1) The insight was structural, not opportunistic

Real estate is the world’s biggest contributor to CO₂ emissions. That’s not a new fact. What was different about Vincent Bryant and his co-founder Emmanuel was how they saw it.

They weren’t building another consultancy or planning to push another energy audit report. They believed the data already existed (scattered across public databases, utility bills, regulatory filings) and that the right technology could turn it into decisions.

The idea didn’t come from a lab. It came from years inside companies like ENGIE’s subsidiaries and Carbone 4, watching the same problem repeat itself with no good answer in sight.

“We were both engineers. We could see how much data was available. And we couldn’t understand why nobody was using it.”

That shared frustration became the foundation.

2) They spent a year testing before jumping

Vincent and Emmanuel spent a full year exploring the idea - while still employed. They used Vincent’s auto-entrepreneurship structure to run small consulting jobs on the side and took vacation days to visit a prospect in Toulouse: a LED lighting reseller who needed to know which industrial buildings were the best targets for his product.

Deeepki crunched the public data, ran the algorithms, and delivered a ranked list. The client was happy.

On the train back, something clicked.

“I told Emmanuel: let’s go. We probably don’t have enough proof yet. But let’s do it.”

Emmanuel’s answer: “Let’s have dinner with our wives first.”

3) The dinner that made it real

Before quitting their jobs, both founders sat down with their partners for dinner. Only after walking out of that dinner did they truly commit.

It’s not a trivial ritual. Starting a company is a family decision. It changes income, weekends, mental load, and the whole fabric of daily life. Neither of them had done this before.

“We needed the sign-off from our Chief Family Officer.”

4) First priority: clients, not product

From day one, their focus wasn’t on software. It was on customers. They had a clear vision of what they wanted to build (a data platform for real estate) but wanted to build it in response to real problems, not assumptions.

First clients came through warm introductions from former employers: Carbone 4, ENGIE subsidiaries. The first fully external client was Picard, found at a real estate trade show: 1,000 stores, €20M in annual energy bills, and a need to identify which buildings to instrument first.

“We helped them identify the barycenters, the most statistically representative stores. That was our first real mission.”

5) No regulatory tailwind. No obvious ICP. Just hustle.

The EU taxonomy didn’t exist. The Décret Tertiaire hadn’t been passed. Real estate owners would say: “Energy? The tenants pay that. Not our problem.”

The companies that bought early were the ones who had already decided to care, for reputation, for tenant retention, for cost savings. Deepki’s job was to find the 5-10% of the market that had already felt the urgency.

“We weren’t selling urgency. We were finding the people who already felt it.”

Thirty clients, averaging €20-30k per year, in 18 months. Entirely self-funded.

6) Fifty apps, then a real product

The early version of Deepki wasn’t a SaaS platform. It was a constellation of mini-applications, each built for a specific client, running on R Project analytics and custom ETL pipelines. It worked, until it didn’t.

“We hit the ceiling on maintainability. And we could also see what all these use cases had in common.”

In 2016, 18 months after launch, they raised €2.5M from Demeter and High-Inov to hire a CTO and rebuild on a unified, scalable infrastructure.

7) The co-founder relationship was managed like a product

Vincent and Emmanuel brought in coaches from the very beginning, not to address a crisis, but to preempt one.

“Three quarters of companies fail within five years for financial reasons. And the same proportion fail because founders couldn’t get along. We didn’t want to fall into those traps.”

They still do it today, now extended to the full executive team.

8) What they’d do differently

Some client engagements were taken just for revenue, pulling the product too far from its core. Some partnerships didn’t deliver. They went international too early and not carefully enough.

“There were things I would have avoided. But I also think we were right to go all in.”

The takeaway

Deeepki’s story doesn’t follow the startup playbook. No viral launch. No obvious market pull. No regulatory wave pushing them forward.

What they had was a structural insight, a disciplined approach to building around real customer problems, and a co-founder relationship they treated as seriously as the product itself.

Thirty clients. One million in ARR. Self-funded. Then a raise. Then a proper SaaS. Then the regulation caught up and the market they’d been building for finally arrived in force.

“At 400 people, we’re less agile than when we were two. But we have far more impact.”

1) The structural problem that no one had bothered to solve

In 2013, Thibault Lauzeral was trying to set up a phone system at Fotolia. What he found was a landscape of expensive, legacy, on-premise hardware, completely disconnected from the SaaS tools the business already ran on. Nobody had built a cloud-based business phone system that integrated natively with CRMs and help desks.

The idea that came out of this experience was narrow and concrete: let companies spin up a call center anywhere in the world, in minutes, connected to the tools they already use. Not another telecom. A software product that sat on top of existing telecom infrastructure.

This framing (software, not telco) would turn out to be one of the most consequential decisions Aircall ever made. Years later, Jonathan reflected: “If we had positioned ourselves as a telco, our valuation would be terrible. By staying pure software, we built something completely different.”

“We decided from day one: we are not a telco. We are software. And history proved us right.”

2) The months before launch: a free MVP and a founding disagreement

The first version of Aircall was free and stripped down to almost nothing: you could receive calls. No outbound. You got a phone number, a basic greeting message, and hours of operation. That was it.

Within a month and a half, they had about a hundred users, mostly small businesses that wanted a professional-looking phone presence without a landline. The team watched how people used it, then added outbound calling.

Then came the first real test of the founding team: should they charge? Jonathan pushed back. The product felt too immature. His co-founder Olivier disagreed. They switched to paid. Fifty percent of free users churned immediately.

Jonathan now describes that moment as one of the most valuable early lessons. The half who stayed had a genuine problem to solve. Their feedback was actionable in a way that free users’ never was. “When people pay, what they tell you actually means something. Free users can say anything.”

“Once people pay, their feedback is real. Free users can tell you anything. You have no idea if it actually matters.”

3) The human moment: selling on Skype, scrapping the Web Summit

Before there were sales processes, there was Jonathan doing demos on Skype, sharing his personal account ID with prospects and hoping they’d connect. This was 2014. Zoom barely existed.

For lead generation, he scraped the full attendee list of Web Summit: 3,000 names, companies, and guessed email addresses constructed as firstname@companyurl.com. He emailed all of them, claiming to have met them at the event. Some replied.

He also built prospect lists by scraping every website in France and Europe that displayed a phone number (indicated by a “+” in the HTML). If they had a number, they were a potential customer. “We had tens of thousands of prospects this way. It was completely artisanal, but it worked.”

Alone, he reached €10K MRR. Then two interns joined. Then came the TechCrunch article following their first fundraise of €900K, and 20 new sign-ups appeared overnight.

“I emailed 3,000 people from the Web Summit pretending I’d spoken to them. Some responded. That’s how we got our first customers.”

4) First clients: what those early deals actually looked like

Early Aircall contracts were priced at €20/month for three users. That’s €6.66 per seat. Closing ten companies in a week meant €200 in new MRR.

The economics were painful but revealing. The customers who stayed at that price had a real need. The company’s ICP at this stage was clear: startups and SMBs who wanted a professional phone system without the friction of traditional telco. Companies that already had a number on their website but nothing real behind it.

Jonathan’s sales motion was relentless outbound: cold email, “warm calling” (which meant the prospect had received at least two emails before being called), and constant iteration on messaging. He describes it as “cambouille”: getting your hands dirty, grinding through the work no one else wants to do.

One of his most counterintuitive findings from this period: most of his “inbound” leads were actually outbound-seeded. Prospects who had received his cold emails but never responded would show up later as self-serve sign-ups. The outbound created awareness; the conversion happened silently.

“70% of what I thought was inbound was actually people who had received a cold email weeks earlier and signed up on their own.”

5) Why people bought before the product was obviously good

For four years (from 2014 to 2018), Aircall’s voice quality was poor. NPS was around -60. Customers would tell prospects it didn’t work. Jonathan openly walked into sales meetings knowing the word-of-mouth was negative.

And yet they closed deals. Revenue grew from €1M to €3M to €9M to €18M ARR, year over year.

Three things made this possible. First, the market was pulling: nobody had a better cloud-based alternative. The timing was exceptional. “The market carried us. Like saying Anthropic isn’t carried by AI. It’s the same thing.”

Second, the product’s value proposition was genuinely compelling even when the voice quality dropped. What Aircall could do (spin up a multi-user call center anywhere, integrated with Zendesk and Salesforce) was new enough that customers tolerated the frustration.

Third, and most surprisingly, the customer success team. Jonathan describes clients who wanted to leave but stayed because of how Aircall’s team treated them. “We had clients say: ‘I want to leave, but your customer success team treats us like family. They roll up their sleeves and fix things.’ That team saved us.”

“Customers told us: ‘I hate you as much as I love you. I want to leave but I can’t. Your product does things nothing else does.’”

6) The product evolution: from broken voice to the integration layer

The moment Aircall became a different company was in San Francisco, in the summer of 2015, during their 500 Startups cohort. US customers were asking for something French customers hadn’t pushed for yet: native integrations with their CRM and helpdesk tools.

In France, the SaaS ecosystem was still catching up. In the US, Zendesk was exploding. Salesforce was everywhere. Customers wanted a phone call to automatically log into the contact record. They wanted call outcomes to sync with their pipeline.

Aircall built the integration layer. Partners (Zendesk resellers, Salesforce consultants) started recommending Aircall to their customers. The company went from €10K to €50K MRR in four months, ending the year at €60K MRR.

Jonathan remembers closing that year’s target at 8pm on December 31st. “We hit exactly what we said we’d hit. That created something in the team: when we say we’ll do something, we do it.”

“In the US, we realized the product was missing its real layer: integrations. That was the unlock.”

7) The founding team dynamic: roles, equity, and the tractor tattoo

Four co-founders, clearly separated roles from day one. Olivier handled product and finance. Pierre-Baptiste did backend. Xavier did frontend. Jonathan owned revenue: go-to-market, sales, everything commercial.

The clearest expression of the founding team’s culture came in 2016, during a particularly dark stretch. The product wasn’t working. They went to get a kebab from the corner shop, came back, and made a pact: if the company ever reached €100M valuation, they’d all get a tractor tattoo.

The tractor had started as an emoji. A departing intern had signed off with a 🚜 in a Slack message, and Jonathan had felt something. “We’re not a rocket. A rocket hits its target once every ten tries. We’re tractors. We till the ground, we do the work, we get there.”

By 2018, the company hit €100M. They went to Corsica and found a tattoo parlor. Jonathan (who hates tattoos) spoke Bulgarian to the Corsican tattooist and asked him to dilute the ink. He got it on his hip, where no one would see it. Olivier got a tractor the size of a fist on his ankle.

“On se tatoue un tracteur. We are not a rocket. Rockets hit their target one time in ten. We till the ground.”

8) What he’d do differently: sell yourself, ignore the noise

Jonathan’s most direct advice: founders who refuse to sell are making a self-destructive choice. He watched too many CEOs delegate sales to avoid looking like a small company. He did it himself at his next venture, Offstone, and caught himself in the mistake in real time.

His second piece of advice: ignore geographic constraints. Aircall built its first €30K of MRR without leaving Paris. The US expansion accelerated things, but wasn’t the cause of product-market fit. “You can be in Clermont-Ferrand, build a team, prove it works, generate revenue. Geography is in your head.”

He’s also skeptical of the pressure to grow fast. Aircall’s path from 0 to €1M ARR took two years. “Don’t tell yourself your company is bad because someone else did 0-to-1M in three months. Focus on your product and your commercial execution.”

On the question of overwork: Jonathan is honest that Aircall ran seven days a week for years. But he’s careful not to present that as the secret. “Working more makes you better faster. But I don’t think success requires 24/7 grind. What it requires is resilience: showing up every day for years, even when you’re not sure it’s going to work.”

“C’est pas le meilleur produit qui gagne, c’est malheureusement le meilleur vendeur. The best seller wins, not the best product.”

The takeaway

Aircall’s 0-to-1M story is not a story about product excellence. It’s a story about what happens when you combine an anomalous market window, a sales culture built from scratch, and a founding team that treated customers like family even when the product let them down.

The NPS of -60 for four years is the detail that most founders skip over, because it’s uncomfortable. It forces the question: what actually drives early revenue? The answer from Aircall is blunt: timing, relentless outbound, the right integrations at the right moment, and a customer success team that made clients feel too taken care of to leave.

What Jonathan calls “being a tractor” isn’t just a metaphor for hard work. It’s a philosophy of compounding. Tractors don’t get airborne. They cover ground, row by row, and the harvest comes. The decade-long trajectory of Aircall (from a €6.66 SaaS with a broken voice layer to a billion-dollar company) is what that philosophy looks like at scale.

Read up on the 10 takeaways from the episode:
1) The real problem wasn’t spending - it was how decisions were made

Before Spendesk existed, Rodolphe had already lived through the problem inside fast-growing companies. Teams were encouraged to move quickly and take ownership. That autonomy was essential to growth. But the moment money entered the equation, everything slowed down.

Finance, understandably, needed control. Visibility over cash, approval processes, compliance. The result was a system where every expense had to be validated and justified - often manually. And as the company scaled, that system became heavier and heavier.

At some point, the contradiction became obvious.

“I was rebuilding the exact bureaucracy I had hated before, just to regain control.”

That’s where the real insight came from. The issue wasn’t that employees were spending too much, or that finance lacked tools. It was that companies were trying to operate in a modern, decentralized way with systems designed for centralized control.

Spendesk was conceived as a way to resolve that structural mismatch: to give teams the ability to act, while allowing finance to keep a clear, real-time view of what was happening.

2) The first version failed and clarified everything

Like most early-stage products, the first version of Spendesk was far from what it would eventually become. The initial idea was to solve a very specific pain point: the sharing of company cards.

The solution they built was simple. A browser extension that allowed employees to access a shared company card, while logging who used it and when. It worked, technically speaking. But it didn’t solve the core issue.

Finance teams didn’t trust it. Security concerns remained. And most importantly, the product sat on the surface of the problem rather than addressing its root.

That failure turned out to be decisive. It forced a realization that would shape everything that followed: you can’t build a meaningful finance product if you don’t control the transaction itself.

In other words, visibility without control isn’t enough.

3) Building under constraints and learning faster because of it

At that stage, nothing was in place to make things easier. There were no obvious payment partners in Europe, no infrastructure readily available to issue cards, and no fully formed technical team to build the product quickly. Rodolphe himself went back to coding to push things forward, even though he hadn’t done it in years.

“I would spend four days building what a good engineer would do in one hour.”

From the outside, this looks inefficient. But in practice, it created something far more valuable than speed: proximity to the problem.

Eventually, they found a partner that allowed them to issue virtual cards. It wasn’t perfect, and the setup introduced friction that would normally kill adoption. But it was enough to test the next hypothesis.

4) The first real signal: customers taking irrational risks

The early product allowed companies to generate virtual cards for online purchases. To use it, however, customers had to pre-fund accounts - through a structure that was far from reassuring.

They had to wire money to Gibraltar.

And yet, customers did it.

That moment revealed something fundamental: the pain was strong enough that companies were willing to accept friction, uncertainty, and risk to solve it.

When users behave in ways that seem irrational on paper, it’s often because the problem you’re solving is much bigger than you thought.

5) The second signal: usage that grows instead of fading

If the first signal proved that the problem existed, the second showed that the solution was working.

Typically, early SaaS products see a burst of activity followed by decline. Users test, explore, and then slowly disengage when the product doesn’t integrate deeply enough into their workflows.

That didn’t happen here.

Week after week, usage increased. More transactions, more teams involved, more use cases emerging organically inside companies. What started as a tool for marketing teams quickly expanded to other departments, without any push from the product itself.

“Normally usage declines. For us, it kept increasing.”

6) The final signal: willingness to pay - without friction

Up to that point, everything had been free. Which meant one critical question remained unanswered: would companies actually pay for it?

When they introduced pricing, the expectation was cautious. Losing a significant portion of users would have been normal, even healthy. Instead, almost all customers converted.

Out of around a hundred companies, only a handful chose not to continue.

“We thought we would lose half. Almost everyone stayed.”

7) When investors don’t believe what customers already know

With these signals in hand, fundraising should have been straightforward. It wasn’t.

For several months, the response from investors was largely negative. The core idea — giving employees direct access to company spending — went against deeply ingrained assumptions about financial control.

“Companies will never give cards to employees.”

The breakthrough came when they changed their approach. Instead of trying to convince traditional venture capital, they turned to people who understood the problem directly: founders, operators, and early customers.

Within three weeks, the round was closed.

8) Building in Europe means building complexity from day one

Spendesk’s early development was rooted in Europe, largely because that’s where the founders had access to their first network and customers. But this choice came with a structural challenge that shaped the company’s trajectory.

Unlike the US, Europe is not a single market. Expanding means navigating multiple countries, each with its own regulations, behaviors, and operational constraints. What works in one market doesn’t automatically translate to another.

“You end up duplicating your effort in each country.”

This makes growth slower and more complex, but it also creates a different kind of strength. Companies that succeed in this environment build resilience early, and develop systems that can handle fragmentation.

9) The role of intuition when everything becomes less clear

Looking back, one of the most striking lessons doesn’t come from the early product decisions, but from how decisions evolve as the company grows.

In the beginning, founders rely heavily on instinct, simply because there are no established rules to follow. As the company scales, that instinct is often diluted by external input — advisors, investors, benchmarks, best practices.

That shift can be dangerous.

“Every time I ignored my intuition, it was a mistake.”

The challenge isn’t to reject advice, but to avoid overfitting to it. The ability to filter signal from noise becomes a critical skill, especially when navigating uncharted territory.

The takeaway

What makes Spendesk’s early story interesting isn’t just the outcome, but the sequence of signals that shaped it.

A problem that felt operational turned out to be structural. A failed MVP clarified what truly mattered. Customers validated the product not through feedback, but through behavior — by taking risks, increasing usage, and ultimately choosing to pay.

Behind it all, there’s a consistent pattern: progress didn’t come from perfect planning, but from continuously testing assumptions against reality, and adjusting quickly when they didn’t hold.

In the end, building a category isn’t about having the right idea from the start.

It’s about recognizing, early enough, when you’re onto something real and having the discipline to follow it all the way through.

For this episode of PMF Collector, Matthieu Vaxelaire sits down with Fabrice, co-founder of Yoobic — a platform used by retail networks to manage store operations, from communication and training to daily execution across hundreds of locations.

Built with his two brothers, Yoobic didn’t start with a product idea.

It started with a conviction: most people don’t work behind a desk — yet the tools built for them are still designed for headquarters.

What follows is not a story about luck or timing. It’s a story about method.

You can watch the full conversation above, or read the key themes below.

1) Starting from a conviction — then forcing it through reality

Yoobic didn’t begin with a feature.

It began with a simple observation: the vast majority of workers are “deskless”, yet almost no software is designed for them. That gap was the starting point.

But instead of jumping into building, Fabrice and his co-founders approached it methodically. First, identify a technological shift — in this case, mobile. Then apply it to a vertical where the gap is obvious.

Retail quickly emerged.

Still, a top-down conviction wasn’t enough. It had to survive contact with reality.

2) The only way to understand a problem: live it

So Fabrice went on the ground.

He spent weeks shadowing store managers, observing how they worked day to day. What he found wasn’t inefficiency — it was chaos. Hundreds of emails from HQ, constant onboarding due to high turnover, fragmented tools, and processes that simply didn’t scale.

In one case, sales reports were still being sent by fax.

“When you’re inside the system, you don’t see how broken it is. From the outside, it’s obvious.”

This wasn’t a feature gap. It was a structural problem.

3) Narrowing down to one painful, recurring process

Instead of trying to fix everything, they made a key decision: go extremely narrow.

They focused on a single process — executing promotions in stores.

Not because it was glamorous, but because it met three critical conditions. It happened frequently, it had a direct impact on revenue, and someone inside the organization was accountable for it.

That was enough.

“If you try to solve everything, you don’t have an MVP anymore.”

So they built only that.

4) The only validation that matters: a paying customer

From the beginning, Fabrice had a clear rule: no free validation.

The MVP had to be paid.

Even if the amount was small, the act of paying forced a different level of commitment. It meant budget approval, internal alignment, and a real problem being solved.

After a short pilot, the signal was immediate.

The client came back and said: “I want this in all my stores.”

That was the first real proof.

“The amount doesn’t matter. The fact that they pay does.”

5) Growth came from focus, not expansion

Retail is a huge market. Too huge.

So instead of expanding, they kept narrowing.

Through repeated sales conversations, they refined their positioning, their messaging, and their target. Eventually, they landed on fashion retail — a segment where operational complexity is high and execution directly impacts revenue.

This focus made everything easier: clearer messaging, faster sales, stronger traction.

“The more precise your target, the easier it is to sell.”

6) Customers as the only real growth engine

In the early days, Yoobic didn’t rely on paid acquisition.

They relied on their customers.

Case studies, testimonials, conferences — in a small ecosystem like retail, word spreads quickly. Once a few brands adopt a solution, others follow.

This created a compounding effect.

“Your best marketing is your customers.”

7) Enterprise sales: slow, but predictable

Selling to large retail networks meant dealing with long cycles — often six to nine months.

To make this work, Yoobic built a structured approach around proof. Start with a limited deployment, measure the impact, and turn that into a quantified business case.

Once the ROI is clear, the conversation changes.

“If you can prove ROI, you don’t need to wait for next year’s budget.”

8) The US expansion: rebuilding everything except the product

After finding traction in Europe, Yoobic expanded to the US.

But almost nothing transferred.

No brand, no references, no network. So Fabrice made a deliberate decision: start from zero again. Small team, narrow ICP, outbound-driven approach.

The product didn’t change.

Everything else did — from messaging to pricing to the sales process.

“Just because it works in Europe doesn’t mean it works in the US.”

It took months to land the first clients. Then momentum came back.

9) A company is built through iterations, not plans

At every stage, the approach stayed the same.

Form a hypothesis, test it quickly, define success in advance, and iterate. Short cycles. No endless discussions.

First, open debate. Then, once a decision is made, full execution.

“You need both: open debate first, then ruthless execution.”

This discipline compounds over time.

10) Keeping the founder mindset at scale

Today, Yoobic is a large organization.

Which creates a tension: you can’t operate like an early-stage startup anymore — but you can’t lose that energy either.

Fabrice’s solution is to carve out small spaces to keep experimenting. Not on the core business, but on the side. Fast loops, quick iterations, then reintegrate what works.

“If I do this on the core business, I break the company.”

It’s how he keeps the “day one” mindset alive.

11) Persistence vs. stubbornness

Behind everything, there’s one constant tension.

You need persistence to push through uncertainty. But too much of it becomes stubbornness — and keeps you going in the wrong direction.

The real skill is knowing when to adjust.

“You need to keep going — but not in the wrong direction.”

This episode is a masterclass in building vertical SaaS from the ground up: starting with a real-world problem, validating through payment, focusing relentlessly, and rebuilding your playbook at every stage.

For this episode of PMF Collector (ex Start Mode), Matthieu Vaxelaire sits down with Émeric Ernoult, co-founder & CEO of Agora Pulse - the #1 social media management tool in France, and the 5th largest worldwide, built without raising a single euro.

It took 15 years. A failed first company. Two men working evenings and weekends for a decade. A co-founder relationship that survived 24 years, therapy included.

But what makes this conversation worth your time isn’t the destination. It’s the honesty about how ugly the road actually was.

You can watch the full conversation above, or read the key themes below.

1) The co-founder you didn’t choose — and why it worked anyway

Émeric didn’t meet Benoît at a startup event or through a warm intro.

They crossed paths at a student house in New York in 1993, during their first internship in the US. Years of reconnections followed — Paris, Washington, chance encounters through Benoît’s twin brother — until 1999, when both were back in Paris, restless, and caught up in the internet bubble.

“We made three PowerPoint slides and thought we’d become multi-millionaires.”

They didn’t. But they stayed together for 24 years — longer than Émeric’s marriage.

2) Ten years of evenings and weekends — and why he doesn’t recommend it

From 2000 to 2010, both founders kept their day jobs.

Émeric worked at a law firm. Benoît at a web agency. Their startup — Affinity, a white-label community platform — ran on whatever hours were left. Morning commutes, Saturday afternoons, Sunday evenings.

At peak, the business was doing €150K in annual revenue. After a decade.

“It should have died. We should have moved on. It wasn’t normal what we did.”

He doesn’t romanticize it. The model was broken — every client needed custom development, it was agency work dressed up as a product, and it never scaled. But something kept them going: the belief that the next pivot would be the one.

3) The insight that created Agora Pulse: sell what people use every day

In 2010, they stumbled into Facebook contest apps by accident — a campaign for the FIA featuring Sébastien Loeb went viral, and suddenly brands were calling.

They sold bespoke Facebook contests at €15,000 each. Within a year, fifty agencies were doing the same thing in Paris, and prices collapsed to €6,000. Then lower.

Meanwhile, they noticed a US startup called Wildfire productizing the exact thing they were doing manually — turning it into a self-serve platform. Google later acquired it for $450M and shut it down within a year.

The real question wasn’t contests. It was: what do community managers actually use every day?

“We asked ourselves: who are our buyers, and what do they need constantly — not once a year?”

The answer was a publishing and engagement tool. And that became Agora Pulse.

4) The ultimatum: 1,000 customers at €49/month — or we shut it down

By late 2011, Émeric was done with uncertainty.

He set a hard target: 1,000 paying customers at €49/month within one year. If they missed it, he was walking away. He’d spent too long, earned too little, and couldn’t keep convincing himself the breakthrough was coming.

It took three years, not one. But every single month — except December and August — the number went up. Three new customers. Four new customers. Slow, but always forward.

The thing that kept him going was a talk by Gail Goodman, founder of Constant Contact, called The Long, Slow SaaS Ramp of Death. Her message: the only question that matters is whether you’re moving in the right direction. Not how fast.

“Every month it progressed. Two months a year it went backwards. That was enough to hold on.”

5) Bootstrapped to €23M ARR — and what that actually costs

Agora Pulse never raised a meaningful round. A €250K grant from a public fund. That’s it.

By 2015 they crossed €1M ARR and reached profitability. By 2016, they grew 150%. By 2019, 90%. Today the company sits around €23M ARR, with 160 employees across France, the US, and internationally.

But Émeric is clear: the growth wasn’t linear. 2017 was nearly flat — Benoît spent the year rebuilding the entire tech stack from scratch, and zero new features shipped. When you stop shipping, customers stop growing.

“Once you’re profitable, the emotional rollercoaster is less violent. But you never feel like you’ve arrived. That’s the nature of the job.”

6) Building on someone else’s land — the risk nobody talks about

The existential vulnerability of a social media tool is obvious once you say it out loud: everything depends on Meta, Twitter, LinkedIn keeping their APIs open.

Émeric has lived through accidental outages that froze €23M in ARR for days at a time, with no clear escalation path, waiting for an engineer in India to push the right button.

“My next company will not be built in someone else’s garden. I know that for certain.”

The diversification across multiple platforms helps — losing one network doesn’t kill the business. But the dependence on Meta in particular remains a structural risk he carries every day.

7) Twenty-four years with one co-founder — the honest version

Émeric and Benoît never did the thing advisors tell you to do: write down what each of you expects, read it back, sign at the bottom.

They fought constantly. They went through stretches of barely speaking. At one point, they hired a coach — couples therapy, essentially — to get back to a functional working relationship.

What saved them wasn’t process. It was a deliberate decision Émeric made about how to look at his partner.

“I decided to look at him as the person without whom none of this exists. Full stop. The rest doesn’t matter.”

His advice to founders pairing up today: write down your expectations before you start, revisit them every quarter, and give each other actual feedback. He admits he never did any of it properly. That’s exactly why he knows it matters.

8) The one thing he’d fix if he started over

Émeric Ernoult’s answer comes from the book he’s writing.

The biggest mistake he made — consistently, across fifteen years of hiring — was bringing people on without ever defining precisely what success looked like for them. Not a vague job description. Not a list of KPIs. A specific, co-authored document that the person reads, agrees to, and can be held accountable against.

He calls it a What Does Success Look Like document.

“We spend so much time hoping people will figure it out on their own. They don’t. Because we never told them what ‘good’ actually means in their role.”

He’s been running this at Agora Pulse for eight years. He believes it’s the single biggest lever most founders leave untouched.

Share

Watch on YouTube for English subtitles (settings → auto-translate → English).

For this episode of Start Mode, Mat Vaxelaire (Partner @ Hexa) sits down with Jérémy Goillot, founder & CEO of Mobile First Company: a startup with a rare level of ambition: rebuilding the software stack used by millions of SMEs, through 40 mobile-first and AI-native applications.

It’s been only 18 months.

In that time, Jérémy raised 12M€, shipped Allo (a modern take on business telephony), signed 5,000+ customers and moved the company to the U.S.

But what stands out in this conversation isn’t the traction.

It’s the way Jérémy works: fast, intense, obsessed, and very honest about what it costs.

You can watch the full conversation above, or read the key themes below.

1) A founder defined by intensity, not credentials

Jérémy didn’t grow up in Paris or SF.

He grew up in the Pyrénées Ariégeoises, in a family of hard-working employees who never asked for raises, never pushed for more.

He saw what “lack of ambition” looks like.

He decided he wanted the opposite.

“If I didn’t work twice as much as everyone else, I’d stay stuck. It was that simple.”

This intensity is how he built his career, how he attracts talent, and why he pushes Mobile First.

2) The insight behind Mobile First: the SME stack is old, slow and overpriced

The initial thesis was almost obvious once you saw it:

• SMEs use outdated tools

• products are designed for desktop workflows

• mobile usage is where attention actually is

• AI destroys price structures overnight

Jérémy wanted to build tools that feel consumer-grade but run business operations.

Then ChatGPT arrived and confirmed the intuition: software pricing, onboarding, activation — all of it was about to be rewritten.

Mobile First set out to rebuild the suite.

3) The 40-app idea and why it isn’t crazy

Most founders would try to get one product right. This team is building forty.

Why?

Because SMEs don’t want 15 vendors.

They want:

• the same UX everywhere

• predictable pricing

• simple onboarding

• tools that talk to each other

And also because, as Jérémy says bluntly:

“I don’t want to spend seven years on one feature. I need a place where the product cycles keep moving.”

Mobile First is designed as a multi-product company from day one.

4) The first product was wrong. The second was wrong. The third clicked.

The early days looked nothing like the current product.

• They started with inventory management → not good enough.

• Then a voice AI answering assistant → people weren’t ready to pay.

• Then they discovered the real frustration: business telephony.

Teams wanted:

• call recording

• SMS

• transcripts

• AI notes

• CRM updates

• and a modern UI

So they glued together everything they had built and Allo was born.

5) The cofounder mistake (and the uncomfortable fix)

One of the hardest parts of the discussion: Jérémy picked the wrong CTO.

He granted equity without a cliff. Two months in, he knew it wouldn’t work.

He had to undo everything:

• buy back shares

• tell investors

• tell the team

• restart the technical foundation

No spin, no sugarcoating.

“It was my first big decision as a founder. I got it wrong.”

It almost killed the company, but making the decision early saved it.

6) PLG wasn’t enough — sales showed them why they win

Mobile First started as a self-serve SaaS. Thousands of signups. Lots of usage.

But not enough revenue. Churn was high. Cohorts weren’t healthy.

Hiring the first salesperson changed the picture:

• 46% close rate

• bigger accounts

• expansion inside teams

• a clear sense of who they outperform

“PLG tells you everything that breaks. Sales tells you why customers choose you.”

7) The actual job of a zero-to-one CEO

With a coach’s help, Jérémy defined his role through five responsibilities:

1. Communicate — clearly, constantly, inside and out

2. Maintain product quality — nothing ships below the bar

3. Drive distribution — ensure a predictable flow of customers

4. Hire and fire — raise the average quality of the team

5. Own revenue — by selling or by fundraising

Every Monday, he rewrites these five principles.

8) Ambition requires energy and energy requires structure

The intensity has a cost.

To avoid burning out or burning his team, Jérémy leans on:

• sports

• routines with his partner

• leaving the office early and working later from home

• weekend escapes outside Miami

• reducing emotional volatility

9) Advice for founders in the 0→1

He believes founders underestimate how much work it actually takes and overestimate how much balance they can keep while trying to build something meaningful. His view is simple: if you want to create an ambitious company, something has to give. You need to correct mistakes fast, even when it’s uncomfortable. You need to stay close to customers every single day because it’s the only place where the signal is real. And you need to accept that the emotional swings — the days where you think you’ve cracked it followed by the days where you want to burn everything down — are part of the job.

Watch on Youtube for English subtitles (settings > auto-translate > English).

In this new episode of Start Mode, Mat Vaxelaire (Partner @ Hexa) sits down with Evan Testa, cofounder & CEO of Roundtable, the platform that structures and automates private market investments — from SPVs to full investment funds.

In just two years, Roundtable has become a reference in Europe’s private markets investing, structuring nearly half a billion euros and scaling through a unique viral loop embedded in the product.

Unlike many founders who struggle to sell, Evan has a rare “superpower”: the ability to build trust instantly. He doesn’t “sell”, he listens, advises and brings people with him. It’s this ability, combined with deep resilience, that shaped Roundtable’s zero-to-one.

This conversation breaks down how Roundtable found its wedge, why the MVP had to be legally perfect before being technical, and how a trio of founders stays aligned under extreme pressure.

Watch the full episode above, or read the key takeaways below 👇

1) The superpower: selling without selling

Evan’s strength has always been human connection.

“I’m never trying to sell my product. I’m offering advice. The product follows naturally.”

He learned this early, from selling on open-air markets at age 13, waking up at 5am every weekend. It shaped his resilience and his instinct for reading people.

This trust-first mindset became fundamental: in private markets, credibility is everything.

2) The pivot before the launch

The original idea wasn’t Roundtable.

Evan and his cofounders wanted to build a platform that democratized access to private equity funds. Then a competitor raised €100M before they even shipped.

“We told ourselves: we’re not going to spend our lives running behind others.”

They pivoted instantly, before writing a real line of code, and discovered an unexpected pain: structuring, not distribution, was the real bottleneck.

3) A legal MVP — not a technical one

Roundtable operates in a world with zero tolerance for error.
A buggy SPV? A misstructured vehicle? Unthinkable.

“Our MVP wasn’t product. It was legal. We spent a year making the SPV structure bulletproof.”

Only once the legal and regulatory backbone was perfect did the team build around it, with no-code at first, then real automation.

Meanwhile, they ran 400 discovery calls in two months to pinpoint the real needs of the market.

4) Viral by design: one client brings twenty

One overlooked insight: an SPV is a naturally viral product.

“When you create a vehicle, you bring 10, 20, sometimes 50 people with you.”

Early users discovered Roundtable through other investors inside each SPV, creating a compounding growth loop.

From one SPV a month… to a hundred. This organic virality became Roundtable’s engine.

5) The power of saying no

One of Evan’s strongest convictions:

“Founders must be comfortable saying no.”

Especially in the early days, when every customer feels vital and every request is tempting.

Roundtable refused features, refused custom work, and even refused some customers to protect the integrity of the product and the company’s reputation.

The paradox? Often, after a “no,” the customer still came back.

6) The founder trio: alignment above ego

Roundtable’s three cofounders share a rare dynamic.

“We challenge each other a lot. But once we align, we go all-in, together.”

They invest heavily in communication: weekly founder syncs, multi-day retreats, constant re-alignment on priorities, roles, and decisions.

Their principle: the project comes before any individual. Meaning that you have to put your ego to the side for the benefit of the company.

7) Fundraising with angel-customers

When raising their first round, Roundtable deliberately chose not to raise from VCs.

Instead, they brought in dozens of angel investors who were also users, from multiple European countries.

This created a network of local ambassadors that fueled their expansion.

“It’s the most powerful go-to-market engine we could build.”

8) Advice for 0→1 founders

Evan’s motto, one that we hear a lot but that is actually very hard to implement:

“If there’s doubt, there’s no doubt.”

Whether it’s recruiting, choosing a customer, or evaluating an idea, hesitation is almost always a signal.

His advice:

• Move fast, but with focus

• Protect your credibility like an asset

• Stay close to your users

• And above all: trust your intuition

Watch on Youtube for English subtitles (settings > auto-translate > English).

In this new episode of Start Mode, Mat Vaxelaire (Partner @ Hexa) sits down with Laurent Perrin, CTO and cofounder of Front (HX14), the customer communication platform that reinvented the shared inbox.

Born at Hexa in 2014, Front has since become a global reference in customer communication software.

While many founders eventually drift away from the product, Laurent never stopped coding. Ten years later, he’s still in the codebase, still talking to users, and still learning from what happens on the ground.

In this conversation, he shares how Front made it from zero to one, by operating in tight three-month cycles, listening obsessively to users, and holding the bar of quality impossibly high. It’s a masterclass in how to stay alive, lucid, and product-driven through the most uncertain phase of company building.

Watch the full episode above, or read the key takeaways below 👇

1) From default dead to default alive

At the beginning of Front, Laurent and Mathilde operated in three-month cycles.

“In my head, the project ended in three months. Every day, I asked myself: what do I need to do today so that in three months we decide to keep going?”

This rhythm helped them balance short-term focus with long-term vision, a simple framework to stay grounded through uncertainty.

“If from the start you say the goal is to build a successful company, it’s too far away. Three months is something you can actually see, and you move forward step by step.”

2) Build fast, understand faster

Front’s original concept wasn’t what it is today.

“We wanted to build a simple support tool, but the market we thought existed didn’t really exist.”

By listening to early users, they uncovered a different opportunity, a collaborative inbox that could connect everyone in the company around customer interactions.

“What we’d built ended up solving a much broader need, a tool anyone could adopt.”

3) Mimicking email: blessing and curse

Front’s email-like interface made adoption easy, but also raised expectations sky-high.

“People would get into the product easily, but make shortcuts. They saw an email client, so they behaved like it was an email client.”

That meant zero tolerance for bugs.

“With email, users expect perfection. Everything you think is simple, isn’t.”

That relentless standard created a real moat, and a culture of extreme engineering rigor.

4) When coding is still leadership

Laurent never wanted to become a distant executive.

“I told myself, I don’t want to be just another exec. I want to do the things an exec can’t do.”

By continuing to code, he keeps a vertical view of the company, from the boardroom to the codebase, and stays connected to reality.

“I see things even managers might miss. It gives me a very global perspective.”

5) The cofounder as anchor

His partnership with Mathilde Collin is built on deep trust and mutual respect.

“We trusted each other from day one. If she was working on something, I knew I could go focus on something else.”

They share the same ethic of work and the same conviction that this project had to work.

“There were too many strokes of luck in how we met. If it didn’t work, we couldn’t just ‘redo it somewhere else.’”

6. Advice for 0→1 founders

Laurent’s philosophy: move fast, stay lucid, and don’t cling to bad ideas.

💡 His advice:

• Test fast, kill fast.

• Don’t over-engineer. “For six months, Front didn’t even have a database.”

• Beware of the sunk-cost fallacy: the more you invest in an idea, the harder it is to drop it.

In this episode, Mat Vaxelaire (Partner @ Hexa) sits down with Titouan, CTO and co-founder of Dotfile (HX21), a startup launched with Hexa in 2021 that simplifies KYB management and already powers more than 80 clients such as Spendesk, DeFacto, and Payset.

Titouan embodies resilience and technical ingenuity: from a 12-year-old kid drilling a hole in the wall to plug in his Ethernet cable, to a founder capable of pivoting in the face of negative market signals. He looks back on his journey, from teenage hacks to meeting Hexa, and finally to the decisive pivot that turned Dotfile into a reference for onboarding and compliance management.

💡 What’s inside:

• Turning every constraint into an opportunity — starting at age 12

• First steps in coding and entrepreneurship at 15

• Lessons learned from working as a CTO-as-a-Service

• Meeting Hexa and the early FileX project

• The foundational pivot to Dotfile and finding design partners

• The first “aha moment” with clients like Spendesk and DeFacto

• Behind the scenes of raising funds in a tough macro context

• The importance of shared values between co-founders

• The long-term vision: becoming a true category leader in KYB

🔥 An inspiring episode for entrepreneurs reflecting on resilience, timing, and the ability to pivot in order to build a solid startup.

👇 Let us know in the comments what stood out to you most — and don’t forget to subscribe so you don’t miss the next episodes of Start Mode with Mat Vaxelaire.

Timecodes

00:00 – Introduction and presentation of Titouan
00:36 – Early days in tech and first hacks at age 12–15
03:00 – First entrepreneurial experiments and creative business models
04:40 – Experience as CTO-as-a-Service and key learnings
05:55 – Meeting Hexa and the FileX project
07:40 – Early negative signals and the decision to pivot
09:50 – How the idea for Dotfile was born and validated with pilots
14:20 – First signed client and key learnings
15:45 – The real “aha moment”: convincing major accounts (Spendesk, DeFacto…)
16:30 – When clients can’t work without your product
17:45 – Fundraising in a tough macroeconomic environment
19:30 – Lessons and advice from the 0 → 1 phase
20:50 – The crucial role of co-founder dynamics with Vasco
22:40 – Long-term vision: becoming a KYB market leader

In this Start Mode episode, Mat Vaxelaire talks with Guillaume Marquis, CTO and co‑founder of Basalt (HX24) - a startup helping companies ensure the quality of their AI features.

Guillaume perfectly illustrates the reality of building in AI today:

When OpenAI drops a new model on Friday night, it’s not enough to wait until Saturday afternoon to put it in production.

He shares how speed, precision, and relentless iteration are the keys to keeping your customers in the world of AI.

In this episode:

• Why integrating new AI models in 10 minutes can be a make‑or‑break

• How they closed a client by shipping their requests in just 48 hours

• Pivoting from a product‑oriented MVP to serving technical teams

• The brutal truth of tech hiring: 100 interviews → 2 hires

• The Europe vs US gap: 1 AI feature vs 80 in production

• Growing from 2 to 8 clients in 3 months — with no fundraising

• Finding the balance between speed and quality

🔥 A must‑watch for founders building AI products and grappling with the pressure to ship fast without breaking trust.

In this Start Mode episode, Mat Vaxelaire talks with Théo Rouer, CEO and co‑founder of Elba (HX21) — a startup reinventing cybersecurity for the era of collaborative tools.

From selling at street markets at 14, to running a cybersecurity training agency, to pivoting into a bold SaaS product, Théo shares a rare founder journey: pragmatic, commercial, and impact‑driven.

In this episode:

• How to move from services to SaaS without crashing

• Why an early “commercial superpower” makes all the difference

• How regulation and Covid shaped the cyber opportunity

• The role of a strong cofounder duo and the timing to hire a CTO

• Why Hexa was key to giving Elba early credibility

• The future of security: shadow IT, shadow AI, and co‑pilot tools

🔥 A dense, hands‑on episode for founders building ambitious products in complex spaces.

In this Start Mode episode, Mat Vaxelaire sits down with Hugues, co‑founder and CEO of Tengo, the AI‑powered tool reinventing how B2B companies handle RFPs.

From ultra‑intense discovery work to a makeshift MVP built on Retool, and all the way to their first 1,000 users, we break down every step of the early journey.

In this episode:

• Hugues’ “superpower” that sets great founders apart

• How to run field discovery (100+ interviews, 10 in‑person immersions)

• Launching a product when all you have is… a deck

• How post‑ChatGPT timing shaped their opportunity

• Behind the scenes of their first fundraise

• Why a tight cofounder duo is a startup’s biggest strength

🔥 A goldmine of insights for anyone looking to build, validate, and scale a B2B startup.