Client-voiced sessions: putting your clients on someone else's stage

Why this is its own format

The chapter 04 menu has a speaking slot (INTO speaks) and chapter 12 has owned dinners and workshops (INTO hosts). A programmed session at someone else's event where three of INTO's clients do the talking is neither, and the differences are the ones that matter operationally:

Chapter 12 gets closest — "a client speaker anchoring the theme is the highest-credibility draw" — but there it is one component of a room INTO controls. Here INTO controls neither the room nor the microphone, and the people carrying the argument do not work for INTO.

Add to the chapter 04 menu: client-voiced session · best for authority + qualified pipeline in a named market · all-in low (the slot is usually won editorially, not bought) · CPQC good if the room is right · INTO posture: pursue where the organiser's audience is an exact BU match and at least two clients are willing.

Doctrine

1. The client's voice is the product; INTO's job is the questions. If INTO's capability appears anywhere on that stage as an assertion, the format has failed and reverted to a vendor session.
2. A pattern persuades; three testimonials do not. Ask every guest the same questions in the same order. The audience discovers the shape in the third answer, and the shape is the operating model — surfaced without a framework slide.
3. The unflattering question is the credibility spine. Ask what they tried before the thing that worked. No client puts a false start in a testimonial, so an honest answer reclassifies the session from marketing to conversation in one exchange.
4. Never ask a client to sell. No lead-gen ask, no "tell them about the ROI", no working the stand. The session converts because the guests are visibly not selling.
5. A guest's relationship outranks the event. Any conflict between what would make a better session and what a client is comfortable with resolves to the client, immediately and without negotiation.

The decision: should this event get a client-voiced session?

Three conditions, all required:

• The room is an exact BU match. This format spends client goodwill; spend it where the audience converts. A generic room does not qualify no matter how large.
• At least two clients are willing and have a measurable change to describe. One client is a solo talk with a guest — re-scope it as one rather than run a thin panel. The kill-condition checkpoint belongs at T-6w, while re-scoping is still cheap.
• Client-naming approval is obtainable. This is Class C (SOP §8): Jay sign-off logged before any guest-facing artefact. Guest consent as panellists is the cleanest form of approval available, but it does not substitute for the recorded sign-off.

Roster composition

Compose for the room, not for the logo. An audience needs at least two companies it can plausibly be: one recognisable name buys attention, but three big names buy nothing, because the room concludes the method needs a big-company budget.

• Recognition seat (at most one): the name the room knows. Buys the first two minutes of attention.
• Relatability seats (at least two): companies the audience could be. Carries transferability, which is what the response objective actually rests on.
• Partner/adviser voice is a different chair again — valuable ("we recommend them to our clients") but it is not a like-for-like operator story and does not count toward the relatable two.

The question set is the deliverable

Not the slides — there usually are none. A deck behind a client conversation makes it look like a vendor session.

Four questions, same order, every guest:

1. The before, concretely. Force an operational detail, not a strategic abstraction. Naming an ordinary moment ("a typical Monday morning") does the work.
2. What they tried first — the spine. See doctrine 3.
3. What shipped, and how long it took. Dated and checkable. Where the delivery cadence becomes visible without being named.
4. What changed that they can measure. In the client's numbers, never ours.

Keep three rescue prompts in the moderator's pocket for abstract answers: give me one example / how long did that take before? / who on your team noticed first? The last one moves the answer from process to people, which is what an owner-operator room remembers.

Moderating is the position, not a demotion

The recurring internal objection is that a moderator cannot talk about the company. That is true, and it is the advantage.

The moderator who asks the sharpest questions in the room reads as the most authoritative person on the stage — expertise shows more in the question than the answer, because only someone who has run the work knows that the first attempt is where the learning is. So the exec moderates, does not pitch, and speaks as an operator exactly once: a 60–90 second closing synthesis naming the pattern across the stories. That is the entire commercial payload, and by then it is earned.

Write the close after the rehearsal, from what the guests actually said. Writing it first and bending their answers to fit is the failure mode.

What the guests receive, and what they are never asked

Send, at T-6w: the frame (what the session is and is not), the four questions verbatim, an explicit brief on the hard question, their one-sentence introduction for approval, an explicit and separate recording/reuse consent, and a written invitation to name anything off-limits — a list then honoured absolutely, including in clips and follow-up content.

Never asked: why they chose us; anything they marked confidential; competitor comparisons; to sell.

Offer, unprompted: their own footage, for their own channels. It costs nothing and is the single most appreciated thing in the pack.

Rehearse the clock, not the content

A programmed slot is a hard stop, and three guests plus Q&A in 40 minutes fails on timing long before it fails on substance.

• Lock the script at T-2w. No content changes after.
• Full timed rehearsal at T-1w, all guests. Non-negotiable; it is the largest execution risk in the format.
• The pre-event huddle is not a rehearsal. Ten minutes on nerves and running order; content is settled.
• Hold the clock hard on the day — move on mid-sentence at the arc limit if you must. Cut the cross-panel round first if the slot shortens; never cut the spine question.
• Plant two audience questions before doors. A dead Q&A reads as nobody here relates to this, and it is entirely preventable.

The 20 minutes after

A session that lands produces a queue with a half-life of about twenty minutes — the densest pool of qualified attention the day will produce. Staff it: nobody in a session, the moderator stays near the stage for senior conversations, someone works the queue and triages, someone captures vox-pops while people are still animated. Under-staffing this is the most expensive available mistake, and per chapter 03 it is also the difference between sitting at the top of the cost-per-conversation band and comfortably inside it.

Measuring the format specifically

Beyond the standard P1–P5, this format has three questions of its own for the readout (chapter 11):

1. Did the repeated-question structure produce a recognisable pattern for the audience — what did the queue and the vox-pops actually say?
2. Did the spine question get honest answers on stage? If not, was it the briefing or the guest?
3. Was moderating enough for the exec's positioning, or did the room finish unsure who INTO was? This is the bet the format makes; this is where it gets settled.

Add the "which story landed" column to the capture sheet — three seconds per row, and it is the cheapest possible read on which arc did the persuading, which sets next year's roster and arc order.

Standing caution

This chapter is argued, not measured. Every stance above is INTO operating doctrine derived from reasoning about one event that has not yet run. It carries no external benchmark because none was found for the format. Treat it as a hypothesis with a scheduled test: the first instance's T+90 readout either supports it or amends it, and the amendments come back here via the post-event report's pack-improvements field — never by assertion.