Pitch One Liner

Pitch one-liner

Turn a company's fuzzy self-description into one sentence a stranger can repeat correctly after hearing it once, then push that sentence into every place it needs to live.

Why one sentence carries so much weight

The best companies grow by word of mouth, which is the cheapest distribution that exists. Word of mouth only works if the sentence survives retelling. If a user cannot repeat what you do at a dinner table, the growth loop never starts.

The sentence is also a diagnostic. People who have thought deeply about an idea can explain it in fewer words. A muddy one-liner is usually not a copywriting problem. It is a sign the founder has not decided who the customer is, or is hedging between two products. When you hit that, name it. Handing back polished words that paper over an undecided strategy does the founder no favors.

Environment check

Run this first. It takes one line of thought and it prevents the two failure modes that ruin this skill in an unfamiliar environment: stalling on a tool that is not there, and claiming to have read something you never read.

| If this environment can | Use it to | If it cannot | |---|---|---| | Read attached or uploaded files | Read the deck, the doc, the screenshots | Ask for the text pasted in | | Fetch a URL | Pull the live hero, subhead, meta description, nav labels | Ask for the hero copy pasted in, or a screenshot | | Drive a browser | Render a client-side site the fetch returned empty | Say the page did not render and work from what you have | | Run code | Nothing here needs it | Ignore |

Nothing in this table is required. This skill works on pasted text alone. State in one line at the top of your output what you actually worked from, and never describe a page you could not open.

Workflow

1. Get the raw material

Work from what they actually publish, not from what they tell you they publish. The gap between the two is often where the problem lives.

| They gave you | Do this | |---|---| | A URL | Fetch it. Pull the hero headline, subhead, meta description, and nav labels. If the fetch returns an empty shell because the site is client-rendered, try a browser tool if you have one, otherwise ask for a paste | | A deck (.pptx, .pdf, Slides link) | Read it. Focus on slide 1, the problem slide, and any what-we-do slide. Note the exact wording. If you cannot extract the text, say so and ask for the title slide as text | | Pasted copy | Use as-is. This is enough. Do not ask for more access than the job needs | | Screenshots | Read what is legible. Do not guess at small type, call it unreadable | | Nothing, just a conversation | Skip to the interview in step 2 |

Look for the artifact wherever this environment puts user files: attached to the message, earlier in the conversation, in the working directory, or in an uploads folder. If you cannot find it, list what you can see and ask, rather than assuming a path.

Then find the current de facto one-liner: the sentence they are already leading with. Quote it verbatim before you critique it. If there are several competing versions across the site, deck, and LinkedIn, list them side by side. Inconsistency across surfaces is itself a finding worth reporting.

2. Fill the four gaps

You cannot write the sentence without four facts. Extract what you can from the material, then ask only for what is genuinely missing. Do not ask for all four if the landing page already answers three.

1. What is it, literally? If a competent engineer heard your sentence, could they sketch what to build? Marketplace, API, browser extension, managed service, physical device, model.
2. Who is it for? A specific buyer, not businesses. Compliance leads at mid-market manufacturers is a customer. Enterprises is a rounding error.
3. What changes for them? The job it replaces or the pain it removes, stated concretely.
4. What is the wedge? The one thing it does better or differently than the obvious alternative, including doing nothing.

If they cannot answer 2 or 4, stop and say so. That is the real blocker.

3. Diagnose before you write

Score the current line against the rubric below. Then name the specific failure mode and quote the offending fragment. Vague feedback (it is a bit unclear) does not help anyone rewrite. The phrase intelligent workflow layer could describe forty companies does.

4. Generate three distinct angles, not three rewordings

Three near-identical options are a waste of the founder's time. Make them structurally different so the choice reveals something about how they want to be understood:

• Angle A, the mechanic. Leads with what the thing literally is. Safest, most legible, best for investors and technical buyers.
• Angle B, the outcome. Leads with what changes for the customer. Best for landing pages and buyers who do not care how it works.
• Angle C, the category claim. Leads with the space they intend to own, which is the boldest and the easiest to get wrong. Only offer this if the wedge in step 2 is genuinely strong. If it is not, say so and offer a second outcome variant instead.

Score all three. Recommend one and say why, in one line. Do not hedge across all three; founders have enough ambiguity already.

5. Convert the winner into the assets

The sentence is worthless sitting in a chat window. Ship it into place. Always produce these, using the recommended line as the spine:

Landing page hero

• H1: the one-liner, or a tightened version under 10 words
• Subhead: one sentence adding the specificity the H1 dropped, naming the buyer
• Primary CTA: what a stranger should do in the next 5 seconds

Deck slide 1

• Company name
• The one-liner, verbatim, as the only body text
• Note: this line goes on the title slide, before the problem slide. Context first, backstory second.

Cold email opener

• Two sentences. The one-liner adapted to second person, plus one line on why this specific recipient.

LinkedIn or bio line

• Under 120 characters, works in a headline field.

Dinner table version

• How a happy user would say it out loud to a friend, in casual speech. If this one is hard to write, the primary line is still too stiff. Fix the primary line, do not just soften this one.

Keep every asset consistent with the chosen line. If the hero says one thing and slide 1 says another, you have reintroduced the original problem.

The rubric

Score each candidate 0 to 2 per criterion. 10 or above ships. Below 8, rewrite rather than tweak.

| # | Criterion | 0 | 1 | 2 | |---|---|---|---|---| | 1 | Reproduction | Listener has no idea what you'd build | Rough guess | Could sketch it | | 2 | Buyer named | No one identifiable | Broad category | Specific buyer | | 3 | Jargon-free | Multiple abstract nouns | One buzzword | Plain speech | | 4 | What before why | Opens with problem or mission | Mixed | Opens with what it is | | 5 | Concision | Over 25 words or two clauses fighting | 15 to 25 words | Under 15 words | | 6 | Repeatability | Nobody can retell it | Retold approximately | Retold accurately |

Criterion 1 is the load-bearing one. A line can pass everything else and still fail if the listener cannot picture the product.

Three fast tests, useful to run out loud:

• The rebuild test. Could a competent builder sketch what to make from this sentence alone?
• The airport test. Could a stranger with no context repeat it to someone else an hour later?
• The swap test. Replace the company name with a competitor's. Does the sentence still read as true? If yes, it describes a category, not a company.

Failure modes

Name these explicitly when you see them. Founders fix what they can name.

Why-first. Opening with origin story, market pain, or mission. It feels natural because we tell stories chronologically, and ambitious founders want to establish stakes. But the listener has no container to put the story in yet. Give them the what, and the why lands harder thirty seconds later when they ask.

• Weak: Founders waste months on compliance paperwork and it is holding back the whole sector, so we built...
• Fixed: lead with the product, let them ask why.

Jargon fog. Abstract nouns doing the work concrete ones should. Platform, solution, ecosystem, enablement, synergy, know-how, seamless, end-to-end, next-generation, AI-powered when the AI is not the point. A line built from these could describe almost anything. Test it by asking what you would have to build to reproduce it.

Rambling. Two or three sentences of setup before the point. The listener has already left. Press, investors, and other founders get pitched dozens of times a week.

Feature listing. Naming three capabilities instead of one job. Readers remember one thing. Pick it.

Tech-attribute padding. Leading with how it works when the buyer does not care. ML-driven mental health for your team is weaker than personalized digital therapy programs for your team, because the buyer wants personalized, not ML. Strip implementation details unless the implementation is the differentiator.

Category inflation. Claiming a category the company has not earned. Fails the swap test immediately.

Audience vapor. For businesses, for teams, for everyone. Specificity is not a limit on the market, it is what makes the sentence spreadable inside one.

Hedging. Two products joined by and, because the founder has not chosen. This is a strategy problem wearing a copywriting costume. Say so directly.

The X for Y construction

Uber for X and Airbnb for X are overused, and every company should have a primary line that does not rely on the analogy. Still, it works in narrow conditions. All three must hold:

1. X is a household name to this specific listener. If they have to look it up, it has failed.
2. It is obvious why Y would want X. If the analogy needs a sentence of explanation, it is not doing its job.
3. Y is a large market. A precise analogy pointed at a small market wastes the listener's attention.

Note that requirement 1 is audience-dependent, not absolute. Stripe for former Soviet Union countries works with Silicon Valley investors and fails with a regional bank. When talking outside the bubble, payment processing for former Soviet countries is stronger. Always give the founder both versions and tell them which room each belongs in.

A line like Airbnb for dance and movement classes is legible, but a marketplace for dance and movement classes is usually just as clear and does not borrow someone else's brand. Default to the plain version and offer the analogy as a situational tool.

Worked examples

Example 1: why-first, from a compensation startup

Before: Many of my friends are super confused about their stock options and it stresses them out, so we are fixing pay transparency. Score: 1/2/1/0/1/0 = 5 After: Pave lets companies plan, communicate, and benchmark compensation in real time. Score: 2/2/2/2/2/2 = 12 What changed: the what moved to the front, the buyer became companies rather than friends, and the backstory got saved for the follow-up question.

Example 2: jargon fog

Before: A know-how and synergy platform. Score: 0/0/0/1/2/0 = 3 Why it fails: this could be an edtech collaboration tool or a marketplace for expert consultants. Nothing in the sentence narrows it. Short is not the same as clear. Fix path: ask what the user does in the product on day one, and write that.

Example 3: tech-attribute padding

Before: Adaptive ML-driven mental health for your team. Score: 1/2/1/2/2/1 = 9 After: Personalized digital therapy programs for your team. Score: 2/2/2/2/2/2 = 12 What changed: the buyer wants personalized care, not a model architecture. The ML is a means, not a benefit, so it belongs on the technology page rather than the hero.

Example 4: a clean baseline to aim at

The first online marketplace that lets travelers book rooms from locals instead of hotels. (Airbnb, YC application) Score: 2/2/2/2/2/2 = 12 Why it works: names the mechanism (marketplace), the buyer (travelers), the job (book rooms), and the alternative it displaces (hotels), in eighteen words with zero abstraction.

Output

Return the report as markdown in your reply. Lead with the recommendation, because founders skim. If this environment delivers files, or the user asks for a document, write the same report to a file named -one-liner.md, hand it over, and keep the reply to the recommended line plus two sentences.

## Recommended
> [the line]
[One sentence on why this one.]

## What you're saying now
> [verbatim current line, and where it appears]
Score: X/12. [Failure mode named, offending fragment quoted.]
[If versions conflict across surfaces, list them here.]

## Three angles

| # | Angle | Line | Score | Best for |
|---|---|---|---|---|
| A | Mechanic | ... | /12 | ... |
| B | Outcome | ... | /12 | ... |
| C | Category | ... | /12 | ... |

[Two lines max per angle on the tradeoff.]

## Ready to paste

**Landing page hero**
H1 / Subhead / CTA

**Deck slide 1**
...

**Cold email opener**
...

**LinkedIn headline**
...

**Dinner table version**
...

## What to fix beyond the words
[Only if diagnosis surfaced a real positioning or strategy problem. Be direct. Skip this section if there isn't one.]

Tone

Write like a good editor, not a brand agency. Direct, specific, quoting the actual text. Do not praise a line that scores 7 out of politeness. A founder who ships a weak sentence pays for it in every conversation for the next year, so the honest read is the useful one. Point at the fragment, name the mode, hand back the fix.

Never fabricate. Do not invent traction, customers, competitors, or a product capability the source material never claimed, in order to make a line land better. An invented detail in the hero is a promise the company has to keep.

What to offer next

Pick one, based on what the diagnosis surfaced:

• Rewrite the full hero section, or the whole landing page, around the chosen line
• Rebuild the deck title slide and the problem slide that follows it
• Write the cold outreach sequence that opens with this line
• Work the positioning question underneath, if that turned out to be the real blocker

If a pitch deck review or landing page review skill is installed, hand off to it here. If not, you can do the work yourself with the same standards.

---

Where this comes from

Built from published guidance on positioning and pitch clarity, including Y Combinator's public writing and talks on explaining what your company does, plus the application and demo day material where the one-sentence discipline originated. The judgment layer comes from a marketing leader who has written and rewritten these sentences with founders and watched which versions actually spread.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: lelephi
• Source: lelephi/marketing-agent-skills-for-founders
• License: MIT

Install and usage instructions live in the source repository linked above.