---
name: research-presentation-development
description: >
  Builds a research presentation as an argument rather than a set of sections:
  one idea per slide, headlines that state the finding, a sequence that follows
  the narrative, an opening that establishes the decision at stake, anticipated
  questions placed deliberately, and ruthless cutting. Use for "build the
  deck", "turn the report into slides", "the debrief presentation", "readout
  deck", "present the findings", "this deck is too long", "slides for the
  steering group".
category: 11 Reporting and Storytelling
ref: 11.05
tier: 1
inherits: [K2, K3, K4, K5]
---

# Research Presentation Development

## 1. One-line description
Turns a research argument into a presentation where the slide sequence is the argument, each slide carries one idea under a headline that states the finding, and the deck is built for how it will actually be used, whether delivered live or read alone.

## 2. What this skill is used for

**The research problem it solves.** Most research presentations are reports with page breaks. They are structured by section rather than by argument, headed by labels rather than findings, dense enough to be unreadable at three metres and thin enough to be uninterpretable when forwarded, and roughly twice as long as the session that will hold them. The audience leaves having seen the study and not having received it.

The root confusion is specific and worth naming first, because almost every other failure descends from it. **A presentation delivered live and a presentation read alone are two different artefacts.** The live deck is a support for a person talking: sparse, visual, sequenced for spoken delivery, with the argument carried by the speaker and the slides carrying the evidence. The read-alone deck is a document in landscape format: it must contain everything the speaker would have said, because the speaker is not there. The commonest failure in research delivery is building one and using it as the other. A sparse live deck forwarded to a stakeholder is a set of orphan charts. A dense read-alone deck projected in a room is a wall of text nobody reads while someone reads it to them.

Content and design are treated here as one pass, because the slide structure is the argument. Deciding that a point takes three slides rather than one is a design decision and an analytical one simultaneously, and separating them produces decks where the layout fights the reasoning.

**Where it sits.** Reporting. Downstream of narrative and analysis, and frequently in parallel with the written report rather than after it. Upstream of the session itself and of the debrief workshop.

**Typical use cases.**
- Building the debrief deck for a completed study.
- Converting a written report into a presentation without simply pasting it in.
- Preparing a steering group or board readout where a decision will be taken in the room.
- Cutting a deck that is twice the length of the session booked for it.
- Building a deck that will be presented once and then forwarded widely.
- Preparing for a session where the findings will be challenged.

**Who uses it.** Researchers and insight leads who present their own work; research directors preparing a client debrief; consultants whose deck is the deliverable; anyone who has watched a good study land badly in a room.

## 3. When to use it

- A completed study has to be presented, and the deck does not exist yet.
- A written report exists and a presentation is also owed, and the temptation to export the report is live.
- The session will end in a decision, so the deck has to argue rather than describe.
- The deck will be forwarded to people who were not in the room, which is nearly always.
- A previous debrief lost the room, and the reason was structural rather than analytical.
- Time is short and the deck must be cut on principle rather than by deleting the last section.
- The audience is expected to challenge the findings, and the answers need placing rather than improvising.

## 4. When NOT to use it

- **The narrative is not settled.** A deck built before the argument is decided becomes the place the argument gets decided, which is the most expensive available venue: slide order substitutes for reasoning, and the version that survives is the one that was easiest to lay out. Settle it in **11.01 Research Narrative Development** first.
- **The analysis is not finished.** Slides create slots faster than any other format, and a slot demands content. Do not build toward a cut that has not been run.
- **The right deliverable is a document.** Where the material is reference, where the audience will consult rather than attend, where the reasoning needs full sentences to be defensible, or where the output must be citable, write the report. Landscape format does not make a document a presentation, and slides are a poor container for a nuanced methodological argument. That is **11.03 Research Report Writing** and **12.01 Research Report Architecture**.
- **The deliverable is a short executive artefact with no session behind it.** A board note or a five-slide decision paper standing alone is **12.05 Executive Research Reporting**.
- **The question is document-level visual design.** Template construction, typographic system, palette definition, grid and the house visual language are **12.02 Research Report Design**. **The boundary, stated in both directions: this skill owns slide structure, hierarchy, density and sequencing as part of the argument, and hands the visual system to 12.02.** Where a design team owns the template, this skill works inside it.
- **The question is which chart to use.** Encoding, ordering, scale and base disclosure are **11.04 Data Visualisation and Chart Selection**. A slide with a misleading chart is not fixed by better layout.
- **A workshop is what is actually needed.** Where the purpose is to get stakeholders to work with the findings rather than receive them, a presentation is the wrong instrument and will produce polite agreement and no action. That is **14.02 Research Debrief and Workshop Design**, and choosing correctly between the two is worth more than any amount of slide craft.
- **The deck is being built to justify a decision already taken.** Where the session exists to ratify a conclusion and the research is the ratification, say so per K4 §4.2. A deck is a persuasive instrument and this is the format where a predetermined conclusion is easiest to dress and hardest to detect.

## 5. Required inputs

**Required.** Without these the skill cannot run. If absent, ask. If no answer is available and work must proceed, state the assumption at the point where it bites, per K5 §5.

- **The narrative statement or spine**, from **11.01**, with its sequence. The slide order is the argument, so it comes from the argument.
- **The findings with references, bases and confidence levels.** A slide is an extractable artefact, so every claim on it needs its evidence attached at the point of use.
- **The decision at stake, its owner, and whether it will be taken in this session.** Determines the opening, the sequencing and how much time is reserved for discussion.
- **The session parameters**: length, audience, seniority, room or remote, and who else will see the deck afterwards. A 30-minute session with a decision at the end and a 90-minute working session are different decks.
- **Whether the deck must also work as a read-alone document**, and for whom. This is the single most consequential input and it is almost never volunteered, so ask.

**Optional, and what each one adds.**

- **The written report**: lets the deck be genuinely sparse, because the detail has somewhere to live and can be pointed at rather than reproduced.
- **The chart set** from **11.04**: removes the largest single time cost and prevents charts being rebuilt for the deck in a way that quietly changes their claims.
- **The client's template**: fixes the visual system so this skill works only on structure and hierarchy.
- **Knowledge of the stakeholders' positions**: identifies which findings will be resisted, which lets the question anticipation at step 6 be specific rather than generic.
- **A record of what has already been said in interim sessions**: prevents the deck presenting as new what the audience heard six weeks ago, which is the fastest way to lose a room.
- **The previous wave's deck**: fixes conventions and lets the session lead on change rather than re-establishing the basics.

## 6. Questions to ask before starting

1. **Will this be delivered live, read alone, or both?** Determines the entire artefact. *Default if unanswered:* assume both, and use the two-artefact resolution in step 1 rather than compromising into a deck that fails at each.
2. **How long is the session, and how much of it is discussion?** Determines slide count, which determines what is cut. *Default:* reserve a third of a decision session for discussion, and build for the remainder.
3. **Is a decision being taken in the room, and by whom?** Determines the opening and where the recommendations sit. *Default:* assume the decision is live and open on it.
4. **What does this audience already know, and what have they already been told by us?** Determines where the deck starts. *Default:* assume everything presented in an interim session is known, and lead with what is new.
5. **Which findings will be resisted, and by whom?** Determines question anticipation and where evidence needs to be immediately reachable. *Default:* assume the finding with the largest cost implication will be challenged on its method.
6. **Who else will see this deck, in what context, without the presenter?** Determines how self-sufficient each slide must be, particularly the headlines. *Default:* assume every slide will be forwarded alone.
7. **Is there a template, and does it constrain anything this skill requires?** Determines what can be negotiated. *Default:* work within the template on everything cosmetic, and never on base disclosure or confidence marking.

## 7. Step-by-step methodology

**Step 1. Decide which artefact you are building, and resolve the two-artefact problem explicitly.** Three legitimate resolutions, and the failure is choosing none of them. **Two decks**: a sparse live deck and a fuller leave-behind, which is the best outcome and costs the most. **One deck with substantive speaker notes**, where the notes carry what the presenter would say and the forwarded version includes them, which is cheap and works only if the notes are actually written. **One deliberately designed read-alone deck**, denser than a live deck, presented by talking around it rather than reading it, which is honest and requires the presenter to accept that people will read ahead. Choose one, say which, and build to it. **Do not build a deck that is 60% dense: it will be too heavy to present and too thin to forward, which is where most research decks land.** *Correct result:* an explicit decision recorded before the first slide, and a slide count budget that follows from it.

**Step 2. Open by establishing the decision at stake.** The first substantive slide is not the objectives, not the method, and not an agenda. It states the decision the audience faces and what the research was asked to settle, in their words. This does three things: it tells the room what the next thirty minutes are for, it establishes that the study is about their problem rather than about research, and it sets the standard the deck will be judged against, which you want to be a standard you meet. The second slide is the answer, at its confidence level. **Do not withhold the conclusion for the end.** An audience that does not know where the argument is going evaluates each slide independently and challenges early, which is how sessions get lost in the first ten minutes on a methodological point that the fourth slide would have made irrelevant. Method goes late, in an appendix or a single slide, unless the audience is technical or the method is itself a finding. *Correct result:* by slide three, the room knows the decision, the answer and how firm it is.

**Step 3. One primary idea per slide, without exception.** A slide carrying two ideas has one that is being ignored, and neither the presenter nor the audience knows which. The test is mechanical: write the slide's idea as a sentence. If the sentence contains "and" joining two claims, it is two slides. This is the discipline that most improves a deck and the one most resisted, because splitting feels like inflating the count. It is not: splitting one crowded slide into two clear ones typically saves time in the room, because the presenter stops navigating their own slide. Where a slide genuinely needs supporting detail, the detail supports the one idea rather than adding a second. *Correct result:* every slide's idea statable in one clause.

**Step 4. Write headlines that state the finding.** The headline is the top line of the slide and it is the most-read text in any deck, because it is what people photograph, forward and remember. It states the finding, not the topic. Worked pairs:

| Label headline | Finding headline |
|---|---|
| Shopper Frequency | Weekly shoppers have become fortnightly shoppers, and basket size has not compensated |
| Barriers to Purchase | Two barriers account for most of the loss, and both are in-store rather than in the range |
| Segment Overview | The segment the range was designed for is the only one that has not grown |
| Awareness and Consideration | Awareness is not the constraint: nine in ten know the brand and one in five considers it |
| Recommendations | Fixing availability before adding lines is the higher-return sequence |

Three tests for every headline. Is it a claim rather than a topic, with a verb? Is it true of this slide and nothing more? Would it survive being read alone, with no chart attached, in someone else's email? That third test is the one that catches overreach, because a headline that quietly added causation or a trend passes the first two. Watch specifically for causal verbs the design does not license, per K4 §3.2, and for behaviour claimed from stated preference. A headline that names the finding also disciplines the slide: if the chart does not support the headline, one of them is wrong, and it is usually easier to see on a slide than in a report. *Correct result:* a deck whose headlines, read in sequence with nothing else, give the argument.

**Step 5. Sequence the slides as an argument, not as sections.** Take the sequence from **11.01** and lay it out slide by slide, checking the same local test the narrative used: at the end of each slide, does the audience hold a question that the next slide answers? A deck structured as sections ("here is the quantitative, here is the qualitative, here are the segments") makes the audience assemble the argument, which they will do incompletely and at different speeds. A deck structured as an argument carries them. Two devices help. Use a **transition slide that states the question**, not the section name, so the deck's own structure is a chain of questions. And **build cumulatively**, so a slide that adds a dimension to a finding shown earlier reuses the earlier visual rather than introducing a new one, which lets the audience track a developing claim rather than parsing a new chart each time. *Correct result:* headlines and transitions that read as a continuous argument.

**Step 6. Anticipate the questions and place the answers deliberately.** List the questions this audience will ask, specifically rather than generically: the methodological challenge to the finding with the largest cost implication, the "did you look at X" from the stakeholder whose area is not covered, the "that contradicts what we see in our data" from the operations lead, the "is that significant" from the analyst in the room. For each, decide where the answer lives. **In the flow** if the question is central and asking it should not derail the session. **On the slide, in a small annotation**, for base sizes, test status and exclusions, which is where most methodological challenges are pre-empted for free. **In the appendix**, findable in seconds, for anything a subset of the room will ask. **In the speaker's head**, for what is genuinely peripheral. The judgement is about the whole room: a defensive deck that pre-answers everything is unreadable and signals that the researcher expects to be attacked. *Correct result:* a question list with a placement against each, and an appendix built from it rather than from leftovers.

**Step 7. Build the appendix as a deliberate artefact.** The appendix holds what a specific question would need, not what did not fit. It is ordered and navigable, because an appendix nobody can search in a live session does not exist. It contains: full data tables, the questionnaire or guide as fielded, method and sample detail, findings analysed but not central, wave comparison detail, and the evidence behind any contested claim. **It does not contain a finding that belongs in the main flow**, and the tell of that error is a presenter who says "there is more on this in the appendix" about something material, which per K4 §4.3 usually means a limitation or an inconvenient finding has been relocated rather than reported. *Correct result:* an appendix with a contents slide, ordered by anticipated question.

**Step 8. Set the visual hierarchy and the density.** At the level a researcher can apply without a designer, four decisions carry most of the quality. **Hierarchy**: exactly one dominant element per slide, which is the thing carrying the idea, with everything else visibly subordinate. Three competing elements at the same visual weight is the commonest slide defect and it makes the audience choose where to look. **Whitespace**: the strongest available emphasis device and free. A slide that feels empty in the file reads as confident in the room; density reads as anxiety. **Typography**: two sizes and two weights are enough, applied consistently, with the smallest text on the slide readable at the back of the room (in practice, larger than most templates default to). **Density**: a live slide carries roughly one chart and one headline, and a read-alone slide can carry a chart, a headline and a short paragraph, and neither carries a paragraph of body text plus three charts. Every slide carries its base and question reference in a consistent position, per K4 §4.3, because slides are the most-forwarded research artefact in existence. Where a template governs the visual system, apply it and hand any conflict to **12.02**. *Correct result:* a deck where each slide's most important element is unambiguous at a glance.

**Step 9. Time it, then cut.** Estimate honestly: a substantive slide with a chart and a point to make takes two to three minutes to present properly, and a complex or contested one takes longer. Reserve a third of a decision session for discussion, because discussion is where a decision session does its work and a deck that consumes the whole slot has prevented the outcome it was built for. Then cut to fit, and cut in the right order: **whole slides before content, supporting findings before the argument, detail before evidence, and never the base sizes, the confidence markers or the contradicting evidence.** Almost every research presentation is too long, and the cause is almost always that the deck reports the project rather than making the argument. The material that comes out goes to the appendix or the report, which is why both should exist. **A useful discipline: cut to two thirds of the available time before the first rehearsal**, because rehearsal always adds. *Correct result:* a deck that fits the session with discussion time intact.

**Step 10. Rehearse against the argument and the question list.** Two passes. First, present the headlines alone, in order, out loud. If the argument does not hold from the headlines, it will not hold in the room, and the fix is structural rather than verbal. Second, run the question list with a colleague playing the audience, checking that each answer is where step 6 said it was and reachable in seconds. Then a final check on the artefact decision from step 1: forward the deck to a colleague who was not in the room and ask what they think it says. If the answer is not the spine, the deck is a live artefact being used as a read-alone one, and the notes or the leave-behind need writing. *Correct result:* a deck whose argument survives the headlines alone and whose answers are where they were planned to be.

## 8. Analytical framework

The deck as argument:

    Decision at stake
      → The answer, at its confidence
        → Slide 1..n  [each: one idea, headline as finding, evidence, base]
            with transitions phrased as the question the next slide answers
              → What we could not establish
                → Recommendations, with confidence and sign-off
                  → Appendix, ordered by anticipated question

**Applying it.** Build the headline sequence first, as text, before any slide is laid out. It takes twenty minutes, it exposes structural problems that are expensive to fix once slides exist, and it is the only artefact in deck-building that can be revised cheaply.

**The two artefacts, compared.** Deciding between them is step 1 and everything downstream depends on it.

| | Live deck | Read-alone deck |
|---|---|---|
| Argument carried by | The presenter | The slides |
| Density | One idea, one visual, minimal text | Headline, visual and a short paragraph of reasoning |
| Headlines | State the finding, and the presenter elaborates | State the finding, and it must be sufficient |
| Caveats | Spoken and marked on the slide | Written on the slide, in full |
| Failure if misused | Forwarded as orphan charts nobody can interpret | Projected as text nobody reads while it is read to them |

**The three-metre and the forwarded test.** Two checks that catch most slide defects. Can the slide be read from the back of the room? Would a person who receives this slide alone, with no presenter, understand what it claims and on what basis? The second is the more demanding, and passing it is mostly a matter of the headline and the base note.

## 9. Output format

**A. Opening (two to three slides)**
The decision at stake, in the audience's words. The answer, at its confidence level. What the session will cover, only if the audience needs orientation.

**B. The argument (the bulk)**
One idea per slide. Headline states the finding. One dominant visual or evidence element. Base, question reference and test status in a consistent position, per K4 §4.3. Interpretation marked as interpretation, per K2 §3.2, including on slides where space is tight. Transitions phrased as the question the next slide answers.

**C. What we could not establish (one slide)**
Objectives not answered, questions the evidence cannot settle, contradictions unresolved. Per K3 §5.2. It belongs in the flow, not the appendix, and it does more for a researcher's credibility in a room than any other slide.

**D. Recommendations (one to three slides)**
Each states the action, the findings it rests on by reference, its confidence per K3 §5.2, and what would change it. `RESEARCHER SIGN-OFF REQUIRED` per K5 §3.1 before the deck leaves.

**E. Appendix**
Contents slide, then material ordered by anticipated question: full tables, method and sample, the instrument as fielded, wave comparison detail, findings analysed but not central.

**Speaker notes**, where the one-deck-with-notes resolution was chosen at step 1: substantive, carrying what the presenter would say, not a restatement of the slide.

**When the evidence is thin.** The deck gets shorter, and the "what we could not establish" slide gets longer and moves earlier, because an audience that learns late that a question was not answered stops trusting what came before. A finding on a small base is presented with its base visible and a directional reading, and never as a headline slide, per K4 §7. Where a section of the argument has no evidence, it is not a slide with a plausible-looking framework on it: frameworks are the visual form of gap-filling and they are almost impossible to challenge in a room, which is exactly why they must not be used, per K4 §1. A twelve-slide deck that says what the study established is a better session than a forty-slide deck that fills the hour.

## 10. Quality checks

Run before the deck is presented or sent. Sits on top of K4 §8.

1. Was an explicit decision made about whether this is a live deck, a read-alone deck, or both, and does the density match it?
2. Does the deck open on the decision at stake rather than on objectives, method or an agenda?
3. Is the answer stated in the first three slides rather than withheld?
4. Does every slide carry exactly one primary idea?
5. Is every headline a claim with a verb, true of its slide and nothing more, and safe if forwarded alone?
6. Does any headline claim causation, behaviour from stated preference, change over time, or reach beyond the sample?
7. Read as a sequence with nothing else, do the headlines give the argument?
8. Does every slide showing a proportion carry its base description, base size and question reference?
9. Are untested differences marked as observed rather than implied to be real, per K4 §3.1?
10. Is interpretation marked as interpretation on the slides, per K2 §3.2, and not flattened into finding language by compression?
11. Is there a "what we could not establish" slide in the main flow rather than the appendix?
12. Does every recommendation name its findings, carry a confidence marker and a sign-off flag?
13. Is any material finding or limitation sitting in the appendix rather than the flow?
14. Does the deck fit the session with a third of the time reserved for discussion?
15. Has the question list been built, and is every answer where it was planned to be, reachable in seconds?
16. Would a colleague who receives this deck with no presenter identify the spine?

## 11. Common failure modes

| Failure | How to recognise it | How to prevent it |
|---|---|---|
| **The report with page breaks** | Paragraphs on slides, section-shaped structure, no headlines | Build the headline sequence first, as an argument. Steps 4 and 5 |
| **The undecided artefact** | Too dense to present, too thin to forward | Step 1. Choose two decks, one deck with notes, or a designed read-alone deck |
| **The withheld conclusion** | The answer arrives on slide 30; the room challenges from slide 4 | Answer in the first three slides. Step 2 |
| **The label headline** | Noun phrases at the top of every slide; the reader must read the chart to learn anything | Every headline is a claim with a verb, safe read alone |
| **Two ideas per slide** | The presenter says "there are two things here" | Write the slide's idea as a sentence. If it contains a joining "and", split it |
| **Section structure** | "Quantitative findings", "Qualitative findings", "Segments" as the top-level shape | Sequence by argument. Transitions state the question the next slide answers |
| **The overlong deck** | Forty slides for a forty-minute session with a decision at the end | Two to three minutes per substantive slide, a third of the session for discussion. Step 9 |
| **Cutting the wrong things** | Bases, caveats and the contradicting finding removed to save time | Cut whole slides, supporting findings and detail. Never evidence or qualification |
| **The framework slide** (the signature AI failure) | An attractive two-by-two or model where the evidence is thin, unchallengeable in a room | K4 §1. Where the evidence supports no slide, there is no slide |
| **Appendix as a hiding place** | "There is more on that in the appendix" about something material | The appendix answers anticipated questions. It never holds a finding or a limitation from the flow |
| **The unreadable slide** | Body text below the size the back row can read; three elements at equal weight | One dominant element, two type sizes, generous whitespace. Step 8 |
| **The orphan slide** | A forwarded slide nobody can interpret: no headline claim, no base, no context | Every slide self-sufficient in headline and base note. The forwarded test |
| **Improvised answers** | A challenge derails the session while the presenter searches the deck | Build the question list and place every answer. Step 6 |
| **Confidence flattened by compression** | Slide language stronger than the report it came from, because slides have less room for hedge | Mark confidence on the slide, per K3 §5.2. Compression is not permission |

## 12. AI guardrails

Skill-specific only. Universal prohibitions are inherited from K4. Confidence marking follows K3 §5.2; interpretation marking follows K2 §3.2.

1. **Never write a slide headline that claims more than the slide's evidence supports.** Headlines travel further than any other research text and they travel without their charts.
2. **Never omit a base size, a base description or a test status from a slide because space is tight.** Space constraints are not a disclosure exemption, per K4 §4.3. Cut the slide or cut other content.
3. **Never let compression upgrade confidence.** A finding stated at moderate confidence in the report is stated at moderate confidence on the slide, in fewer words, not in stronger ones.
4. **Never build a framework, model or two-by-two that the evidence does not populate.** A conceptual visual is nearly unchallengeable in a live room, which makes it the most dangerous available form of gap-filling, per K4 §1.
5. **Never move a material finding, limitation or contradicting result to the appendix.** The appendix answers anticipated questions; it does not hold anything the audience needs in order to read the argument correctly, per K4 §4.1 and §4.3.
6. **Never present a slide as new when the audience has already been shown it in an interim session**, and never present an interim finding that later analysis has revised without saying it has been revised.
7. **Never write a recommendation slide without the findings beneath it named by reference**, per K2 §5, and without a sign-off marker per K5 §3.1.
8. **Never let a chart be rebuilt for the deck in a way that changes its encoding, ordering, scale or base.** Charts come from **11.04** and are re-styled, not re-argued.
9. **Never produce slides to fill a session length.** A deck is as long as its argument, and the residual time belongs to discussion.
10. **Never present a quote on a slide without its participant reference and, where a theme is claimed, its prevalence.** Quotes on slides are the most-extracted evidence in research and the easiest to strip of attribution.

## 13. Best-practice principles

1. **The slide sequence is the argument.** Structure decisions are analytical decisions, which is why content and design are one pass and why the headline sequence is written before any slide is laid out.
2. **Decide whether it will be presented or read, and build that thing.** The undecided deck fails at both, and it is the most common deck in existence.
3. **Lead with the answer.** Withholding the conclusion for effect invites the room to challenge every step of a journey whose destination they cannot see.
4. **Every headline is a claim that survives being forwarded alone.** Assume every slide will be photographed and sent to someone who was not there, because it will be.
5. **One idea per slide, and the second idea gets its own slide.** Splitting costs a slide and saves time in the room.
6. **Whitespace is the cheapest emphasis available.** Density reads as anxiety; space reads as confidence, and the confidence is easier to justify when there is less on the slide to defend.
7. **Anticipate the challenge and place the answer.** Improvised answers in a room cost more credibility than any weakness in the finding itself.
8. **The "what we could not establish" slide belongs in the flow.** It is the slide that most improves how the rest of the deck is heard.
9. **Cut whole slides, never qualifications.** The base size, the confidence marker and the contradicting finding are not compressible, and they are the first things a rushed edit removes.
10. **Reserve a third of a decision session for discussion.** A deck that fills its slot has prevented the decision it was built to inform.
11. **Rehearse from the headlines.** If the argument does not hold from the headlines alone, it will not hold in the room, and no amount of delivery fixes a structural problem.
12. **Almost every research presentation is too long, and the cause is reporting the project rather than making the argument.** The project is in the report. The room needs the argument.

## 14. Worked example

*Fictional scenario, used for illustration only. The organisation, findings, figures and quotes below are invented for the purpose of demonstrating method.*

**INPUT.** A grocery retailer has completed a study into declining performance of its own-label chilled range. The report is written and QA'd. The debrief is a 45-minute steering group session at which the category director will decide whether to invest in range extension or in in-store availability. Inputs: the narrative statement (spine: the range is not the problem, availability is, and the two are being confused because the availability failure is concentrated in the newest lines), 24 findings, a chart set from 11.04, and a draft deck of 52 slides built by exporting the report.

**PROCESS.**

*Step 1, and the first decision.* The deck will be presented once and then circulated to three regional teams who were not in the room. The draft is 60% dense: too heavy to present, too thin to interpret alone. The resolution chosen is one live deck plus substantive speaker notes, because a second leave-behind cannot be produced in the time available. The notes are budgeted as real work rather than an afterthought.

*Step 2, and the restructure.* The draft opens with four slides of objectives and method. They move to the appendix, except one method line that stays on the closing slide. The new opening states the decision: "You are choosing between range investment and availability investment for the next financial year." Slide two states the answer at its confidence: "The evidence points to availability, at moderate confidence, and the two options are not equally reversible."

*Step 4, and the headlines.* All 52 headlines are labels. Rewritten, five carry the argument: "Shoppers are not rejecting the range, they are not finding it"; "Out-of-stock rates on the six newest lines are three times the range average (client-supplied stock data, weeks 12 to 30)"; "The shoppers most affected are the ones the range was built for"; "Adding lines increases the number of lines that can be out of stock"; "Fixing availability before adding lines is the higher-return sequence". The client-supplied label travels onto the slide, per K2 §6, because the strongest evidence for the second headline is not evidence the team produced.

*Step 6, and the judgement call.* The question list identifies that the category director's deputy has publicly backed range extension and will challenge the availability finding on method: the stock data is operational, not research, and covers 18 weeks. The temptation is to lead with the shopper evidence and hold the stock data back. That would be a selection, per K4 §4.2. Instead the limitation goes on the slide itself, in a one-line annotation, and a "what would change this" line is added to the recommendation. The challenge is pre-empted rather than avoided, and the answer sits in the appendix as three slides on the stock data's coverage and its known gaps.

*Step 9, and the cut.* 52 slides becomes 14 in the flow plus 21 in the appendix. The cuts are made by whole slide: nine slides of segment profiling that support no part of the argument, six wave comparisons where question wording changed and the comparison is not valid, and the entire methodology section. Nothing is cut from a surviving slide. Two findings on bases under 100 keep their flags and their directional framing. At 14 slides and roughly two and a half minutes each, the deck runs 35 minutes against a 45-minute slot, leaving discussion time before the decision.

*Step 10.* Rehearsing from the headlines alone exposes a gap between headline three and headline four: the argument jumps from who is affected to what range extension would do, with nothing establishing that the newest lines are the ones being added. A transition slide is added, phrased as the question: "So would more lines help?" The forwarded test is run on a colleague, who identifies the spine correctly from the deck and its notes alone.

**OUTPUT.** A 14-slide flow opening on the decision and answering it by slide two, five finding headlines that read as a continuous argument, every slide carrying base, source and test status, a "what we could not establish" slide naming the two objectives the study did not answer and the stock data's coverage limit, two recommendations with confidence markers, findings named and sign-off flags, substantive speaker notes for the forwarded version, and a 21-slide appendix ordered by the anticipated question list.

## 15. Advanced usage

**Presenting to a hostile or sceptical room.** Front-load the method exposure rather than hiding it: a single slide early on that states the sample, the bases and the known limits removes the challenge that would otherwise arrive at slide ten and derail the argument. Then hold the strongest evidence for the most contested claim in the flow rather than the appendix. Counterintuitively, a deck that names its own weaknesses is harder to attack than one that does not, because the attack has already been made and answered by the person who knows the study best.

**Remote and hybrid delivery.** Attention falls faster and side conversations are invisible, so slides need to change more often, and a slide held for four minutes loses the room. Split further than you would in person, use the deck's own transitions to signal progress, and build in explicit check points where the argument pauses for questions, because remote audiences will not interrupt. Assume the deck is being read ahead of the presenter by at least some of the room.

**Decks that will be presented by someone else.** Where a client will re-present the deck internally, the read-alone standard applies regardless of the original decision at step 1, and the headline discipline becomes the entire deliverable, because the headlines are what will survive re-telling. Add a short presenter's guide covering the argument, the confidence levels and the questions to expect, and mark clearly which claims must not be simplified further.

**Multi-audience programmes.** One study frequently owes an executive readout, a working session and a team-level debrief. These are three decks and building one that serves all three produces a deck for nobody. The efficient pattern is a shared, checked evidence base and a shared headline set, with three different sequences and depths built on it. The headlines stay identical across all three, which is what keeps the versions from diverging into different claims.

**When the standard approach does not fit.** Where the study produced two incompatible readings, per **11.01** step 9, do not present a synthesis. Build the deck to present both, with the evidence for each and what would adjudicate, and open on that structure so the room is not waiting for a resolution that is not coming. This is a harder session to run and a more honest one, and it usually produces a better decision than a confident recommendation the evidence does not support.

## 16. Skill chain

**Recommended previous skills:**
- **11.01 Research Narrative Development.** Hands over the spine and the sequence, which become the slide order and the headline chain.
- **11.04 Data Visualisation and Chart Selection.** Hands over charts already correctly encoded, ordered and based, so the deck restyles rather than rebuilds them.
- **12.03 Research Report Compilation.** Hands over the checked report and evidence map, so every slide claim is already traceable.
- **11.02 Executive Summary Development.** Hands over the distilled conclusion set with confidence attached, which becomes the opening two slides.

**Recommended next skills:**
- **12.02 Research Report Design.** Applies the visual system, template and typographic hierarchy to the built deck without altering claims, bases or structure.
- **14.02 Research Debrief and Workshop Design.** Takes the argument and designs the session around it where stakeholders need to work with the findings rather than receive them.
- **12.06 Research Report QA.** Checks the deck's claims, headlines and bases against source before it is presented.
- **14.01 Insight Activation and Socialisation.** Takes the deck as the artefact that will circulate, and works on whether it produces action.

**Runs well alongside:**
- **11.03 Research Report Writing**, for the sentence craft of headlines, annotations and speaker notes.
- **07.04 Quote and Evidence Extraction**, because quotes on slides travel without their context and need their attribution and prevalence attached at the point of use.
- **K4 §4.3 and §7**, which govern base disclosure and limitation placement on a format where space pressure is constant, and **K5 §3.1**, for the sign-off every recommendation slide carries.

---
A Yazi Supplied Skill and resource.