How every engagement works
The method
Every analysis, research project, and design engagement at Irrational Technology runs on the same method: a systematic way of reading what a digital product is actually doing to its users, and a clear path from a team’s stated values to its UX. This page documents it. Read it before we ever speak — if the method doesn’t convince you, the call won’t either.
01 — The problem it solves
Why a method exists
Most digital products didn’t set out to manipulate their users. They inherited the patterns from the platforms that came before them — the infinite scrolls, the manipulative defaults, the engagement extraction dressed up as good UX, the friction in the wrong places and the absence of friction in the others. Anyone building a digital product today is building on top of two decades of design decisions that nobody at the new product would consciously sign off on.
The teams I work with have already decided they don’t want any of that. They have a stated philosophy. They have values about how users should be treated. What they don’t have is a systematic way of finding the patterns hiding in their own interface. The method is that systematic way — developed over two years of public research, published on Substack, presented at industry events, and applied in client work.
02 — The instrument
Reading the interface: the five lenses
An interface analysis reads every surface in scope against five lenses. Each lens has a diagnostic question, and findings that surface under more than one lens are a signal of significance, not a duplicate.
Lens 1
Defaults
What’s pre-selected, what’s checked on first load, what’s opted into without explicit consent. Notification permissions, marketing checkboxes, data-sharing toggles, auto-renewal.
Would the user knowingly choose this if it were presented as an opt-in?
Lens 2
Friction asymmetry
Where the product is easy and where it’s hard. Sign-up against sign-out, subscribe against unsubscribe, account creation against deletion. The shape of the friction tells you whose interests are being served.
Measure clicks-to-complete on opposing actions — anything past 2:1 is engineered.
Lens 3
Visibility
What’s prominent and what’s buried. Pricing before commitment, data permissions before grant, switching costs before lock-in.
At the moment of decision, does the user have what they need to choose?
Lens 4
Cost & power
What the user trades, and whether they can see it. What’s collected and why, what real control the user has against apparent control, how much of the experience they steer against how much is steered for them.
If the user knew the total cost of an action — in data, attention, money, social exposure — would they still take it?
Lens 5
Stated vs. actual
The lens that ties the others together: the gap between what a product says about itself and what its interface does. Mission statements and founder essays on one side, the findings from lenses one to four on the other.
Where exactly does the product stop doing what it says it does?
03 — Adjacent practices
What this isn’t
Interface analysis gets confused with adjacent practices. The short version:
Interface analysis
The behavioural architecture — patterns shaping user behaviour, often inherited from extractive platforms
Your product has a stated ethical position and you want the UX to actually reflect it
UX audit
Whether users can complete the tasks they came to do
You have usability problems and want them fixed
Ethics consultancy
Policies, principles, governance, risk
You need a framework or an external position, often for regulatory reasons
A UX audit asks whether users can do the thing the product wants. Interface analysis asks whether the thing the product wants is something the user would have asked for, knowing the full picture. They share methods; they’re chasing different problems.
04 — Beyond the audit
The same method, in research and design
The lenses aren’t only an audit instrument.
In research
The lenses become research questions
A finding under friction asymmetry becomes a usability session watching real users hit that friction; a stated-vs-actual gap becomes interview questions about what users believe the product does. Research engagements take whichever question matters and answer it with evidence — desk research and structured walk-throughs at the light end, user testing and qualitative interviews at the heavy end — so design decisions get made from a grounded account of how users actually move through the product, not from assumption.
In design
The lenses become constraints
Defaults the user would choose knowingly. Friction that’s symmetric — as easy to leave as to enter. Costs visible at the moment of decision. Power where the user can see it. Design work here doesn’t start from a blank canvas and a mood board; it starts from the same five questions the analysis asks, applied in reverse — and a retainer keeps applying them to every new feature as the product evolves.
Every engagement, whatever the mode, runs through the same three stages.
Stage 1
Discovery
A free call where we scope the work and fix the price.
Stage 2
Diagnosis
The deep read, visible to you as it develops.
Stage 3
Delivery
The scoped output, handed over in a working session so it lands inside your team.
Next step
The method is fixed. The engagement is scoped to you.
Four offers, each priced and timeboxed, from a £750 single-flow analysis to ongoing design work.
→