From Products to Profits: The Startup Journey Through Tepilo, Twilert, and ScreenCloud's Nine-Figure Exit

The Trigger: From Services to Products

We'd been running the agency for five years and wanted something different. We'd always talked about products. Luke had created Red5 (a Flash media server that Facebook ended up using), and we kept building the same things for different clients. Seemed inefficient.

Two things pushed us to change. First, I went to Future of Web Apps conference in 2007. Few hundred people in a room, Kevin Rose and the Etsy founders talking about building products that scaled globally. Made agency work feel small.

Second, the 2008 financial crisis. Teachers TV, our big government client, got cut overnight. Budgets disappeared. Half the team quit. We had six developers in Bangkok with nothing to do, still on payroll for six months due to contracts. Instead of letting them go, they taught themselves mobile development.

We decided to invest the agency profits into building our own products. Time to see if we could actually do it.

Mobile Apps: 19 Million Downloads and Hard Lessons

Our first product was Learn Thai in 2011. Just a language phrasebook app. Nothing clever. But it worked, so we made 20 more languages. Over 10 years, the Learn apps made £1.6m revenue and hit 19 million downloads. User acquisition cost: zero. Apple Store was 80% of revenue. iPhone users actually pay for things. Japanese was the big earner. Business travellers needed it.

Flow for Instagram was our breakout. Instagram didn't have an iPad app, so we built one. Apple featured us, tech press covered it, thousands of users daily. Then Instagram changed their API to protect their ad model. We were done. Platform risk.

But we learned loads. Users don't just want functional apps. Looking at our reviews, after filtering basic words, the most common were "great", "good", and "love". Then "learn", "easy", "helpful", "useful", "simple".

Simple is hard. But worth it.

Kizzu was different. Over 3 million downloads of educational iPad apps for kids under 5. Probably the most rewarding thing we built. Took ages to design, revenue model didn't work, but parents loved it. An education hedge fund bought it for £400k as their first deal. Good timing.

Tepilo: The Sarah Beeny Rollercoaster

Tepilo started with Sarah Beeny calling me. She was our client, known for property TV. "I should be doing something online with property," she said.

We launched a joint venture (should've gone 50/50, but you learn). Sarah was brilliant at PR and distribution. Got her on social media early, she built a massive following. Press coverage constantly. Traffic was good.

One problem: no business model.

So we pivoted. Started as 'for sale by owner' with 8,000 listings but no revenue. Became an online estate agent in 2013, advertising on Rightmove and Zoopla, single call centre, automated everything. The pivot worked initially.

Then Northern & Shell invested £3m to use their media channels. Bad idea. We were first to market but Purple Bricks out-executed us. The board didn't work. Nobody had the business's interests at heart.

Tepilo merged with Emoov and another company. Three businesses with no cultural fit, couldn't raise money as the market shifted. From £100 million merger headlines to complete collapse in 2018.

Right idea, wrong investors, poor execution. But we learned about consumer businesses and why having investors who actually care matters.

Twilert: Our SaaS Education

Twilet started as a client project. Original developer couldn't handle the hosting as it grew, so we took over the IP and infrastructure. Twitter monitoring was hot, and we'd built something people wanted.

2012: Twitter changed their API, killing most Twitter apps. While others gave up, we rebuilt Twilert as a subscription service. It worked. We had a profitable SaaS business.

But we didn't double down. Twilert was successful, but we treated it as a cash cow rather than reinvesting. Maybe we didn't care enough about it, or maybe building on someone else's platform always felt temporary.

By 2018, with ScreenCloud taking off, I sold Twilert to an American solopreneur. It taught us SaaS fundamentals. That was worth more than the money.

ScreenCloud: The One That Changed Everything

ScreenCloud came from frustration. Luke wondered why he couldn't easily develop apps for screens. I wanted our office screens to do more than show PowerPoint. Meanwhile, Amazon Fire TV and Chromecast were launching. Consumer hardware that could change digital signage.

The industry was stuck: expensive, complex, sold through resellers who only wanted big projects. Why wasn't this easy? Why wasn't it cheap? Why couldn't an iPad control any screen?

2015: Luke and I were done with agency work. Tepilo was falling apart, I was burnt out. We knew we were good at products. Time to commit properly. We spun out ScreenCloud in September 2015, raised funding by June 2016, and I became full-time CEO.

It took 10 years, but it was worth closing the agency for this.

The journey to April 2024 is its own story, but the results:

• Over $30 million ARR
• Majority buyout for over $100 million
• Every investor got returns
• All employees got equity rewards
• I rolled forward and stayed as CEO
• 10,000+ customers
• 130+ employees across 5 offices

Credit to David and Luke, plus the original UK and Thailand teams. Everyone who worked on it built something special. Through Brexit, COVID, the tech recession. We're still here.

The Product Studio Philosophy

Between 2009 and 2018, we launched dozens of things. Webcam Snapper (Airbnb used it), Slottr, Twilert, Learn apps, Kizzu, Flow for Instagram, loads more. By 2017-2018, we sold everything to focus on ScreenCloud.

Most failed or faded. That's fine.

Running a product studio taught us:

• Volume matters: Can't predict winners
• Resources get stretched: App teams kept getting pulled into client work
• Platform risk is real: Instagram and Twitter killed successful products overnight
• Business model first: Traffic without monetisation is just expensive
• Care about what you build: We didn't care enough about Twilert. We care about ScreenCloud

Agency vs Products: Different Games

Agency: Selling time. Every client matters. You say yes to survive. Customer service is everything. Unhappy clients mean no revenue tomorrow.

Products: Selling value. Many customers. You can say no. But go enterprise and you're back to professional services. Basically agency work again.

The agency years made us customer-obsessed because we had to be. No investor stories or pivot excuses. Either clients are happy or you don't eat. That discipline helped with products.

Lessons from the Journey

Each venture taught something:

• Tepilo: Distribution isn't everything. Need the right investors and a business model from day one
• Twilert: If something works, double down. Half-hearted success is worse than productive failure
• Mobile apps: Platform dependence is dangerous. Emotional connection beats features. Simple is worth fighting for
• ScreenCloud: When you find product-market fit with something you care about, commit fully

The bigger lesson: going from agency to products isn't linear. Every failed app, every pivot, every hard lesson prepared us for ScreenCloud. Skills compound, networks grow, pattern recognition improves.

The Product-Founder Fit

ScreenCloud succeeded where others didn't because we genuinely cared about the problem. We understood the technology. Our agency background meant we understood enterprise needs. We'd learned SaaS from Twilert, distribution from Tepilo, product development from dozens of apps.

Most importantly, we were ready to commit fully. No hedging, no side projects.

You can't fake caring about a problem. When you find what you're good at, what the market needs, and what you actually want to fix, that's when it works.

The journey from that first Thai app to ScreenCloud's exit wasn't planned. But each step taught us something we needed for the next. Sometimes you just start building and see where it goes.

Building products is different from agencies, but one prepares you for the other. After dozens of attempts, thousands of lessons, and millions of users, it's simple: solve real problems, care about them, and when you find something that works, commit properly.