Study Group
A one-hour role-play where employees forget they work at your company and try to use your product
- Difficulty
- Easy
- Time to result
- ~days to results
- Steps
- 6
- Confidence
- 95%
Study Group is a structured empathy ritual Jeff Weinstein created at Stripe to fight product entropy. Four to eight volunteers from any function meet on a call, invent a fictional company with a real goal, and attempt to accomplish that goal using the company's own products under two hard rules: you have never worked here, and you are not allowed to solve anything. The output is not a bug list but a felt, shared experience of how hard your product actually is to use — which turns out to be far more motivating to teams than customer complaints they can rationalise away.
Origin
Invented by Jeff Weinstein at Stripe (2024) as a lower-barrier successor to Stripe's existing 'friction log' practice, which required a single highly-motivated individual with broad context and organisational standing to walk a flow end-to-end and publish a critique. Weinstein wanted something more fun and far more frequent. Over 25 sessions and 250+ Stripe participants ran in the first few months; teams then 'franchised' it internally with their own study group captains. The pantomime lineage traces to his father, who ran an IT business in Baltimore and would physically place an empty chair in meetings to represent the customer.
Core principles
- 01Product decay is not a code problem — internal lingo and internal mental models leak into the application over time and nobody notices from the inside.
- 02You need an unnatural, safe, fun ritual to break internal knowledge, because internal knowledge is exactly what your customers don't have.
- 03Watching a colleague fail to use your product is more motivating than a customer failing, because you can't dismiss a colleague as 'a special setup'.
- 04Separate observing from fixing — the moment the room starts solving, the empathy work stops.
- 05Non-punitive by design: no team is named, no one introduces which team they work on, and there is no mandate attached to the output.
- 06Preference order for perspective: be your own customer > sit next to a real customer > pantomime the customer with zero internal knowledge.
How to run it
- 1
Open sign-ups to anyone, not just product people
Recruit four to eight participants from anywhere in the company — support, sales, events, infrastructure engineering, PM. Deliberately do not compose the group from the team that owns the product. Do not have people introduce which team they work on.
Pro tip Give it a slightly cutesy, vague name so you have room to evolve the format later — Weinstein picked 'study group' precisely because it didn't over-commit him to a shape.
- 2
Invent a fake company and an outcome goal
Together, make up a company (e.g. 'Dolphin Aquarium Industries') and a concrete outcome it wants: accept money in person at a farmers market, run a multi-country global business, sell tickets. Assign roles from within the group — someone is CEO, someone is the designer, someone is the CTO.
Pro tip Pick an outcome that is genuinely complicated and spans multiple products — the value is in the seams between teams, which is where entropy accumulates.
- 3
Enforce Rule One — you do not work here
Participants have never worked at the company. The facilitator polices this firmly but kindly: any time someone uses insider knowledge ('the docs link is wrong but the other one works'), stop, and redo the sentence or the action with zero internal knowledge.
Pro tip It takes a few corrections before the group settles in; after an hour people genuinely start to believe they work at the fake company.
Watch out Insider shortcuts are the single most common failure — people don't realise they're using them. Without a maestro actively catching them, the session degrades into a normal demo.
- 4
Enforce Rule Two — solve nothing
No critiquing, no solutioning, no suggesting, no filing bugs during the session. The session is recorded; all of that can happen afterwards. The hour is only for practising empathy and moving painstakingly slowly through the product as a customer would.
Pro tip Go deliberately slowly. The slowness is a feature — rushing is what produces the internal-optimised view you're trying to escape.
- 5
Funnel outputs into existing processes, not a new one
Route anything found into the bug/priority processes you already have. Stripe tags craft-related bugs against an existing rubric; a T0 craft bug (not an incident) must be acknowledged within seven days — acknowledged, not necessarily fixed. Teams are allowed to relabel severity themselves.
Pro tip Give teams great information and trust them with it. Attaching a mandate ('you must fix this') is what makes internal critique programmes get resented and die.
Watch out Resist the urge to add scoring rubrics, SLAs and regulation to the programme itself. That's the reasonable-sounding step that kills the fun and collapses the frequency.
- 6
Franchise it
Once it has an internal brand, teams will proactively request study groups on their own products (including before launch) and on internal tools. Appoint study-group captains so others can run their own, and maintain a backlog of requested targets.
Pro tip Anything can be study-grouped — internal sales tools, deal desk, vendor onboarding, a migration service. Nothing is off-limits from 'product'.
In the wild
A group of Stripes drop their real jobs and become the staff of an invented company that wants to sell in-person tickets to a dolphin aquarium. The person playing CEO turns to a colleague who isn't a designer in real life and asks where the colour palette is coming from. They then attempt the real Stripe flows to accept in-person payments, with the facilitator resetting anyone who uses insider knowledge.
→ Participants report reactions like 'I can't believe how hard it was to accomplish that — I thought we were amazing at this.' Over 25 sessions with 250+ participants ran in the first few months of 2024; teams began franchising it themselves.
Weinstein's father, who runs an IT services business in Baltimore, would physically drag a chair into a meeting, declare the customer was sitting in it, and address the invisible customer directly: based on what you've seen today at this meeting, are you more or less likely to pay your bill?
→ Weinstein describes it as an early prototype of study group — the same mechanism of forcing the customer's physical presence into an internal room.
Common mistakes
Letting the room start solving
The instant people begin filing bugs or proposing fixes, they slip back into employee mode and the empathy simulation collapses. The insights come from struggling, not from resolving the struggle.
Staffing it only with the owning team
The owning team cannot un-know its own product. The value comes from fresh eyes with no internal shortcuts — and from the owning team watching, not participating.
Bolting a mandate onto the output
Study group has no authority to demand fixes. Weinstein deliberately keeps it as 'high-signal cultural information' that teams choose to act on. Turning it into an enforcement mechanism makes teams defensive, which is the exact opposite of the intent.
Is it for you?
Best for
Product leaders at multi-team or multi-product companies where a once-smooth flow has silently rotted, and where individual friction logs are too high-effort to happen often
Not ideal for
Pre-product-market-fit teams where nobody uses the product yet — you have no entropy to find, and you should be sitting with real customers instead
From the transcript
“rule one is you do not work at stri and Rule two is we're not here to solve any problems this is just about practicing…”
“we're not here to solve any problems we're not here to critique we're not here to solution or suggest we're not here to file bugs”
“we need unnatural safe fun lore that gets us out of our chair and into the customer's mindset”
“it's you and you're your own customer okay second best would be sitting right next to the customer in the outside world and then like…”
“seeing your fellow teammates struggle to use your thing in some way is more motivating than the customer because you can kind of always say…”
From the episode
Building product at Stripe: craft, metrics, and customer obsession
Jeff Weinstein (Product lead)