find the ground truth of how developers actually work

Research for
AI era tooling.

An independent practice that helps AI-tooling startups understand the developers they're building for — before a wary, time-constrained market decides their product isn't worth the switch.

The thesis · Jevons' paradox means software must be more user-specific

AI has not reduced demand for developer tools — it has exploded it.

01 · more code

Cheaper, faster generation produces more software, more developers, and more product surface area— not less.

02 · more tools

A flood of new entrants competes for one genuinely scarce resource: a developer's trust and time.

03 · the edge

The winners aren't the teams who ship fastest — they're the ones who understand the developer they are building for and build a product specifically for their one usecase.

The argument

Why research developers when developers build the software?

Because the people building the tool are the worst-placed to see it clearly. Four reasons the team's own intuition is not enough — each one a place adoption quietly goes to die.

01

Expertise is a blind spot, not a window.

The person who built the tool can no longer see it the way a newcomer does. The curse of knowledge hides exactly the friction that decides whether anyone stays.

02

The builder is not the buyer's proxy.

Your team is fluent, motivated, and already convinced. The wary, time-constrained developer you need to win shares almost none of those conditions.

03

"Programmatic" does not mean "unattended."

CLIs, APIs, and SDKs must first be understood by a person — read, mistyped, abandoned. A machine-readable surface should still be a human-readable one.

04

The decisive first hour is invisible from the inside.

Whether a developer stays is mostly settled before they ever file an issue or appear in your funnel. That hour rarely shows up in your analytics at all.

Services · productized for startups

Scoped engagements, built for the way startups actually move.

01

DX Diagnostic

A scoped getting-started and time-to-value audit. Where your first hour leaks, ranked by what it costs you.

2–3 weeks
02

Adoption Research Sprint

Interviews plus a longitudinal study that maps where developers stall, switch, or quit — and why.

4–6 weeks
03

DX Strategy

Findings translated into a prioritized roadmap your engineers and PMs will actually act on.

engagement
04

Embedded DX partner

Fractional, ongoing research and prototyping — a researcher who ships alongside the team.

ongoing
The practice

I'm an independent developer-experience researcher. For more than fifteen years I've designed products; for the last seven, the tools developers build with — leading DX at the observability companies whose entire job is helping expert users diagnose complex systems.

Before product, I trained as a fine artist and worked glass — a long apprenticeship in a single conviction: that even the most technical surface is made for a human hand.

See selected work — dqe1.com →
// let's talk

Tell me what you're building and who it's for.

I'll come back with where I'd look first — no deck, no discovery-call theater. Just the questions I think are worth answering, and how I'd answer them.

Book a discovery call → hello@devexresearch.com
// let's talk

Tell me where to look first.

Four quick things. I read every one myself, and reply with where I'd start.

— Add your name and a valid email so I can reply.
{{ errorMsg }}
or — hello@devexresearch.com

Sent.

Thanks{{ sentNameComma }} — I read every note myself, and I'll come back with where I'd look first.