Three phases. One loop.

Assessment → delivers the Opportunity.

The gap between where the stakeholder segments currently stand and what the client could credibly own.

Two-sided, always.

The Opportunity is delivered in two halves: Capture (upside to take) and Protect (downside to prevent). Neither half may be silently empty; if there is nothing to capture or nothing to protect, the Assessment says so. One noun, two halves — no separate "exposure" or "risk" deliverable, no eighth risk lens.

Protect is board-legible.

Each Protect-side risk is expressed in enterprise-risk language: the lenses generating it, likelihood, and magnitude of impact if unaddressed. This is how corporate affairs feeds the board's existing risk register rather than reporting beside it. Protecting licence to operate is an opportunity to define the ground before someone else does; framing defence as opportunity is more strategic than framing it as exposure.

A conclusion, not a scan.

A landscape deck that describes without concluding has not completed Assessment.

Segmentation happens here.

Stakeholder groups are carved in Assessment so they can be selected in Strategy.

Assessment is run through seven lenses. Each lens is an input; the Opportunity (Capture + Protect) is the single output. (They are lenses, not forces: inputs to a conclusion, not a competitive model.)

Four forcing questions, applied to every lens

1. 01

   What changed?

2. 02

   Who did it change things for? (Which stakeholder segments.)

3. 03

   Does it open or close a gap the client could own?

4. 04

   What is the probability and magnitude of harm if nothing changes?

Question 3 finds the gap (feeds Capture). Question 4 prices the downside (feeds Protect). Only lenses that answer 3 or 4 contribute to the Opportunity. Everything else is context: filed, not presented. This is the guard against Assessment degrading into monitoring. Trends are context inside the lenses, never the source of the Opportunity; the failure mode is strategy driven by what is new rather than by what stakeholders actually do.

Strategy → delivers the Idea.

Idea = Position × Target × Intended Outcome. The position the client takes, with the groups chosen, to produce a specific change.

• All three components must be present. An idea without a chosen target or a quantified intended outcome is a campaign theme, not a strategy.
• The Idea is the position, not the narrative. Messaging, creative platform and story are Execution outputs. Corporate affairs habitually merges these; the Standard keeps them apart.
• The intended outcome is quantified against the baselines set in Assessment. This is where measurement enters the process.

Research designed to serve the Build purpose and an external positioning purpose with one instrument. The industry calls it thought-leadership research or data-driven thought leadership; the classic form is a survey you publish on a topic to draw attention to the topic, with your name on the sponsor line.

POV. Done properly it is the best value in the grid: one field cost, two outputs. Internally it is Build. Externally it is Earned and Organic at once (POET), and it is the receipt under Provoke, Persuade and Position (P3). Most thought leadership in any category looks the same; a number nobody else has is the one thing that doesn't.

The one rule. The same instrument serves both purposes only if every question would survive publication. Design as if it will be published, even the parts that won't be. Not all research needs to be published; the split between the published module and the internal module is decided before fielding, labelled internally, and never revisited to keep only the flattering half.

Best practices, synthesised

AAPOR disclosure standards; Pew and Gallup practice on wording and reporting; established polling-firm practice on design; kept to core, simple concepts rather than a new methodology.

1. 01

   Topic first, brand second.

   The subject draws the attention; the sponsor borrows it. Published questions are about the category and the audience's life, not about you. Brand questions go in the internal module.

2. 02

   Neutral wording, balanced scales.

   Wording moves results more than sampling does: Pew found changing “welfare” to “assistance to the poor” shifted support by more than 20 points, and support for military action in Iraq fell from 68% to 43% when the question mentioned casualties.^{S16, S17} Scales carry equal positive and negative options; response options are mutually exclusive and exhaustive; no double-barrelled or leading questions; offer “don't know” where honest; randomise order because preceding questions change answers.^{S16, S18}

3. 03

   Design for surprise, not for the answer.

   Build the instrument in modules, each capable of carrying its own story and audience cut; include a few simple questions with a plausibly surprising answer; take a forward-looking angle; quantify a big idea rather than a small one. The Q2 discipline on top: write the headline hypotheses down before fielding (the “ideas to be pretested” of any good research report) so the surprise is found, not engineered.

4. 04

   Disclose like a pollster, the same day.

   AAPOR's immediate-disclosure items: who sponsored and who conducted (and who funded, if different); the exact wording and order of every reported question, including anything preceding it that could influence the answer; the population and geography; sampling frame; dates; mode; sample sizes by frame; margin of error and whether it's adjusted for weighting; weighting method.^{S20, S21} The method note ships with the headline, not on request.

5. 05

   Report faithfully.

   Stay close to the question wording when writing up; paraphrase puts words in respondents' mouths.^S22 Show subgroup n's; don't headline a subgroup below ~100.

6. 06

   Credibility is borrowed, so decide whose.

   This is the Partnered pillar of POET applied to research: shared trust, multiplied reach. A brand that publishes beside the right institution borrows trust from the room and becomes part of a bigger conversation than its own channel could create. Considerations rather than a rule: work with a research firm that has a strong track record in the domain, not just a name; partner with a relevant third-party organisation that already holds the audience's trust; partner with a known expert or academic who can stand behind the method. In every case the brand sits on the sponsor line and the partner sits on the method. In regulated or contested categories at least one of these is not optional.

7. 07

   Repeat it.

   An annual index beats a one-off: it buys t over time and delivers Prove for free.

8. 08

   Be generous with the data.

   Publish the tables and the instrument; make it shareable. That is the receipt.

9. 09

   Voice makes it coverable.

   Quant makes a finding citable; open-ends and interview verbatims make it a story. A published survey without voices is a press release.