customers hire a product to make progress. The job decides which events matter, so interviews come before the tracking plan.
  • Customer health score: A customer health score blends usage, support and sentiment to predict renewal risk. Usage carries the most weight, and nothing in the score checks the usage events.

Source and disclosure

Read Issue 4 on A Founder's Friend. The full interview is for subscribers. This page paraphrases it and quotes nothing that sits behind the paywall.

Skene sponsored this research and commissioned it on a single question. A Founder's Friend ran the interviews, wrote them up and published them, and the series states that Skene paid for the research and did not direct the questions, the analysis or the conclusions. Each of the thirteen interviews was reviewed and signed off by the person in it, named or anonymous by their choice. Appearing in the series is not an endorsement of Skene, and nobody here is presented as a Skene customer.