sales-page-fixer

Sales Page Fixer

Turns a sales page that gets visits but few buyers into one that sells to the right people. It works the way David Ogilvy worked: facts first, then the headline, then one big idea, then every vague word swapped for proof, then one final test: does it sell? Use it when a sales page, landing page, offer page, or booking page is not converting, when you launch a new offer and need the page written, or when your hero section sounds like everyone else in your market.

Before you start

1. Look for a Business Context file in this project or chat (it may be called Business Context, business-context.md, or similar). If it is there, read it first and pull: what you sell, who it is for, price and what is included, client results with numbers, testimonials, guarantee, brand voice, words you never use, and funnel numbers (visits, booked calls, show rate, close rate).

2. If the user gave a URL and web browsing or search is on, read the live page now. Check that you got the headline, the offer, the price or the call details, and the testimonials. Many page builders load text with scripts, so a fetch can come back nearly empty, and proof often sits in images or video. Never conclude "the page has no facts" from a thin fetch. If there is no page yet, you will write one from scratch.

3. Ask for the page, the transcript if needed, and every missing item in ONE message. Skip anything the Business Context file or chat already answers. If nothing is missing, start. The message can cover only these:

   • The page text or screenshots, if you do not have the page or the fetch came back thin.
   • The video transcript, if the page leans on a video. You will judge its first 30 seconds like a headline.
   • The one action the page must get (book a call, buy, apply, join a waitlist).
   • Who the ideal buyer is, and the words they use to describe their problem.
   • Your best proof: client results with numbers, named testimonials, case studies. If you already have the page and the facts you have are thin (step 1 defines thin), put the 5 to 10 fact-hunt questions from step 1 here instead.
   • The offer: what they get, for how long, price or price range, guarantee.
   • Where visitors come from (cold ad, email list, referral, webinar, social post), and per month, roughly: visits, booked calls, calls held, and sales (if known).

   End the message with this line: Or reply just go and I will work with [ADD] placeholders.

4. After one round of answers, or "just go", start and deliver the full output. Never send a second round of questions before the rewrite. Mark every missing fact as a placeholder, like [ADD: number of clients served]. Never fill a gap with a made-up number. See "Output variants" below for what to deliver in each case.

How it works

Work through the steps in order. Skip a step only when it truly does not apply.

1. Facts audit (before judging a single word)

Ogilvy: "What you say is more important than how you say it."

• List every fact on the page the reader could check: numbers, names, timeframes, deliverables, credentials.
• Count the proof facts. Thin means fewer than three checkable proof facts. These are client results with a number, named testimonials, or concrete deliverables with a count or duration. "Thin" means exactly this everywhere in this skill. If the page is thin, that is the diagnosis.
• List the facts that exist but are missing from the page. Mine the Business Context file, testimonials, client messages, call notes, case studies, and the offer itself. Ogilvy's most famous car headline came from a line in a technical review of the car, not from a brainstorm. Your best headline is probably hiding in a client result or a product detail.
• If everything you have (page, Business Context file, the user's answers) is still thin, write a fact-hunt list: 5 to 10 pointed questions that pull the facts out. Example: "What is the fastest result a client got, and in how many days?" It goes in the intake message if the intake has not been answered yet, and at the top of the full output after that (see Output variants).

2. Positioning check

Positioning is what the offer does and who it is for. Ogilvy said results depend less on how you write than on how the product is positioned. Dove was not sold as soap for dirty hands. It was sold as a bar for women with dry skin.

• Write the positioning in one line: "[Offer] helps [specific buyer] get [specific result] without [main pain or fear]." This line goes in the verdict.
• A page that speaks to everyone flags no one. If the positioning is vague or wrong, say so before polishing sentences. Copy cannot rescue a fuzzy offer.
• If web search is on, read 2 or 3 competitor pages and note the claims everyone makes. Those claims are invisible to buyers, so avoid leading with them. If search is off, ask the user what competitors promise, or skip this.

3. Headline test (eighty cents of the dollar)

Ogilvy found that, in print ads of his day, about five times as many people read the headline as the body copy. So writing the headline is spending most of your dollar. On a web page, "headline" means the first screen on a phone: headline, subhead, one proof line, button. Judge it first and hardest.

Check the current headline against each point:

• Flags the right buyer. Ogilvy called the headline "the ticket on the meat."
• Promises a benefit tied to the reader's self-interest.
• Carries news if there is any (a new method, a new result, a new offer).
• Is specific: a number, a timeframe, a named outcome.
• Makes it clear what is being sold without reading further.
• Is not blind, clever, or negative. Puns and riddles lose.
• Length is fine if every word says something. A long headline that says something beats a short one that says nothing.

Match the angle to the traffic. Cold traffic: lead with the problem or result and the buyer call-out. Warm list or referral: lead with the offer and the next step.

Then write at least five new headlines, each from a different angle: result, speed, mechanism, buyer call-out, proof number, news. Pick one, say why, and name a runner-up for the test plan.

4. Big idea test

Without a big idea, Ogilvy warned, an ad passes like a ship in the night. Ask his five questions:

1. Did it make me gasp?
2. Do I wish I had thought of it?
3. Is it unique?
4. Does it fit the strategy perfectly?
5. Could it run for years? (For a coach or consultant: could it anchor the brand, not just this launch?)

One page, one idea. If the page argues three ideas, it has none. Name the one idea the whole page will serve, and the facts from step 1 it rests on. Usually it is a mechanism (why this works when other fixes failed) or a promise backed by proof. A big idea grows out of the facts. No facts, no idea: if the facts are thin, say so plainly here and in the verdict, and name the facts that would unlock it. Never invent an idea the facts cannot carry.

5. Specificity pass: swap adjectives for facts

• Hunt every adjective, superlative, and empty phrase: world-class, proven, cutting-edge, life-changing, seamless, next level, transform, unlock, game-changer, results-driven, high-quality, tailored, holistic.
• Replace each one with a fact the reader can check, or delete it. Not "proven results" but "[number] clients, average [result] in [timeframe]." Not "personalized support" but something like "two 45-minute calls a week, plus replies within one business day." That last one is an example of the form only. Use the real deliverables the user gave, or an [ADD: ...] placeholder.
• Use the exact number the user gave. If they gave a round or vague number, ask for the exact count or leave [ADD: exact number of clients]. Never turn a round number into a precise one.
• Testimonials are not copy to edit. Leave their words alone (see Rules).
• Write to one reader who is alone. Ogilvy told writers to pretend they were writing each reader a letter. Use "you," not "business owners who."
• Write the way you talk. Short words, short sentences, short paragraphs. No jargon.

6. Length: cut emptiness, not length

• Long copy sells when every sentence is interesting and factual. "You cannot bore people into buying your product." A high-ticket offer sold on a booked call usually needs more proof and more answered objections than a short page can hold.
• Cold traffic needs more proof and more answered doubts. Warm referral traffic can be shorter.
• Cut: warm-up paragraphs, repeated claims, founder story that does not serve the buyer, filler FAQ.
• Keep: anything that adds a fact, a proof point, or answers a real doubt.
• Make it work for phone scanners: write subheads that tell the whole story on their own, so someone who reads only the subheads still gets the pitch.

7. Attack and defense

Ogilvy's early sales manual split selling into two jobs: attack (build desire) and defense (remove objections). Never argue with a prospect. Let the facts close.

• Attack: the result, what the buyer's business or life looks like after, and proof that people like them got it.
• Defense: list the top 5 to 7 doubts the buyer has before they act. Common ones: "Too expensive." "I don't have time." "Will this work for my kind of business?" "I tried something like this before." "Who are you?" "Will I get pitched hard on the call?" Answer each on the page with a fact, a proof point, or a guarantee the business has agreed to.
• Place each answer where the doubt shows up, not only in the FAQ. Example: put the "will it work for me" proof right after the promise.
• There is no average buyer. If the Business Context file lists more than one buyer type, make sure the page speaks to the main one first.

8. Does it sell, or is it just clever?

Ogilvy's credo: "We sell, or else."

• One page, one action. Remove links and buttons that pull away from the main action.
• The case must be complete: offer, proof, risk reversal, and the next step spelled out (what happens after the click, how long the call is, what they get from it even if they do not buy).
• Sell to the right buyer, not to everyone. On a booked-call page, more bookings from poor-fit buyers means more no-shows and fewer sales. Say plainly who it is for and who it is not for.
• Kill lines that entertain the writer but do not move the reader.
• Family test: never write a page you would not want your own family to read. If a claim is not backed by facts, the fix is a better fact or a smaller claim, never a bigger adjective.
• Brand check: every page adds to or takes away from how people see you. Match the voice in the Business Context file. Consistency beats novelty.

9. Write the rewrite

Use this section order for a page that sells a high-ticket offer or a booked call. Hero, Problem, What you get, Proof, Offer and Final call to action are required. When facts are missing, keep them with [ADD: ...] placeholders. Drop Who it is not for, About you, Guarantee and FAQ only when there is nothing real to put in them (for Guarantee, leave the item 9 placeholder so the user decides).

1. Hero: headline, subhead (who it is for plus the result; it doubles as the qualifying line, so a poor-fit buyer can tell it is not for them), one proof line, and a button that names the action in plain words.
2. Who this is for, and who it is not for.
3. The problem, in the buyer's own words, with what it costs them.
4. The big idea: why this works when other fixes did not.
5. What you get: the deliverables the user gave, with their numbers (sessions, weeks, assets, access).
6. Proof: results with before and after numbers, and testimonials quoted word for word with name, role, and result. You choose which testimonial goes where. You never change its words.
7. About you: credentials as facts, not adjectives.
8. The offer: the price or investment range the user gave, or "what happens on the call" if the price is not shown. If price is hidden, sell the call, not the program.
9. Guarantee or risk reversal, only one the business will honor. If the user gave no guarantee, write [ADD: a guarantee you will honor, or delete this section]. You may suggest guarantee options in the verdict, clearly marked as suggestions.
10. Objections and FAQ.
11. Final call to action: repeat the promise and say exactly what happens next.

Put a button after the hero, after the proof, and at the end.

10. Read-aloud check

Ogilvy's rule: never send an important memo the day you write it. Read it aloud the next morning, then edit it. Before handing back:

• Read the rewrite as the buyer would. Cut any line that sounds like an ad instead of a person.
• Check that every number, price, guarantee, bonus, deadline, and deliverable traces back to something the user gave. Anything else becomes a placeholder.
• Check that every testimonial matches its source word for word, with any trim marked by an ellipsis.
• Tell the user to read it aloud tomorrow before it goes live, and to ask one trusted person to improve it.

Adapting to today

Ogilvy wrote for print. Keep his principles (facts, self-interest, one idea, one action) and adapt the tactics to phone screens, video on the page, and readers who scan. When you adapt a tactic, say so in one line.

What you get

Use these headings, in this order.

1. The verdict One sentence: the single most important change. Then two required lines:

• Positioning: [Offer] helps [buyer] get [result] without [pain]. Right, fuzzy, or wrong, and why.
• Big idea: the one idea the page will serve, and the facts from the audit it rests on. If the facts are thin, say so and name the facts that would unlock it.

Add any flags here (earnings claims that need a typical-results note, suggested guarantee options marked as suggestions).

2. Facts audit

Fact	On the page now?	Where it should go

Then a short "Missing facts" list, each with the question that will get it.

3. Headline verdict Current headline: PASS or FAIL, with the reason in one line.

#	Headline option	Angle	Why it works

Pick: which option, and why, in two lines. Runner-up: which option you would test next.

4. Cut and prove

Before (current line)	Problem	After

At least five rows on a full page, taken from the page itself. Never put a testimonial in this table. Testimonials are quoted, not rewritten.

5. Objection map

Objection	How the page answers it	Proof used

6. The rewritten page Full copy, section by section, labeled with the section names from step 9. Missing facts appear as [ADD: ...].

7. Test plan

The numbers to watch. Fill in the baseline if the user gave one:

Number	Old page (baseline)	New page
Booked calls per 100 visitors
Show rate (calls held per call booked)
Close rate (sales per call held)
Sales per 100 visitors

• A win needs both: more booked calls per 100 visitors, and more sales per 100 visitors (or at least as many calls held and closed from the new page). If the page sells with no call, sales per 100 visitors is the one number. Sales come after bookings, so count closes from calls booked on the new page even when they close after the test window.

• If bookings rise but show rate or close rate fall, the page is attracting the wrong buyers. Tighten the Who it is not for section and the qualifying line in the hero before you call it a win.

• No baseline: tell the user to note current visits, booked calls, calls held, and sales before the new page goes live. Direct-response rule: hold the page to a number.

• How to test: before replacing, duplicate or save the old page so you can roll back. Keep traffic sources and ad spend the same during the comparison. Judge only after the new page has had at least as many visitors as the baseline period, and treat small differences on small numbers (a few calls either way) as noise. Rule of thumb, not a law: split test only if each version can get roughly 100 or more conversions in a month. Otherwise, replace and compare. When you do split test, test the headline first: headline A against headline B.

• What to change next if the numbers do not move, in this order:

  1. Test the runner-up headline from section 3.
  2. Move the strongest proof directly under the hero.
  3. Make the offer and the next step clearer: show the price range (only if the business agrees to show it) or what happens on the call, and the guarantee (only one the business will honor).

  If still flat after these, the problem is likely the offer or the traffic, not the page. Run the offer-stress-test or funnel-doctor skill from this pack.

Output variants

• Thin facts and the intake has not been answered yet: your reply is the one intake message from Before you start. Open it with 1. The verdict, put the 5 to 10 fact-hunt questions in place of the proof question, end with the just go line, then stop and wait.
• After one round of answers, or after "just go": always deliver the full output. If facts are still thin, put the remaining fact-hunt questions at the top and keep [ADD: ...] placeholders in the page.
• No page yet: skip the current-headline PASS or FAIL and section 4. Deliver the verdict, facts audit, headline options, objection map, rewritten page, and test plan.
• Hero or headline only: deliver sections 1, 3, and 4, plus the rewritten hero. Cut and prove gets up to five rows.

Rules

• Never invent numbers, results, testimonials, client names, credentials, prices, guarantees, bonuses, deadlines, capacity or deliverables. Missing facts become [ADD: ...] placeholders.
• Quote testimonials word for word. You may trim with an ellipsis, but never reword, merge, or add numbers to them. Choose which testimonial goes where. If a stronger quote is needed, write the question to ask the client, not the quote.
• Never use client names or client numbers the user has not cleared for public use. Ask if unsure.
• No hype words, no exclamation marks, no fake urgency or fake scarcity (timers that reset, "only 3 spots left" when it is not true).
• No income or health promises the business cannot back with typical results. If the page makes one, flag it and offer a smaller, provable version.
• When proof shows client income or revenue, do not present the best result as normal. Pair it with the typical or average result if the user has it, or flag it in the verdict: "This result may need a typical-results note. Check with your advisor before publishing." Keep it a flag, not legal advice.
• One page, one big idea, one main action.
• Every swap replaces a vague claim with a fact, or deletes it. Never swap one adjective for another.
• If the offer or the positioning is the real problem, say so plainly. Better copy will not fix it.
• Never call the page a win on bookings alone. Check show rate, close rate, and sales too.
• Quote Ogilvy sparingly, using only the short lines in this file. Never invent a quote. No costume, no fake British phrasing. The value is the method.
• Keep the prose around the tables short. A full rewrite can be as long as the page needs. Never pad.
• Write the new copy in the user's voice from the Business Context file, not in a generic AI voice.

Part of the free 20-skill pack by Ruben Davoli at BeaverMind (beavermind.ai).

Built on the published work of David Ogilvy. Not affiliated with or endorsed by David Ogilvy, the Ogilvy estate, or the Ogilvy agency.