Seven Years. Same Question.

I wrote my Master's thesis in 2019 about organisational culture surveys, the kind most companies run once or twice a year, a set of standardised questions, a five-point scale, an aggregate score fed back to leadership a few weeks later as a single number meant to represent how an entire organisation feels about itself.

The argument I made, stripped down to its core, was that these tools claim to capture what's actually going on inside an organisation but end up flattening it instead. A pulse-check survey takes a measurement. It doesn't start a conversation. Employees are asked to rate their engagement on a scale of one to five, and whatever nuance existed in their actual experience, the specific frustration, the thing that would have taken an actual conversation to explain properly, gets compressed into a number that sits somewhere between two and four on a dashboard nobody outside HR will ever look at closely. People wanted a voice in how their workplace was understood. What they got was a score.

The alternative I argued for in the thesis wasn't complicated. It was closer to just asking people directly and letting them answer in their own words instead of someone else's categories. That's a harder thing to aggregate, which is exactly why organisations don't do it at scale. A five-point scale produces a chart. An open conversation produces a transcript nobody has time to read across a thousand employees, so the scale wins by default, not because it's more honest, but because it's more manageable.

I spent the five years after that thesis working in operations, building processes, coordinating launches, the kind of work where you spend most of your time making systems function rather than writing about how organisations should measure themselves. I wasn't thinking about the thesis during any of it. It sat wherever finished academic work sits once you've handed it in and moved on, filed away as something I'd written once rather than something I was still actually working on.

Earlier this year I built Texture Grid, an experience where the only option is to respond, not consume. You're given a word, and you show how it makes you feel through texture instead of describing it in language that's already been pre-approved by whatever survey designed the question. There's no scale to place yourself on. There's no aggregate score at the end.

I didn't connect the two things until recently, going back through old material and noticing the same sentence structure showing up in both. The problem I was writing about in 2019, participation being replaced by measurement, voice being replaced by data, is the exact problem I've spent this year trying to solve, just in a different medium. In 2019 it was an HR department asking employees to rate their engagement instead of asking them what was actually happening. This year it's a platform asking a user to click, scroll, and generate behavioural data instead of asking them to actually respond to anything. The tool changed. The instinct behind the tool didn't.

This isn't only a workplace problem, either. Cultural institutions run the same survey model on their visitors that companies run on their employees, usually without noticing the resemblance. A satisfaction card at the exit, a star rating, a scale from one to five asking how likely you are to recommend the visit to a friend. It produces a number that's genuinely useful for a board report and genuinely useless for understanding what actually happened to someone standing in front of an object that stopped them in their tracks. The instinct is the same one I wrote about in 2019: measure instead of ask, aggregate instead of listen, because a number is easier to report upward than a conversation is.

What strikes me most, looking back, isn't that the problem persisted. Problems persist. It's that I spent seven years circling the same question from two directions that felt completely unrelated to me at the time, an academic argument about workplace surveys and a product built to give people somewhere to actually respond, without noticing they were the same question until I happened to look at both at once. There was no single moment where the thesis was supposed to resurface. It just sat there while I worked on other things, and the only reason I found it again was going back through old material for an entirely different reason and recognising my own argument staring back at me from a different context.

I don't think that's unusual. I think most people are working on the same underlying question for far longer than they realise, in whatever form their current job or project happens to hand them, and the throughline only becomes visible in hindsight, usually by accident, usually much later than it should. It took a Master's thesis and a piece of software built seven years apart for me to notice mine.