What I’m Doing Differently 2nd Time: Product Lessons from Scaling a Profitable B2B Adtech Platform

By Mike Harty, Co-Founder of FunnelFuel Reflections on building smarter, leaner, and more profit-focused in the AI era.

Second Time Around: Building Product the Right Way

My first startup, PowerLinks, came close to being a breakout success. We had top-tier partners, strong market timing, and what could have been a very significant exit to a household name. On paper, it had everything.

Under the surface, we were building the wrong way.

We raised over £10 million from high-net-worths and VCTs. Our burn was at times utterly eye-watering — AWS bills routinely hit low to mid six figures monthly. We ran a second “ghost” payroll just to feed the machine. Every impression we bought cost us money. At peak, we bought <1% of the best 1% of inventory available. The wastage in hindsight was criminal

We built product ahead of revenue, driven by the belief that speculative innovation would create its own demand. That’s the VC model. And sometimes it worked — but only sometimes. It wasn’t sustainable. We chased scale before stability, roadmap before readiness, and funding before focus.

Our funding headway was eroded by the infrastructure burn and we tried to build SO MUCH - ad servers, DSP, ad exchange, enrichment layers and a form of Dynamic Creative. We had our contextual tooling which me and my co-founder Kevin held a patent for. £10m came in drips and drabs over nearly a decade, set against huge costs. Everything - every line of code - was written from scratch.

Powerlinks was a jack of all trades trying to create a channel which was a subset of display really - a feature in a format rather then a channel. That sheer level of spread with under capitalisation was hardly likely to make a great product - yet at varying times it actually very nearly did

We had a talented team and a shot at greatness. But we didn’t get the fundamentals right. When the music stopped, the value didn’t flow to founders or the team. It was destroyed through greed, misalignment, and poor sequencing.

No liquidity. No upside. A masterclass in value destruction.

A challenging backdrop to start again from - yet like a true entrepreneurial lunatic, that’s exactly what I did. But lessons had to be learned

Try #2: FunnelFuel — Built the Opposite Way

Do you really learn when you win? I don’t think so.

FunnelFuel is what happens when you internalise failure and decide to build differently.

We knew from day one:

• Revenue-first

• Customer-led

• Bottom-line positive - control our destiny

• Product-disciplined

• No product without paperwork.

If a client wouldn’t sign for it, we didn’t build it.

That sounds restrictive. And it was — especially early on. But it forced us to build around actual demand. Product became something real people used and paid for. That changed everything.

From Fragile Start to Profitable Engine

Today, FunnelFuel is a profitable, 8-figure revenue platform with a team of 50+. Product and engineering now make up a focused 10-person team — scaled carefully over three years.

We didn’t scale because of funding.

We scaled because our product worked.

And it worked because of how we build:

• Prototypes before platforms

• AI before engineering

• Signals before features

In the product team, we use OpenAI, Claude, Replit, Cursor, and a growing stack of assistive tooling to prototype in days. Clickable demos — not pitch decks — guide internal decisions. These artifacts align sales, strategy, and product in a fraction of the time. We test in-market, prove the need, then invest engineering cycles.

We’re doing more with fewer engineers — 2–3x the throughput of PowerLinks with smarter tooling, tighter process, and clearer ownership.

The Real Product Discipline: Building in a Profit-First Business

This time, the product function doesn’t live in a fantasy.

There’s no blank cheque.

No speculative “wouldn’t it be cool if…” roadmap.

Nothing is passing the roadmap with our some pretty restrictive testing, validating and internal scoring

Everything funnels into support, retention, and revenue.

That makes the job of product harder — but also clearer.

It forces us to sequence ambition.

To align tightly with client needs.

To prioritise speed, value, and measurable outcomes.

It’s not glamorous. But it’s real.

Product vs. Engineering: The Line is Clearer — and Blurrier

At PowerLinks, product and engineering often blended into a single, chaotic build machine. This time, we’ve drawn clearer lines — a seasoned CTO and lead engineer (both CTO-calibre) have created real technical leadership, separating the “what” from the “how.”

And yet, prototyping blurs that line again.

AI tooling lets product own more of the early motion. We turn thoughts into interfaces. We simulate user experiences before code exists. And when it’s time to build for real, engineering isn’t starting from zero — they’re refining and scaling validated needs. They’re solving problems with prototypes - gaps in what they can do, scaling them up, security hardening, making battle (client ready) tooling not pretty UIs

Is FunnelFuel an Adtech Company?

It’s a fair question. And it gets asked enough

We don’t look like most adtech platforms.

And we’re not an agency or a trading desk — even if we sometimes sound like one.

We’ve built FunnelFuel around, on top of, and in tandem with the best in the business:

• Trade Desk & Adform on the buy-side

• Index Exchange & OpenX on the sell-side

• Open source analytics and attribution layers

• Identity infrastructure from ID5 and others

We’re not reinventing the wheel.

We’re making the wheel work better for B2B.

That’s where the real product IP lives - in the layers that unify, augment, and tailor world-class tooling to the unique complexity of B2B programmatic. Adtech and B2B haven’t historically played well together. It’s our job to change that.

We’re not building a DSP.

We’re building an intelligence layer above it — a new way to activate and measure B2B demand.

That makes us a product company. Not a services business. Not an agency. Not just a reseller of media. And every line of code we write reinforces that.

New Problems, Better Problems

At PowerLinks, the problem was always existential.

Would we survive? Would we raise? Would we find product-market fit before burning out?

At FunnelFuel, we face different problems:

• Localising products across APAC, EMEA, and North America

• Navigating privacy regulation from a UK HQ

• Finding data partners who can handle global scale

• Supporting a distributed team across time zones

• Navigating technical blockers in stacks not built for our needs

• Doing it quickly in an organisation scaling very quickly

We originally intended to stay UK-focused. Now we’re live in multiple markets. This global pull was organic — clients took us there — but it introduces real operational and product challenges. Building for international from day one would’ve been smart. We didn’t. So we’re adapting.

But I’ll take these problems.

They’re real. They’re growth problems.

They’re good problems.

What I Wish I Knew Then

PowerLinks taught me:

• Don’t build without validation

• Don’t confuse motion with progress

• Don’t let funding replace fundamentals

• Don’t ignore the economics of delivery

• And never let vision outpace value

FunnelFuel is the answer to that experience. A second-time founder’s blueprint, shaped by mistakes.

We’re not perfect. But we’re pointed in the right direction.

Final Thought: You Don’t Have to Burn to Build

This isn’t just about product. It’s about how you build — and who you build with.

If you’re a founder or product leader trying to do more with less, to scale sustainably, to build something that lasts — I see you.

And I’ll keep writing for you.

This is the second time around. And I’m building it right.