Strategy Deck Checklist: Check the Argument Before the Slides

The most important line was at the bottom

On January 24, 2003, with the space shuttle Columbia in orbit, NASA's Debris Assessment Team presented its analysis of a foam debris strike on the left wing, 81.9 seconds after launch, to the Mission Evaluation Room ¹. The briefing relied on PowerPoint slides from Boeing ¹. One slide was titled “Review Of Test Data Indicates Conservatism for Tile Penetration” ¹.

Only at the bottom of that slide did the engineers state a key fact: one estimate of the debris was 640 times larger than the data used to calibrate the model behind their damage assessment ¹. Later analysis put it at 400 times ¹. The slide stacked six levels of hierarchy onto what was, in substance, 11 simple sentences, and used “significant” or “significantly” five times with very different meanings ¹. The accident investigation board, drawing on an analysis by Edward Tufte, concluded that “it is easy to understand how a senior manager might read this PowerPoint slide and not realize that it addresses a life-threatening situation” ¹.

No strategy deck carries those stakes. But the failure is the same one that sinks strategy decks every week: the decisive fact sits where the reader will not look, under a title that says the opposite. Nothing about the fonts would have fixed it.

A strategy deck checklist is a set of checks to run before a deck that asks for a strategic decision goes to its audience. Run it in five passes, in order: the ask, the storyline, every number, each slide, then the send. The first three catch the errors that change the decision; the last two catch the ones that embarrass the author. Most reviews spend their time the other way around.

The most important sentence in the deck should never be at the bottom of a slide.

Why the argument comes before the slides

Reviewers fix what they can see. A misaligned chart is visible from across the room. A recommendation that answers a question nobody asked is invisible until the meeting, when the most senior person says so. That is why review time drifts to formatting, and why it should be spent in the opposite order.

Tufte's critique of slideware is that it “often reduces the analytical quality of presentations,” and that popular templates “usually weaken verbal and spatial reasoning, and almost always corrupt statistical analysis” ². You do not have to accept all of that to accept the practical point: the format rewards fragments, and a strategy argument is made of connections. The checks below are mostly about restoring the connections.

Two structural ideas underlie them. Barbara Minto's: thinking is easy to grasp when the ideas are presented “as a pyramid under a single point” ³. And Michael Alley's assertion-evidence approach: each slide carries a sentence headline stating its main message, supported by visual evidence. In a study summarized on his Penn State research page, audiences shown assertion-evidence slides comprehended and recalled the content significantly better than audiences shown common-practice slides for the same talk ⁴. The pyramid principle page covers the storyline in depth; this page is the inspection.

The five passes

1. Pass 1: The askBefore opening the file. Is there a decision, a decider and a reason to decide now?
2. Pass 2: The storylineTitles only, read in a row. Does the argument hold, and is the bad news where it belongs?
3. Pass 3: Every numberSource, date, unit and comparison for each figure, and agreement across slides.
4. Pass 4: Each slideOne message per slide, a title that states it, evidence that proves it.
5. Pass 5: The sendPre-read timing, pre-wiring, the summary page and the backup slides.

Do not merge passes. A reviewer reading numbers will not notice that the storyline answers the wrong question, and a reviewer reading the storyline should not be distracted by a chart's colors. Thirty minutes for pass 1 and 2 together saves days when it catches a problem; a pass 4 finding never does.

Pass 1: the ask

• The decision is written as a question with at least two real answers, on the first or second slide.
• One person is named as the decider, and the deck is built for that person, not for the room.
• The recommendation is on slide one or two, in a sentence, not revealed on slide 19.
• The deck says why the decision is needed now: what it costs to wait a quarter.
• Real alternatives appear, including doing nothing where that is a genuine option, with why each lost.
• The deck states what must be true for the recommendation to work.
• The final slide says what you want the audience to do at the end of the meeting: approve, choose, fund or reject.
• You have asked whether this should be a deck at all. A decision with many interlocking reasons may be better argued in a narrative memo.

Pass 2: the storyline

• Read only the slide titles, in order, aloud. They form a paragraph that makes the case on their own.
• Every title is a full sentence that states a finding, not a topic label such as “Market overview.”
• Each title claims no more than its slide's evidence supports.
• Each section opens with a summary that says what the slides beneath it mean, not that there are three of them.
• The bad news is in the main body, next to the claim it threatens, not in an appendix or a footnote.
• The biggest risk to the recommendation has its own slide with its own title.
• The situation and background take no more than a quarter of the main slides.
• Nothing in the main body is there only because it took a long time to produce.

The Columbia slide fails the second and third checks at once. Tufte suggested its title should have read “Review of Test Data Indicates Irrelevance of Two Models” ¹. Same data, opposite message, and the right one. Our action title drill gives ten rounds of practice at exactly that rewrite. If the titles do not read as an argument, stop the review; the fix is a new storyboard, which storyboarding a deck before you build it walks through.

Pass 3: every number

• Every chart and table names its source and the date of the data.
• Every figure that appears twice is the same both times. Market size, headcount and baseline revenue are the usual offenders.
• Units are written one way throughout. The Columbia slide gave the same unit three ways: “3cu. In,” “1920cu in” and “3 cu in” ¹.
• Every number has a comparison: last year, plan, a competitor or a benchmark. A number alone cannot be judged.
• Forecasts are shown as ranges, each tied to the assumption that moves it most.
• Vague quantity words are replaced or defined. “Significant” on a slide means whatever the reader fears or hopes.
• Percentages state their base, and changes say whether they are percent or percentage points.
• Someone other than the author has recomputed the three numbers the recommendation depends on most.

The last check matters most. It is also the one most often skipped, because by the time the deck is built nobody wants to find a problem. Our analysis QA checklist covers how to test the analysis underneath the numbers before it reaches a slide.

Pass 4 and pass 5: each slide, then the send

Pass 4: each slide.

• One message per slide. If the title needs “and,” consider two slides.
• The chart proves the title. Cover the title and ask a colleague what the chart shows; their answer should match it.
• Caveats sit next to the claims they qualify.
• The slide is readable printed in black and white, which is how some directors will see it.
• Labels are on the data, not in a legend the reader must decode.
• Detail that only the author needs has moved to the appendix.

Pass 5: the send.

• The pre-read goes out with enough time to read it, and the date is fixed in advance.
• The decider and the people most likely to object have seen the recommendation privately before the meeting. Managing up covers that pre-wiring.
• A one-page summary leads the deck, and a reader who stops there can act. The executive summary template shows how.
• The question you most fear has a backup slide, and you know its number.
• The file name and cover carry a version and date, so nobody argues from an old draft.

A deck run through the checklist

Notice which findings changed the decision. The pilot condition came from moving the bad news forward in pass 2. The eight-color legend changed nothing.

When checklists make decks worse

The strongest objection to a deck checklist is that it produces defensive decks. Every number footnoted, every risk given a slide, every alternative shown: the deck becomes a legal brief, grows to 60 slides and loses the reader. A second objection cuts deeper. A deck that passes every check can still be wrong, and its polish makes it harder to challenge.

Both are fair. The answer to the first is the order of the passes. Passes 1 and 2 cut; they ask what the decision needs and remove the rest. A team that runs pass 3 on a deck that never passed pass 2 will footnote its way to 60 slides. The answer to the second is that the checklist cannot supply judgment, only expose where it is missing. It catches buried bad news. It cannot tell you that the strategy is a bad one. For that, test the content against what strategy is, and what it gets mistaken for.

There is also a political edge. Moving bad news forward is the check most often vetoed, usually by someone senior who prefers the deck to land well. The investigation board's broader point about NASA was not about one slide: it called the “endemic use of PowerPoint briefing slides instead of technical papers” an illustration of problematic methods of technical communication ¹. In a strategy team, the equivalent habit is letting the slide's tone be negotiated separately from its content. If you run the strategy function, make pass 2 non-negotiable: the risk slide stays where the risk is.

The short version

1. Can the decider see the decision and your answer by slide two?
2. Do the titles, read alone, make the case?
3. Is the bad news next to the claim it threatens?
4. Has someone else recomputed the three numbers that matter most?
5. Does every chart prove its title?

If you only have ten minutes, ask those five. The rest of the checklist protects the author. These five protect the decision, which is the only reason the deck exists.

Questions

What should a strategy deck include?

A decision written as a question, the recommendation in a sentence on slide one or two, the reason to decide now, the real alternatives and why they lost, what must be true for the recommendation to work, the biggest risk on its own slide, and a closing slide that says what you want the audience to do. Everything else supports one of those or moves to the appendix.

How do you review a strategy presentation?

In five passes, in order: the ask (decision, decider, recommendation), the storyline (titles read alone make the case), every number (source, date, unit, comparison, consistency), each slide (one message, a title that states it, evidence that proves it) and the send (pre-read, pre-wiring, summary, backup slides). Do the first two before anyone formats anything.

Where should bad news go in a strategy deck?

In the main body, next to the claim it threatens, with a title that states it. Bad news in an appendix or at the bottom of a slide is easy to miss. The Columbia accident investigation board found a key fact about debris size stated only at the bottom of a briefing slide whose title suggested the opposite.

How many slides should a strategy deck have?

There is no right number. The test is whether every main-body slide supports the recommendation or answers a question the decider will ask. A useful rule of thumb is to keep background to no more than a quarter of the main slides and move detail only the author needs to the appendix.

Where to go next

• The Pyramid Principle for Strategists
• Storyboard Your Deck Before You Build It
• Analysis QA Checklist: Before It Goes in the Deck
• Board Deck: The 7 Components of a Board Presentation
• Slide and presentation tools for strategy teams
• The One-Page Decision Memo