Leaving Domus, Starting Glanaer: Smart Cities Consultancy

Domus.ai didn’t fail dramatically. There was no single catastrophic moment, no running out of money in a blaze of glory, no falling out between founders. It just gradually became clear that the product we were building wasn’t going to find a market in the timeframe we needed.

I’ve written before about the challenges: planning law being more ambiguous than we’d expected, the lack of obvious paying customers, the difficulty of encoding officer discretion into rules. By early 2018, we’d been going for the bones of 18 months and hadn’t found product-market fit. We had a decent prototype, some interesting conversations with local authorities, and exactly zero revenue.

So we wound it down. Quietly, undramatically, like most startups that don’t work out.

What I Learned From Domus

The biggest lesson was about timing and domain fit. The underlying technology (NLP, document parsing, rule engines) was solid. The problem (slow, expensive planning decisions) was real. But the gap between “this problem exists” and “someone will pay for this solution” was wider than I’d anticipated.

Planning officers didn’t want to be automated. Local authorities didn’t have technology budgets. Homeowners wanted human reassurance, not algorithmic confidence scores. And the planning consultants who might have used our tool as a productivity aid were a fragmented market of small firms, each with different workflows and preferences.

I also learned that running a startup while doing a PhD is a bad idea, which I think I already knew but needed to experience to truly believe. The context-switching cost alone is enormous. Your brain doesn’t go from “energy harvesting optimisation” to “planning policy interpretation” without losing something in the transition.

Starting Glanaer

Glanaer started from a simple observation: I’d spent the last few years building expertise in IoT, smart cities, and AI/ML applications, and there were companies that needed this expertise but couldn’t afford (or didn’t need) a full-time hire.

The UK smart cities market in 2018 was an interesting place. There was a lot of government enthusiasm: Innovate UK was funding smart cities projects, the Future Cities Catapult was running programmes, and every council wanted a “digital strategy.” But the actual implementation was patchy. Councils would commission a smart cities strategy document, get a 60-page PDF from a Big Four consultancy, and then have absolutely no idea how to implement any of it because the document was written at 30,000 feet and the council’s IT team consisted of three people who were already busy keeping the email server running.

That’s where Glanaer came in. Practical, hands-on consultancy for startups and SMEs working in the UK smart cities space. Not strategy documents. Not slides. Actual help with things like: which sensors should we use? How do we get our data platform to work with the council’s existing systems? What does the procurement process actually look like? How do we write a bid for an Innovate UK grant?

The Pivot from Product to Services

Switching from product to services is a psychological adjustment as much as a business one. With Domus, I was building something, a product with my name on it, a thing that would exist in the world and (hopefully) grow. With Glanaer, I was helping other people build their things. Less glamorous, but more immediately useful and, critically, people actually paid for it.

The hourly rate for smart cities consultancy in 2018 was the bones of £80-120 for someone with my background, depending on the client and the complexity. Not getting rich, but paying the bills while I finished the PhD. And the work was genuinely interesting. I got to see inside a dozen different companies’ technical challenges, which taught me more about the diversity of the smart cities market than any amount of desk research.

One client was building a noise monitoring platform for construction sites. Another was developing a smart parking system using magnetometer sensors (which don’t work well when it snows, a fact they discovered the hard way during the Beast from the East). A third was trying to sell predictive maintenance for council housing boilers and had built a beautiful ML model that nobody in housing maintenance wanted to use because it required them to change their entire workflow.

That last one was a pattern I saw repeatedly. The technology works. The business case makes sense on paper. But the operational reality (the people, the processes, the politics) doesn’t align. This is the gap that consultancy fills, and it’s a gap that product companies consistently underestimate.

What the UK Smart Cities Landscape Actually Looks Like

From inside, the UK smart cities market in 2018 was characterised by a few things.

First, fragmentation. Every city was doing its own thing. Bristol had a smart city programme. Manchester had a smart city programme. Glasgow had a smart city programme. None of them talked to each other, and all of them had different technology stacks, different governance models, and different definitions of what “smart city” even meant.

Second, procurement pain. Selling to local authorities is slow, bureaucratic, and uncertain. A typical procurement cycle was 6-12 months from initial conversation to contract, with no guarantee of success. For a small company with limited runway, that’s brutal. I watched several promising startups die waiting for a council to complete a procurement exercise.

Third, data politics. Everyone wanted data. Nobody could agree on who owned it, who could access it, how long it should be retained, or what format it should be in. GDPR had just arrived and councils were terrified of doing anything with personal data. Even anonymised sensor data (air quality readings from a lamp post) generated lengthy conversations about data governance.

And fourth, a genuine enthusiasm for improving public services that was often thwarted by lack of budget, lack of skills, and lack of institutional flexibility. The people I worked with in councils were smart, motivated, and deeply frustrated by the gap between what they wanted to do and what their organisation would let them do.

Glanaer operated in this landscape for about a year before I moved on to other things. But the experience shaped my understanding of how technology actually gets adopted in the public sector: slowly, politically, and with far more human challenges than technical ones. That’s a lesson I’ve carried into everything since.