Coffee subscription concept test

It runs the room with you and writes the readout while you talk.

To understand how frequent and occasional coffee buyers react to a $12.99 a month unlimited drinks subscription, what they think it is worth, what makes them hesitate, and what would have to change for them to sign up.

guide Warm-up · Current habit · Stimulus · Key questions · Wrap length 34 minutes

A counter in a coffee shop with cups being prepared — The offer shown to the group at 03:30

notes taken 0

M Marcus daily, commutes 0% P Priya twice a week 0% A Ade daily, works nearby 0% D Dana weekends only 0% S Steph three times a week 0% R Ron daily, retired 0%

Readout

What we heard

Coffee subscription concept test · Group 2 of 4 · Tuesday, 7:00pm · 6 participants · 34 minutes

Hands, notebooks and coffee cups on a long table during a group discussion — The hour you already run. Everything below was written while it happened.

Heavy buyers price it in under a minute and stop evaluating

The daily buyers converted the offer into a per-cup number within forty seconds of seeing it and did not revisit it. For this group the concept does not need selling, it needs to not be spoiled by the terms.

3 of 6

That is what, fifty pence a coffee. It is not even a close call.Marcus · P1 · daily, commutes · 04:09

The resistance is about being locked in, not about the price

Nobody in the room argued that $12.99 was too much for what it gives. The objections were subscription fatigue, an expectation that the price will rise once the card is on file, and a general disbelief in cancel any time.

4 of 6

My first thought was not the coffee, it was, oh, another one.Ron · P6 · daily, retired · 06:11

For light buyers the offer creates a members-only price they are on the wrong side of

The twice-a-week buyer did not read the offer as expensive. She read it as not meant for her, and then reported a small ongoing irritation at paying full price next to someone who is not. This may suggest a smaller tier is worth testing, though one participant is not enough to size it.

2 of 6

Like I am the mug for not being organised enough.Priya · P2 · twice a week · 11:52

What we think it means

Kept separate from what we heard, on purpose. The block above is the room. This block is a reading of it, and you should be able to disagree with it without disagreeing with the transcript.

The pricing question looks close to settled for heavy users, so the risk sits in the terms rather than the number.
A pause control and a written price hold both came up unprompted and both address lock-in rather than cost. They are cheap to offer and they answer the objection that was actually raised.
If a smaller tier is added it changes who the product is for, so it belongs in a separate test rather than as a footnote to this one.

What to do before the next group

Put a pause control in the next stimulus and see whether the lock-in objection survives it.
Test a fixed-price-for-twelve-months line, since it was named unprompted by the most sceptical participant.
Answer the two operational questions before the next group: can it be used to buy for someone else, and what happens to existing loyalty stamps.
Size the per-cup break-even in quant. Six people cannot tell you where the line sits.

Six participants, group two of four in this study. Qualitative and not projectable. Quotes are verbatim from this session and are attributed by code and behaviour rather than by real name, which is the normal convention in a client-facing readout.

Three things your brief does not say, and what I would pick

Four people talking around a table in a cafe — Stock photograph. Six people, one table, one hour is the format this is built for.

Does the group meet live, or answer over a few days?

Live gives you the pile-on and the group energy. Async means everyone answers in their own time, gets probed privately, and the quiet ones say far more. Async is also cheaper to build and has no scheduling problem.

I would build async first and add live after.

Does it run the room, or sit next to you while you run it?

You wrote help me run, so the version above sits beside you. It watches airtime, drafts the next question, and writes the notes. You still decide what gets asked. Full auto scales better but it will miss the follow-up that mattered.

Co-pilot for anything client-facing. Full auto is fine for screening rounds.

Text, voice, or listening in on the call you already run?

Voice feels like the real thing but you take on interruptions, mic quality and working out who is speaking. There is a third option: keep running the group on Zoom exactly as you do now, and this listens on a second screen.

For this budget, listening in on your existing call gives you the most back for the least risk.