Lesson 3 of 7 · 7 min read

Action titles and the titles-only test

Write every slide title as a full-sentence claim, then read the titles alone, in order, to check that they make the whole argument.

A title that states the point

An action title is a full-sentence headline that states the claim the slide proves. “Overtime” is a topic label. “Overtime ran 410 hours but did not hold the on-time rate” is an action title. A reader who sees only the title knows what the slide argues.

A topic label leaves the audience to work out the point. An action title states it, and the body of the slide then proves it. This is the second of the course’s four rules: one claim per slide.

Write action titles yourself, even when an assistant drafts the slides. One Hacker News user described polished AI decks that pleased management. When asked what a term on a slide meant, the presenters could not answer, because the model had chosen it. Each action title is a claim you must defend in the room.

What the experiment found

Researchers call this method assertion-evidence: a sentence headline supported by visual evidence instead of bullet lists. Joanna Garner and Michael Alley tested it in a study published in 2013. They assigned 110 engineering students at random to two groups.

Fifty-nine students saw slides with sentence headlines and visual evidence. Fifty-one saw common PowerPoint slides with topic titles and bullet lists. Both groups heard the same recorded narration about how MRI scanners detect tumors.

The sentence-headline group wrote better essays explaining the process. The effect size was d = 0.81, which means the average scores differed by about eight-tenths of a standard deviation. That group also showed fewer misconceptions and reported lower mental effort. On a surprise test one week later, they recalled more.

One result went the other way. The sentence-headline group rated the amount of text on its slides as less appropriate. The authors suggest the students expected PowerPoint’s usual bullet lists.

Three limits on that finding

The study has three limits. First, Alley, one of the two authors, created the method under test. Second, the two slide sets differed on six design principles at once, so the study cannot show which change made the difference.

Third, the participants were students watching recorded narration. They were not managers in a live meeting, deciding whether to spend money.

Emeri’s research found no controlled study of slide formats with executive or sales audiences. Every slide-format study it found tested students. The step from the classroom to your meeting is a reasonable inference, not a finding.

So the course labels this rule experimental evidence, with those limits. One controlled study supports it, and consulting practice agrees with it. That is enough to teach the method. It is not enough to promise any effect on your own audience.

Rules for writing action titles

Keep each action title to two lines or fewer, as Penn State’s guide to the method advises. A title that runs longer is often carrying two claims.

Put one claim in each title. A consulting guide from Deckary suggests that a title with “and” in it often means two slides. Check each “and”. If both halves need their own evidence, split the slide.

Write titles in the plain words your audience uses. Avoid internal jargon unless every person in the room knows it.

Use numbers when you have them, and state the direction of change. “On-time shipments fell from 97% to 88%” tells the reader more than “On-time performance declined.”

Check the house style before you change a team’s deck. A PrepLounge forum thread notes that some consulting partners dislike action titles. If the deck belongs to someone else, ask what they prefer. If they want topic labels, keep the claim in your speaker notes and in the first sentence you say.

These writing rules are expert consensus. They come from the method’s authors and from consulting practice, not from separate experiments.

Using an assistant for titles

An assistant can turn topic labels into action titles quickly. Give it your one sentence from Lesson 2 and the content of each slide. Ask for one claim per slide, in two lines or fewer.

Then check each title against your data. An assistant writes a confident claim for any slide, whether or not the content supports it. If a title claims more than the slide can prove, narrow it or mark it [TBD].

Watch for titles that use words you would not use. If you could not explain a term on your own slide, replace it with one you can. The first prompt below tells the assistant to use only your facts.

The titles-only test

The titles-only test checks the whole deck at once. Read only the slide titles, in order, and nothing else. They should make the full argument, from the one sentence to the decision. Deckary calls this horizontal logic.

If the titles read like a table of contents, they are topic labels. If one title does not follow from the one before, a step in the argument is missing. If two titles say the same thing, merge the slides.

Run the test again after every major edit, because a slide added late can break the sequence. Then read the titles aloud to a colleague who does not know the project. If they can tell you what you want and why, the titles work.

The titles-only test is expert consensus. It comes from consulting practice, and no controlled study has tested it. It takes two minutes, and it finds gaps before you spend time on charts.

Worked example · Larkfield Supply

Morgan rewrites the titles

An assistant built Morgan’s draft deck from Morgan’s notes, and every title is a topic label. Morgan rewrites each one as a claim that supports the one sentence from Lesson 2.

The draft’s titles read: Background, Peak season performance, Overtime, Customer feedback, Proposal, Costs, Next steps.
What would you do next?Run the titles-only test on these seven titles. What does it show?Think it through, then show the answer
A table of contents. A reader learns the topics but not the argument, or what Morgan wants. No title says that a second shift is the recommendation.
Morgan replaces “Background” with the one sentence from Lesson 2. It fits on two lines, and it states the ask on the first slide.
Morgan rewrites the problem slides. “Peak season performance” becomes “Last peak, on-time shipments fell from 97% to 88% of orders.” “Customer feedback” becomes “Two of our three largest customers complained in writing.”
For the overtime slide, Morgan first writes “Overtime cost $24,000 and hurt morale.”
What would you do next?What should Morgan do with this title?Think it through, then show the answer
Cut the second claim. The title holds two claims, and Morgan has no data on morale. Morgan writes “Overtime ran 410 hours but did not hold the on-time rate,” which the weekly data can prove.
Morgan turns “Proposal” and “Costs” into two claims. One reads “A four-person shift from 2 p.m. to 10 p.m. costs about $62,000 over 13 weeks.” The other reads “My estimate: a second shift brings on-time shipments back to about 96%.”
“Next steps” becomes “The decision today: approve the second shift, or repeat last year’s overtime.”
Morgan runs the titles-only test again. Read in order, the seven titles now make one argument. They give the ask, the problem, why overtime failed, the customer risk, the cost, the estimate and the decision.That answers the question you guessed at: Guess: if you read only a deck’s slide titles, in order, what should they tell you?

Morgan’s deck has seven action titles, each with one claim. Read alone, they make the same case Priya will read in the memo. Lesson 4 puts evidence under each one.

Starter prompts

Rewrite topic labels as action titles

Here is my one sentence:

[paste sentence]

Here are my slide titles and the main content of each slide:

[paste titles and content]

Rewrite each title as an action title: one full sentence that states the claim the slide proves, in two lines or fewer. Use only facts and numbers that appear in my content. If a slide has no claim it can support, write [TBD: no claim] instead of inventing one.

Run the titles-only test

Here are my slide titles in order:

[paste titles]

Read them as one paragraph. Does the sequence make a complete argument for this sentence: [paste one sentence]? List any title that is a topic label, any step that is missing, any titles that repeat, and any title that holds two claims. Do not rewrite the titles and do not add facts. List the problems only.

When not to use AI

Writing a claim you have not confirmed. An assistant will produce a confident title for any slide. Confirm the claim first, then write the title.
Changing a deck owner’s house style. If your manager or client wants topic labels, ask before you change them.

Recall from memoryHow do you run the titles-only test, and what does a failing set of titles read like?Answer it without looking back, then show the answer

Read only the slide titles, in order, and nothing else. They should make the full argument, from the one sentence to the decision. Titles that fail read like a table of contents.

Exercise · your project

Write an action title for every slide in your own deck, then run the titles-only test.

Make your first slide’s title your one sentence from Lesson 2.
Rewrite every other title as one full-sentence claim of two lines or fewer.
Split any slide whose title holds two claims that each need evidence.
Read the titles alone, in order, and fix any gap or repeat.
Check whether the deck owner or your team has a house style for titles.

Mark this lesson as done