Structuring what you say

The 30-second version of anything

You should be able to produce on demand a 30-second version of: what your company does, what you do, any deal in your pipeline, any customer story, and yourself. Not a script, a shape.

For a company or product: who it's for → the problem they have → what you do about it → one piece of proof.

"We work with biobanks and research labs holding decades of samples across mixed systems. The recurring problem is that nobody can prove consent status fast enough to release samples, so access requests stall. We reconcile legacy records — including paper — into a single tracked chain of custody. Calderfield biobank cut access-request turnaround from three weeks to four days."

Four sentences, no adjectives; the proof is what stops it being marketing.

The pyramid principle: top line first

Most people narrate their reasoning: context, detail, more detail, conclusion. That works in a novel. In business it fails, because your listener is deciding whether to keep listening while you're still in the context.

Voice from the field

The Pyramid Principle's core instruction is that ideas in writing should always form a pyramid beneath a single governing thought: the answer first, then the arguments supporting it, then the data beneath those. Barbara Minto developed it over two decades of corporate writing at McKinsey, and it remains the house style of most strategy consultancies.
— Barbara Minto, The Pyramid Principle: Logic in Writing and Thinking (1987)

• Bad: "So we looked at the schema, and there were three naming conventions, and forty thousand records had no consent flag, so we think about six weeks."
• Good: "It's about six weeks. Three things drive that: three naming conventions to normalise, forty thousand records with no digital consent flag, and a reconciliation pass we'd want to run twice."

Answer the question that was asked, first: "how long does implementation take?" is answered by a duration, not by a description of your methodology.

The rule of three

Three is the working number for spoken points: enough to feel substantial, few enough to remember. There's experimental support for treating it as a ceiling — Shu and Carlson's study in the Journal of Marketing (2014), "When Three Charms but Four Alarms", found that when people know the source has a persuasive motive, impressions improved up to three claims and then declined, the fourth triggering scepticism about all of them.

Choose your three and drop the rest. Seven advantages and the buyer remembers zero; three and they might repeat them to a colleague — the actual goal, since most selling happens when you aren't there.

Telling a story that lands

Beginners ruin customer stories by telling them as a sequence of events. Stories need tension: setup → tension → resolution → point.

• Setup: who, and enough context to care. "Calderfield biobank, about 250,000 samples, three access requests a week."
• Tension (the part beginners skip): what was at stake. "Thirty thousand of their samples had consent recorded only on paper. Every request stalled while someone went to the filing cabinet, and two research groups had started sourcing elsewhere."
• Resolution: what changed, with a number. "We ran the reconciliation in phases over six weeks. Access turnaround went from three weeks to four days."
• Point: why you're telling this person this story now. "I mention it because you described the same paper-consent problem, and it's the part most vendors won't touch."

Rules: 60–90 seconds; real numbers, not "significant improvements"; never tell a story without the point.

Answering a question you don't know the answer to

The moment beginners fear most, and much less dangerous than it feels. You aren't expected to know everything; you're expected to be reliable.

1. Say you don't know, plainly. Not "great question", not a fifteen-second hedge, not a guess.
2. Say what you do know, if it's relevant.
3. Commit to a specific person and a specific date. "I'll ask our data lead and come back to you by Thursday." Vague commitments are worth nothing.
4. Then do it, by that date, even if the answer is bad. A seller who says "I checked — the answer is no, we'd need an export step" on the promised Thursday has built more trust than one who was right about everything and never followed up.

Never guess in front of a technical audience. In life sciences and health tech you'll routinely be the least expert person on the call, and a confident wrong answer about regulatory status or data handling destroys credibility in one sentence.

6. Reading the room

Buying signals versus politeness

Beginners over-read politeness, especially in the UK, where "that's very interesting" can mean anything from genuine interest to "please stop".

Real signals involve the buyer spending something — time, effort, political capital or specificity:

• They ask how it would work here: implementation, timelines, who'd need to be involved.
• They introduce someone else, or say "I'd want X in the room for this."
• They raise a concrete obstacle — budget cycle, procurement route, an internal sceptic. Objections are engagement.
• They use future tense about themselves, or correct your understanding of their process — both mean they're installing it mentally.

Politeness masquerading as signal: "This looks great", with no action attached. "Very interesting, send me some information" — the classic soft exit. "Leave it with me." Enthusiasm from someone with no budget and no stated problem.

The test: did they agree to do something specific, on a date, that costs them something? If not, you have interest, not intent. To find out: "What would need to be true for this to become a real project?"

Disengagement signals on video

Video strips out most body language, so what remains matters more: eyes tracking sideways and down (reading email); the camera going off mid-call, having been on, which is the strongest single signal; delayed, generic responses arriving a beat late; a stillness that isn't attention; and several attendees looking away at once, which means you've lost the room, not a person.

The response is always the same: stop talking and ask something. "I've been talking a while — is this the right level, or would X be more useful?" Presenting into a disengaged room never recovers it.

When to stop talking

Four hard triggers, any one of which ends your turn:

1. You've talked for more than about 90 seconds without the other person speaking.
2. You've answered the question. The common failure is answering, continuing to answer, then introducing a doubt that wasn't there.
3. You've made your point and they've agreed. Selling past the yes gives the buyer something new to disagree with.
4. You notice any disengagement signal.

Field note

The hardest version of "stop talking" is stopping after you've asked for something. You propose the next step, the silence is unbearable, so you soften it: "…or whenever suits, no pressure, obviously if it's not the right time…" You have just talked your way out of the yes. Make the ask, then close your mouth and let it be answered. If you build one habit from this module, build that one.

The silent participant, and multi-person dynamics

On a multi-person call someone always says nothing, and the silence means one of several very different things: seniority (senior people often let others run the meeting — their silence is evaluation), scepticism (waiting to say so afterwards, when you're not there), a technical remit (IT, information governance and data protection colleagues stay quiet until something touches their area), or no stake.

You can't tell which from silence, so bring them in — by name, on their territory, inviting a concern rather than an opinion:

"Priya, you're closest to the data side — what would worry you about a migration like this?"

That framing gives a sceptic permission to speak, and sceptics who speak on the call are far less dangerous than sceptics who speak after it. If someone stays silent through a whole call despite an invitation, flag it internally as a risk, not as agreement.

For the call as a whole: ask who's joining and why beforehand, so you can prepare a question for each; address people by name, because with six tiles on screen an unaddressed question goes to nobody; and watch who defers to whom, since the person others glance at is usually where power sits.