Competitors Prove the Claims Your Pages Only Make

Two plumbing service pages. Same city, same query, same intent. Open them side by side and the keyword coverage is nearly identical: emergency plumber, burst pipe, 24 hour, licensed, same suburb names, same schema type, similar word counts.

The losing page says: fast, reliable, fully licensed, trusted by locals.

The winning page says: Licence #12847. Average arrival time 43 minutes across 1,914 callouts last year. 12-month workmanship warranty in writing. Lead technician: Dave Morrissey, 19 years on the tools.

Both pages make the same promise. One of them proves it. That difference is not a keyword gap and it will never show up in a keyword gap tool, because both pages target the identical terms. It is a proof gap, and it is the most common reason two similar pages perform completely differently.

Claims are cheap and buyers have learned that

Every service page on the internet says fast, reliable, professional, trusted, experienced. Those words cost nothing to type, which is exactly why they carry no weight. A buyer comparing three tabs is not reading for adjectives. They are scanning for something they can check, count, or hold you to.

The shift underneath this is worth stating plainly. Local discovery is moving away from term coverage and toward evidence density. When an AI answer engine assembles a response about who to call in your suburb, it is pulling specifics: named credentials, stated response windows, warranty terms, dated outcomes. A page full of adjectives gives that system nothing to extract. A page full of verifiable facts gives it a dozen extractable statements.

There is a measurement problem layered on top. Search Engine Journal's September 2026 reporting found that Search Console under-reports positioning inside AI search surfaces, which means the rank number you are looking at describes less of your actual visibility every quarter. Chasing that number is chasing a shrinking signal. Auditing the proof assets on your own page is auditing something you fully control and something both humans and machines read the same way.

Sampled performance signals across weak-CTR pages make the cost concrete. Tens of thousands of impressions returning roughly one click per hundred is the classic pattern of a page that ranks, gets seen, and gets skipped. Those pages are almost never losing on relevance. They are losing at the moment a buyer looks for a reason to choose.

Build the proof inventory from three competitor pages

Pick one money query. The one that actually produces phone calls, not the vanity term with volume. Open the top three organic results and any business the AI answer names. Three pages is enough. Ten is procrastination.

Read each page line by line and log every claim into three columns.

• Column one: verifiable claims with evidence attached. A licence number you could look up. A stated average response time. A warranty length with terms. A named technician with years of experience. A dated job result with a before and after. A price range with what is included. An insurance limit in dollars.
• Column two: bare claims with nothing behind them. Fast. Reliable. Family owned. Trusted. Award winning with no award named. Best in the area. These are the competitor's soft spots, and they matter as much as column one.
• Column three: claims you cannot legally or operationally match. A trade qualification you do not hold. A 24/7 crew when you run business hours. A manufacturer certification that takes two years to earn. Write these down honestly and leave them alone.

Expect roughly 8 to 20 entries per competitor page. Column three is not failure. It is scope. If a competitor genuinely runs three vans overnight and you do not, no amount of content strategy fixes that, and pretending otherwise creates a claim your operations will break in front of a customer.

Enterprise teams run the same inventory, just wider. A 40-location brand should sample four locations across different competitive intensities, because the proof set that wins in a dense metro is usually heavier than the one that wins in a regional town.

Score your own page and count the proof gap

Now do the exact same pass on your page using the exact same three columns. No mercy on this read. If a sentence says "we pride ourselves on rapid response," it goes in column two, not column one.

The output is a number. Not a keyword gap count, a proof gap count: how many column-one items exist across the competitor set that your page does not have. Under 3 and you are competitive on evidence. Between 4 and 8 is the typical gap for a page stuck on impressions with weak clicks. Over 10 and the page is essentially a brochure.

Two scores worth tracking alongside it:

1. Above-fold proof density. How many verifiable facts appear before a buyer scrolls. Most losing pages have zero. The winner usually has two or three.
2. Claim-to-evidence ratio. Total claims divided by claims with evidence. A ratio of 6:1 means you are writing adjectives. Aim for 2:1 or better.

This scoring pass takes about 40 minutes per page and produces something a keyword tool cannot: a list of specific, ownable facts you are missing. Run it inside the same system you use to track page performance so the gap count sits next to the impression and click data rather than in a spreadsheet nobody reopens.

Convert each gap into a brief, never a copy

This is where most competitive work goes wrong. Finding a competitor's warranty statement and writing your own version of the same paragraph produces a second-rate duplicate. Every missing proof item should become one of three things instead.

An evidence-collection task

The competitor lists average response time. You do not have one because nobody has ever measured it. The brief is not "write about response time." The brief is: pull the last 90 days of job dispatch records, calculate the average, and publish the real figure with the sample size. If the number is worse than the competitor's, publish a different metric you actually win on. Same pattern for licence numbers, insurance limits, dated case results, named staff, and photographed work.

A buyer question the competitor left unanswered

Column two is your opening. If all three competitors say "no hidden fees" and none of them explain what triggers an after-hours surcharge, that unanswered question is worth more than matching their warranty. You are not competing on their evidence. You are adding evidence to a question they ignored.

An internal link to proof you already hold

Often the evidence exists somewhere on your site already. A completed project write-up, a certifications page, a staff bio. The brief is a link and a one-line pull quote on the money page, not a new article. This is the cheapest fix in the entire workflow and it is almost always available.

The rule holds across the whole process: the output is a list of proof the business must earn, not paragraphs to imitate.

The proof audit checklist

Run this against any page carrying real commercial weight.

• Credential: is the licence, registration, or certification number printed as a number, not described as a category?
• Time: is there a response, turnaround, or completion commitment stated in specific units?
• Guarantee: is the warranty length and what it actually covers on the page, not behind a PDF?
• People: is at least one person named with a role and a tenure?
• Result: is there one dated job outcome with a location, a problem, and what was done?
• Money: is there a price range, a call-out fee, or a clear statement of what is included?
• Cover: is the public liability or insurance figure stated in dollars?
• Volume: is there a job count, years-in-trade figure, or service-area count with a real number?
• Proximity: do at least two of the above appear above the fold?
• Machine readability: are these facts mirrored in structured data, not only in prose?

Ten checks. A page scoring 7 or better on a competitive local query is usually the page that gets chosen.

What to fix first

Order matters more than volume here.

1. Start with your weakest-CTR high-impression page. Not your homepage. The page already being seen and skipped has the fastest payback because the demand is already arriving.
2. Add the proof you already own before commissioning anything new. Licence numbers, insurance figures, years in trade, staff names. These take an afternoon and require no new data collection.
3. Move two facts above the fold. Most pages bury their strongest evidence in an about section nobody reaches.
4. Commission the one measurement you are missing. Usually response time or completion time. Pick one, measure it properly, publish it with the sample size.
5. Answer the question all three competitors dodged. Column two tells you what it is.

Do not attempt all ten checklist items at once. Three additions to one page, measured properly, teaches you more than thirty additions across ten pages measured never.

Measure the additions, then keep the inventory alive

Set a baseline before you touch anything: impressions, clicks, click-through rate, and conversions for that page over the prior 30 days. Make your three proof additions in a single dated change so the cause is unambiguous. Then read the result at 30 and 60 days.

Click-through rate usually moves first, because proof additions change what appears in snippets and answer summaries. Conversion rate moves second and moves harder, because a buyer who arrives and finds a licence number and a named technician converts at a rate an adjective never produces. If CTR lifts but conversions stay flat, your proof is attracting attention without resolving the buyer's actual hesitation, which means you added the wrong fact and column two is telling you which one to try next.

Competitors update their pages. A proof inventory built once and never refreshed decays within two quarters. Re-run the three-column pass every 90 days on your top five money queries, and treat any new column-one item a competitor adds as an immediate evidence-collection task rather than a content task. For enterprise teams, this becomes a proof register maintained centrally with location-level fields, so a new certification earned in one market propagates to every page entitled to claim it.

This is the part SEOGOD is built to hold: the page-level evidence state, the gap count, the dated changes, and the performance read sitting in one place rather than scattered across spreadsheets and someone's memory. If you want to see the proof gap on your own highest-impression pages before committing to the work, start with a free audit and score the first page yourself.

Your competitor is not beating you because they found a keyword. They are beating you because they wrote down a number and you wrote down an adjective.