The Ten-Second Problem

It was 17:55 on a Friday, and my children were arguing about whether "pterodactyl" counts as an animal beginning with P. We were playing a word game — the kind that families have been playing on long car journeys for generations. Someone picks a category (animals, countries, fruit), and you work your way through the alphabet: "Antelope, Bear, Cat, Dolphin..." trying to name something for every letter before the other players. It's the sort of thing that keeps everyone entertained for the first hour of a drive to Cornwall and descends into chaos roughly ninety seconds after that.

The game we'd been using had run out of categories. Or rather, it had run out of categories that an eight-year-old and a six-year-old could reasonably attempt. (I have views on whether "European capital cities beginning with X" is a fair ask for anyone, let alone a child who has only recently learnt where France is.)

So I built one. LexiconLeaper is a simple web game — a category, a letter, a ten-second countdown, a button to advance. Nineteen categories ranging from fruit to superheroes to London Tube stops. You can randomise the category, randomise the letter, show hints if you're stuck, and restart when the table descends into the inevitable accusations of cheating. It's written in Vue 3, it's hosted for free, and the source code is open.

I think the whole thing took me an evening to build. The data took longer — it turns out that compiling a comprehensive list of vegetables beginning with every letter of the alphabet is more difficult than it sounds. (Try U. I'll wait.)

And yet.

I mention this not because the game itself is remarkable (it isn't — it's a parlour game with a countdown timer), but because it illustrates something I think about constantly in my day job: the distance between having an idea and having a working thing.

The Distance Problem

That Friday evening, the gap between "we need a better word game" and "here's a working word game" was about three hours. I had a laptop, I had a framework I knew, I had free hosting, and I had two children providing enthusiastic (if not always helpful) quality assurance.

It seems to me that this is the experience we should be aiming for in public sector technology — and it's the thinking behind what we're doing on the National Digital Exchange.

The gap I see most often in local government is not a gap in ideas. Councils are full of people who know exactly what they need: a chatbot that can answer residents' planning queries; a tool that redacts FOI responses; a dashboard that makes sense of their service data. The gap is between that idea and being able to try it. Procurement takes months. Cloud platform requests take weeks. By the time someone has written the business case, the person with the idea has moved on to the next crisis.

NDX:Try is designed to close that gap. We give local government organisations free cloud sandboxes — no procurement, no cost, a working environment in fifteen minutes. Over 50 organisations are currently using it to explore scenarios from council chatbots to planning application AI to FOI redaction tools. Every use case is shared openly so others can learn from it.

I think the parallel to my Friday evening word game is not as silly as it might seem. In both cases, the quality of the outcome is determined almost entirely by how quickly you can get from idea to experiment. My children didn't need a product roadmap for LexiconLeaper. They needed something they could play with immediately and tell me what was wrong with. ("Daddy, there's no Tube stop beginning with X." "That's because there isn't one." "Then why is it showing us X?")

The councils using NDX:Try don't need a twelve-month digital strategy before they can test whether a chatbot would reduce call centre volume. They need to try it for a fortnight and find out.

The Compound Effect

However, I believe the really interesting thing happens after the experiment. When my children played LexiconLeaper, they discovered that they knew far more vegetables than they thought they did (and far fewer currencies). The game didn't teach them anything new — it revealed what they already knew and where the gaps were.

I think NDX:Try does something similar for councils. The experiment itself is often less valuable than what it reveals about the organisation's readiness, appetite, and capability. A council that deploys a chatbot in a sandbox learns not just whether chatbots work, but whether their team can configure one, whether their data is in the right shape, and whether their residents would actually use it. That knowledge is worth considerably more than the experiment itself.

And when those findings are shared openly (as all NDX:Try use cases are), the compound effect is significant. We stop being 400 councils each discovering the same things independently, and we start being a sector that builds on each other's experience. This is also why we've been cataloguing 24,500 UK government repositories and building semantic search across them — so that the next team with an idea can find out who's already tried it.

Back to the Kitchen Table

LexiconLeaper has had a few hundred plays since I built it (mostly from my own family, I suspect, though the analytics suggest at least some strangers have found it). It's not going to change the world. But it solved a real problem, it took an evening, and it's sitting there in the open for anyone who wants to use it or improve it.

Perhaps we should make it that easy for everyone building public services.

If you're in local government and you've got an idea you'd like to try — or if you just want to settle an argument about whether pterodactyl begins with P — the links are below.

The Distance Problem

The Compound Effect

Back to the Kitchen Table

Links