Case: a retail chain's store sales fell 20% after it moved online
A 40-store apparel chain launched its own app last year. Online now takes a fifth of revenue, but total revenue is flat and store footfall is down a fifth. Store staff are on fixed salaries. The board wants a recommendation in one meeting.
An advanced-level session of about 10 minutes. You join 3 other speakers already in the discussion. You get 90 seconds to prepare before it starts.
The group will find the cannibalisation story quickly — online is eating the stores — and the danger is that everyone then spends eight minutes agreeing with it. Total revenue is flat while footfall is down a fifth, and flat revenue is the number that makes the simple story incomplete: something is holding the total up, and the group has not said what. Fixed-salary store staff turn this from a channel question into a cost-structure one, which is the part most rooms leave alone because it is uncomfortable. If the discussion is drifting towards close some stores, the useful contribution is to ask which stores, on what evidence, and what the app does for the ones that stay.
Angles worth taking
- Flat total revenue with falling footfall is the tension. Do not let the group skip past it.
- Fixed salaries make this a cost-structure question, which most rooms avoid because it is awkward.
- "Close stores" will arrive fast. Ask which ones and on what evidence before it becomes consensus.
- Stores as fulfilment and returns points is the option nobody proposes until too late.
- One meeting means the group owes a recommendation, not a list of things worth investigating.
Ready to speak?
Free to start. You need a TalkDrill account to speak — reading this page does not.
Browse all group discussion topicsThe format
Speakers in the room
You plus 3 other speakers
Preparation time
90 seconds
Modes
Learn (guided) · Mock (exam-real)
How a group discussion session works
Read the situation
You get the topic and the real disagreement behind it — who wants what, and why the obvious answer is not obvious. No hunting for context.
Take your prep time
A short prep clock, the same as a real round. Enough to pick a position and one example. Not enough to write a speech, which is the point.
Get into the discussion
The AI participants open with their own positions and keep going. You enter when you find the gap, name what you are adding, and hold your turn.
Read the report
How much of the discussion was yours, whether you built on others or restarted, your fluency and clarity, and the two things to fix next time.
What you're scored on
A GD panel is not scoring the topic. Across every discussion below, the same four things decide the result, and they are what the report measures.
Contribute real points, build on what others say, and hold your share of the discussion without dominating it.
Content
Whether your points are real — an example, a number, a consequence — or a restatement of the topic in longer words.
Entry and airtime
Whether you got in at all, how often, and whether you held your share without running the whole room.
Listening
Whether you answered what was actually said. Building on someone, or disagreeing with their point specifically, scores far above a prepared line.
Delivery
Pace, clarity and steadiness under interruption — the part that decides whether a good point survives being made.
More group discussion topics
All group discussion topics · Case topics · All speaking practice