speech-design

Speech Design

Designs keynotes and conference talks built for provocation, not education. Produces a talk structure with one core argument, one anchor story, one emotional peak, and one memorable close. Tim brings the insight and stories — this skill provides the architecture that keeps him from drifting into training mode.

How It Works

BRIEF (gather context)
    ↓
CORE ARGUMENT (one sentence)
    ↓
STORY SELECTION (anchor + supporting)
    ↓
EMOTIONAL ARC (open → build → peak → resolve → close)
    ↓
TALK STRUCTURE (timed, with delivery notes)
    ↓
THE ONE LINE (what they'll repeat tomorrow)

Step 1: Gather the Brief

KEYNOTE BRIEF
─────────────
Event/Host:          [Who's putting this on?]
Event Name:          [Conference, summit, etc.]
Talk Title:          [If assigned — may need to change]
Duration:            [Minutes — hard stop?]
Audience:            [Who's in the room? Size? Roles? Sector?]
Why Tim:             [What made them pick Tim specifically?]
Audience Pain:       [What problem does this audience have?]
Desired Shift:       [How should they feel DIFFERENTLY when they leave?]
Adjacent Talks:      [What else is on the agenda? Don't duplicate.]
Constraints:         [Topics to avoid, format rules, tech setup]
Tim's Current Angle: [What does Tim think the talk should be about?]

Step 2: Define the Core Argument

The core argument is the ONE reframe the audience hasn't heard. Not a topic. Not a category. A specific claim that shifts perspective.

Bad core arguments (topics, not arguments):

• "Expectations matter"
• "The human side of technology"
• "Change management for nonprofits"

Good core arguments (specific, reframing, provoke disagreement):

• "The nonprofit tech industry is over-investing in experience and ignoring the lever that actually determines outcomes: expectations"
• "Three Disneyland cast members making $450/day outperformed a $994M ride — because they managed expectations, not experiences"
• "Your CRM implementation failed not because the technology was wrong, but because nobody managed what people expected it to feel like"

CORE ARGUMENT
─────────────
DRAFT 1: [First attempt]
DRAFT 2: [Sharper version]
DRAFT 3: [Sharpest — usually the one]

THE ARGUMENT IN ONE SENTENCE:
[This is the sentence Tim should be able to say from memory.
 If he can't say it in one breath, it's not sharp enough.]

REFRAME TEST:
- What does the audience believe BEFORE the talk? [Current frame]
- What should they believe AFTER? [New frame]
- What's the moment of surprise between the two? [The shift]

Step 3: Select the Story Architecture

A keynote needs ONE anchor story that carries the entire argument. Supporting stories are optional and brief.

STORY ARCHITECTURE
──────────────────
ANCHOR STORY: [The one story that IS the talk]
- Setup:     [What's the situation? Who are the characters?]
- Tension:   [What goes wrong or what's at stake?]
- Turn:      [The moment that changes everything]
- Landing:   [What it means — but DON'T over-explain]

CONNECTION TO ARGUMENT: [How does this story PROVE the core argument
                         without Tim having to say "and the lesson is..."]

SUPPORTING STORIES (0-2 max):
- [Brief story 1 — what it reinforces]
- [Brief story 2 — what it reinforces]

DATA POINT(S) (0-2 max):
- [Surprising stat that creates an "oh" moment]

Story selection criteria:

1. Does it carry the argument without explanation?
2. Does it create an emotional moment (surprise, recognition, tension)?
3. Is it Tim's own story or observation? (First-person > third-person > research)
4. Can Tim tell it from memory without notes?
5. Does it have a "turn" — a moment where the meaning shifts?

Step 4: Design the Emotional Arc

EMOTIONAL ARC DESIGN
────────────────────

DURATION: [X minutes]

PHASE    | TIME  | WHAT HAPPENS                  | EMOTION TARGET
─────────|───────|───────────────────────────────|──────────────────
HOOK     | 0-2m  | [Opening moment — question,   | CURIOSITY
         |       |  surprising stat, or in       |
         |       |  medias res story start]       |
         |       |                               |
BUILD    | 2-Xm  | [Anchor story setup +         | TENSION /
         |       |  tension. Audience feels       | RECOGNITION
         |       |  the problem in their gut]     |
         |       |                               |
TURN     | Xm    | [The moment of reframe.       | SURPRISE
         |       |  This is where the core        |
         |       |  argument LANDS through        |
         |       |  story, not explanation]        |
         |       |                               |
PEAK     | X-Ym  | [Highest emotional intensity. | AWE / CONVICTION
         |       |  Usually the data point or     |
         |       |  the story's most powerful     |
         |       |  moment. Should be in final    |
         |       |  third of the talk.]           |
         |       |                               |
CLOSE    | Y-end | [The one line. The call to    | RESOLVE /
         |       |  change. NOT a to-do list.     | DETERMINATION
         |       |  NOT "in summary." A moment    |
         |       |  that echoes.]                 |

Step 5: Produce the Talk Structure

TALK STRUCTURE: [Title]
EVENT: [Event] | DATE: [Date] | DURATION: [X min]

THE ONE LINE: "[The sentence they'll repeat tomorrow]"

TIME | SECTION        | WHAT TIM DOES                        | DELIVERY NOTE
─────|────────────────|──────────────────────────────────────|──────────────
0:00 | HOOK           | [Exact opening — first 30 seconds    | [Energy: start
     |                |  matter most. Script this precisely.] |  at 7/10, not
     |                |                                      |  10/10]
     |                |                                      |
2:00 | STORY SETUP    | [Begin anchor story. Set the scene.  | [Slow down here.
     |                |  Characters. Context. Make them care.]|  Details matter.]
     |                |                                      |
X:00 | TENSION        | [Build the problem. Let it sit.      | [Don't rush to
     |                |  Don't resolve it yet.]               |  the answer.]
     |                |                                      |
X:00 | THE TURN       | [Deliver the reframe. This is the    | [PAUSE after
     |                |  moment. Don't explain it — let it   |  the turn. Let
     |                |  land.]                               |  it breathe.]
     |                |                                      |
X:00 | PEAK           | [Highest intensity. The stat, the    | [This is where
     |                |  most powerful story beat, the        |  Tim's energy
     |                |  moment of awe.]                      |  peaks. Full
     |                |                                      |  conviction.]
     |                |                                      |
X:00 | CLOSE          | [The one line. The call to change.   | [Memorize this.
     |                |  Full stop. Don't add anything       |  Say it. Stop.
     |                |  after it.]                           |  Don't add.]

IF RUNNING LONG: Cut [specific section — never cut the close]
IF RUNNING SHORT: Expand [story detail in build section — never pad the close]

Design Rules

1. Write the close first. Everything else serves the close. Tim's pattern is amazing open, weak close. Reverse it.
2. One argument, not three. Tim's Quickstart 9 generates 5 arguments per minute. Pick the strongest. Kill the rest.
3. Story carries the argument. If Tim has to SAY "and the point is..." the story didn't work. Redesign it.
4. No framework slides. If a 2x2 matrix or a step-by-step process appears, it's a training. Convert to story or cut.
5. The one line must be repeatable. Test: would an audience member text this to a colleague? If not, sharpen it.
6. Peak in the final third. Front-loaded talks die at the end. Build toward the peak.
7. End with the mic drop, not the to-do list. "5 things to do" is a training close. A keynote close is a single haunting challenge.
8. Silence is a tool. After the turn, after the peak, after the close — script the pauses. Tim fills silence with explanation. Don't.

Tim's Keynote Strengths to Leverage

• Disneyland story: Keynote-grade. The niece, the stormtroopers, the rebel spy, the $450 vs. $994M stat. This is a complete talk in one story.
• The reframe: "Your tech is fine. Your humans don't know how to human." This is a core argument.
• Warmth + conviction: Tim's High I creates genuine connection. Don't bury it under slides.
• Data surprises: Tim finds stats that make people lean forward. Use them at the peak, not the preamble.

Cross-Skill Integration

• After speech-audit: Audit identifies structural problems → Design fixes them
• Before speech-rehearse: Design produces the structure → Rehearse coaches delivery
• With story-finder: If the anchor story isn't strong enough, story-finder searches for alternatives
• With tim-voice: Calibrate the talk's language to Tim's authentic voice
• With slide-reviewer: After design, slides should be reviewed for keynote mode (not training mode)

Triggers

• "build a keynote"
• "design my talk"
• "write my speech"
• "speech design"
• "new keynote"
• "I have a talk coming up"

What This Skill Does NOT Do

• Audit an existing talk (use speech-audit)
• Coach delivery (use speech-rehearse)
• Evaluate post-delivery (use speech-evaluate)
• Design training sessions (use instructional-designer: session-architect)