What 18 Months of Building a Mental Health Company Taught Me About My Own Mind

What 18 Months of Building a Mental Health Company Taught Me About My Own Mind

I co-founded Mena.ai, a digital mental-health company that builds software for psychologists and patients, eighteen months ago. For most of that time, I did not use what we were building. I want to write about why, because I think the answer matters beyond my own story - and because Mental Health Awareness Month is a strange time to keep being quiet about it.

The most recent founder surveys put the share of entrepreneurs reporting mental-health struggles at around 72%. Only 7% receive any kind of support. Research from UCSF (Freeman et al.) found founders to be roughly 50% more likely than the general population to experience a mental-health condition. I was, for at least the first year of Mena, comfortably inside the first statistic and refusing to join the second. This is a piece about how that happened and how it ended.

The First Statistic Is Quieter Than the Second

The thing that gets undercounted about founder mental health is how socially adaptive the symptoms are. Anxiety, in the right industry, looks like drive. Hypervigilance looks like attention to detail. The inability to switch off looks like commitment. Insomnia, framed correctly, becomes "I get my best work done at 4am."

For the first year of building Mena, my weeks looked like this: wake at 6, work until midnight, skip lunch, eat dinner at the desk, talk to my partner in five-minute fragments, and call it intentionality. My resting heart rate had drifted up by about ten beats per minute. My sleep had broken into 90-minute chunks. The metric I cared about - investor updates, product velocity, recruitment - kept improving. The metric I wasn't measuring - whether I was still a person I recognized - kept getting worse, silently.

Here's the part I want to dwell on, because it's the part that keeps me up at night. The work I was doing during the day was, literally, building software to identify exactly what I was experiencing. I was integrating clinical screeners (PHQ-9 for depression, GAD-7 for anxiety) into our platform. I was sitting in user research calls with psychologists describing burnout, panic disorders, and the cognitive distortions that fuel them. I was reading clinical literature on emotion regulation. I had every diagnostic instrument the industry uses sitting on my screen - and I never once ran any of them on myself.

That is not coincidence. That is denial with a startup wrapper.

The Tuesday in Autumn

I am going to keep this brief because I'm wary of the redemption-arc problem in founder mental-health writing - the tendency to wrap a breakdown in a story arc that makes it sound, retrospectively, noble. It isn't, and it wasn't.

It was a normal Tuesday in autumn. I sat down for dinner. I started crying. I cried for about an hour. My partner asked me what was wrong and I could not answer - not because the truth was complicated, but because I had spent eighteen months helping other people put words to feelings and had built, for myself, a vocabulary of zero.

There was no insight. No epiphany. No product idea fell out of my tears. What I got was a fact I had been actively avoiding: I was inside the population I was building for, I had been inside it the whole time, and pretending otherwise had cost me something I could no longer un-cost.

What I've Learned Since (And What I Wish Someone Had Told Me)

I'm going to lay out the lessons that have stuck. They are simpler than the experience. That's part of the point - the work isn't intellectually hard, it's emotionally inconvenient.

Productivity is not the same as functioning. Founders confuse these constantly, and the people around us - investors, advisors, even families - are often complicit, because productive misery is more legible than functional rest. During the worst stretch of last year, I shipped more code than at any other point. I closed a round. I gave the keynote I'd been preparing for nine months. None of that was evidence I was okay. It was evidence I had become very good at performing okayness. The two are not the same, and you can spend a decade telling them apart.

Therapy is a competence, not a confession. The framing I had inherited - that therapy is for people who have "really broken" something - is wrong, and it kept me out of a room I should have been in years earlier. Watching real psychologists work, day after day, in user research and product calls, slowly rewrote my sense of what therapy is. The good ones do not "fix" anyone. They teach skills: emotion regulation, cognitive flexibility, distress tolerance, interpersonal repair. These are buildable in the same way you'd build any technical skill. If you wouldn't be embarrassed to learn Postgres, you should not be embarrassed to learn how to sit with discomfort.

The 167-hour problem applies to you, too. If you are in weekly therapy, you spend 1 hour in session and 167 hours in your actual life. Roughly 0.6% of your week is therapy. The other 99.4% is where the work either takes root or quietly fails. Patients dropping out of therapy - meta-analysis puts the rate at 34.8% across 146 studies, with younger adults at the high end - usually aren't dropping out because therapy is bad. They're dropping out because nothing in their week, except that hour, supports the change they're trying to make. Founders fall into the same trap. The session is the easy part. The infrastructure around it - sleep, sunlight, friends who don't only want to talk about your company, a partner you actually see - is what determines whether anything sticks.

The "investment thesis" framing is useful here. I came around to therapy faster once I thought of it the way I think about technical infrastructure. Nobody waits until production is on fire to invest in observability. You build it before you need it, because by the time you need it you can't think clearly enough to set it up. Mental health is the same. The cost of doing this work when you are calm is small. The cost of trying to do it when you are mid-crisis is enormous.

The boundary you most resist is the one you most need. Mine was no Slack after 8pm and no work on Saturdays. I told myself, with absolute conviction, that the company could not survive me doing this. The company has survived it. My relationships almost did not survive me not doing it. The general form: whatever boundary you flinch at hardest is the most diagnostic information you have. Trust the flinch.

The data is unambiguous and we're all pretending it isn't. Founders are about 50% more likely to experience a mental-health condition than the general population. Up to 72% report current struggles. Only 7% receive any form of support. That gap - between how prevalent this is and how rarely it is named - is the cultural problem we're embedded in. We have made admission expensive. We have made vulnerability look weak. Then we wonder why everyone is quietly miserable. The fix is not personal heroism. The fix is making it normal to use the support we already build for other people.

Where Mena Goes From Here

I am not going to pretend that running a mental-health company makes me an authority on my own mental health. If anything, the irony has gone the other way: building this has made me painfully literate about the gap between knowing what's healthy and doing it.

But I will say this. Every product decision we have made in the last six months - the way we think about between-session continuity, the way we design our mood-tracking flow, the way we are building our work for clinicians treating high-risk young adults - has been shaped by a year I spent being a textbook case of the thing we are trying to fix. That is not a marketing line. That is just how it shook out.

If you are a founder reading this and seeing yourself in any of it: the work is not over because you noticed. Noticing is step one. The thing you are calling drive deserves a closer look. Anxiety and drive can be indistinguishable from the outside. They are not the same thing inside, and you can spend a long time not knowing which one is running you.

Therapy is one option. So is a real conversation with a person who knows you well enough not to be impressed by your last update. So is making it boringly normal in your founder community to ask, out loud, how someone is actually doing - and to answer honestly when someone asks you.

A Short Note on Resources

If you are in crisis or thinking about hurting yourself, please reach out before anything else. In Portugal, SNS 24 (808 24 24 24) and the mental-health support line (808 200 204) are both available 24/7. In the US, you can call or text 988. Almost every country has an equivalent service - please use it.

If you are looking for less-acute support: Mental Health Awareness Month is the month many therapists in your country will be running open-enrollment intake. It is a good month to do the thing you keep telling yourself you'll do "after this quarter."

If you are a clinician reading this and thinking about the founders in your practice: thank you. We need you, and we are slowly getting better at admitting it.

Frequently Asked Questions

Is founder burnout really that common, or is it overstated?

The numbers are unambiguous. UCSF's Freeman et al. study found founders to be roughly 50% more likely than the general population to experience a mental-health condition. Recent surveys put the share of entrepreneurs reporting current mental-health struggles at around 72%, with only 7% receiving any kind of support. The gap between prevalence and help-seeking is the central pathology - we've made admission socially expensive, and the costs compound silently.

How do I tell the difference between healthy drive and anxiety dressed up as drive?

Healthy drive is responsive: it ramps up when there's a real problem and recedes when the problem is solved. Anxiety presenting as drive doesn't recede - it scans for a new threat the moment the old one is resolved. Other tells: chronic difficulty switching off, sleep that breaks into 90-minute chunks, a resting heart rate that has crept up without obvious cause, and an inability to enjoy the wins. If success doesn't lower your baseline, you're probably running on the wrong fuel.

What is the "167-hour problem" and why does it matter for therapy?

A weekly therapy session is one hour out of 168. That hour is 0.6% of your week. The other 99.4% is where any actual change has to land. Meta-analyses of psychotherapy drop-out (Swift & Greenberg, 2012) found a 34.8% mean drop-out across 146 studies - usually not because therapy is bad, but because nothing in the week between sessions supports the change. For founders, the takeaway is the same: the session is the easy part. The infrastructure around it - sleep, sunlight, real friendships, boundaries - is what makes it stick.

Does building a mental-health company give you any meaningful protection from your own mental-health problems?

No. If anything, the proximity creates a false sense of immunity. Knowing the screeners, the clinical literature, and the language of distress is not the same as applying any of it to yourself. The pattern I see most often is sophisticated avoidance - turning your emotional life into a product roadmap rather than processing it. Knowledge and self-awareness are different skills, and both have to be practised deliberately.

---

Mena.ai is a digital mental-health platform built with clinicians in Portugal and the UK. We support therapists and patients in the long stretch between sessions. Learn more →

Francisco Ribeiro e Silva is the co-founder of Mena.ai. He writes about the intersection of technology, mental health, and the human side of building a startup. For a companion essay on the same themes, see "The Founder Paradox".

---

References:

• Freeman, M. A., Johnson, S. L., Staudenmaier, P. J., & Zisser, M. R. (UCSF). Are entrepreneurs touched with fire? University of California, San Francisco.
• StartupNation (2024). Startup Pressure Is Real: Why 72% of Founders Struggle with Mental Health.
• Swift, J. K., & Greenberg, R. P. (2012). Premature discontinuation in adult psychotherapy: A meta-analysis. Journal of Consulting and Clinical Psychology. (Meta-analysis of 146 studies; mean dropout 34.8%.)
• De Salve, F., et al. (2025). Dropout in Psychotherapy for Personality Disorders: A Systematic Review of Predictors. Clinical Psychology & Psychotherapy.
• Kroenke, K., Spitzer, R. L., & Williams, J. B. (2001). The PHQ-9: Validity of a brief depression severity measure. Journal of General Internal Medicine.
• Spitzer, R. L., et al. (2006). A brief measure for assessing generalized anxiety disorder: the GAD-7. Archives of Internal Medicine.