A Worse Page Outranks You. What Should You Actually Do?

A Worse Page Outranks You. What Should You Actually Do?

Start by checking whether it is worse for the searcher rather than worse in your judgement. Then look at intent match, site level signals, and links before you touch the content. In most cases we look at, the winning page is not better written. It is better matched.

This is one of the most common questions we get after a site revival. The new site is faster, the writing is sharper, the design is far better, and a page that looks thrown together in 2019 still sits above it. It is genuinely infuriating, and it is also usually explainable.

Here is the order we work through it, and the parts that are worth your time.

Why Does Google Rank a Page You Think Is Worse?

Because Google is not ranking writing quality. It ranks predicted usefulness for a specific query, using many systems at once. Google says plainly that "relevancy is determined by hundreds of factors, which could include information such as the user's location, language, and device." Prose quality is one input among many.

Google publishes a guide to its named ranking systems, and reading it changes how this question feels. There is a passage ranking system to "identify individual sections or 'passages' of a web page", a neural matching system to "understand representations of concepts in queries and pages and match them to one another", and link analysis systems and PageRank to "understand how pages link to each other as a way to determine what pages are about."

None of those systems reads your page and forms an aesthetic opinion. They ask narrower questions. A scrappy page that answers the exact question in the first paragraph can beat a beautiful page that takes four hundred words to get there.

Is It Actually Worse, or Just Uglier?

Separate craft from usefulness before you do anything else. Open the competing page and ask what a person typing that query wanted. If the answer is a price, a comparison, or a yes or no, and that page gives it in ten seconds while yours gives a narrative, theirs is the better page for that query.

We do this with clients and it is uncomfortable every time. A page we would never ship is often doing one thing very efficiently. Design quality and conversion quality are our job. Query satisfaction is a different measurement, and pretending they are the same is how teams stay stuck.

There is a fair version of the complaint too. Sometimes the competing page really is thin, wrong, or out of date. That happens. It is just not the first thing to assume, and assuming it first costs you the diagnosis.

Have You Matched the Intent, or Just the Keyword?

Intent mismatch is the single most common cause we find. Your page targets the words in the query while the ranking pages target the job behind it. If the results page is full of comparisons and yours is a feature explainer, you are in the wrong format, not the wrong depth.

The quickest diagnostic takes five minutes. Search the query, look at the top ten, and write down what format each one is. Tool, list, comparison, definition, template, calculator. If eight of ten share a format and yours is the tenth, that is your answer and no amount of rewriting within your current format fixes it.

Format is a decision you can actually change. We covered the underlying thinking in our piece on search intent, and it is the part of SEO where effort converts most reliably into position.

Could This Be a Site Level Problem Instead?

Often, yes. A strong page on a weak site loses to a weak page on a strong site more often than anyone likes. Google's link analysis systems and PageRank exist precisely to weigh how the web references pages, and a site nobody references starts behind regardless of what it publishes.

Two other named systems matter here. The site diversity system "generally won't show more than two web page listings from the same site in top results", which can cap you if you are competing against yourself. The exact match domain system exists to "ensure we don't give too much credit for content hosted under domains designed to exactly match particular queries", which sometimes explains why an obvious looking domain is not winning.

If the competing site has ten years of references and yours has ten months, the gap is not a content gap. It is a reputation gap, and you close it with coverage and citations rather than with another draft. We argued the link case in do backlinks still matter.

Is the Page Even Eligible to Rank?

Check the boring technical things before the interesting strategic ones. A page that is not indexed cannot outrank anything, and a page that is indexed but buried may be competing with another page on your own site for the same query. Both are quick to confirm in Google Search Console.

Use the URL Inspection tool on your page and read what it says rather than what you expect. Then open the performance report, filter to the query, and see which of your URLs Google is showing. It is surprisingly common to find Google has chosen a page you did not intend to compete with.

We walk through that whole triage in our piece on being indexed but not ranking. Do it before you commission a rewrite, because a rewrite of the wrong URL changes nothing.

How Long Should You Wait Before Judging?

Longer than feels reasonable. Google says "In 2023, we ran over 700,000 experiments that resulted in more than 4,000 improvements to Search", broken down as 4,781 launches, 16,871 live traffic experiments, 719,326 search quality tests, and 124,942 side by side experiments. The results you are watching are moving underneath you constantly.

That volume has a practical consequence. A position that changed this week may be an experiment, a personalisation effect, or noise. Reacting to a single week of movement is how teams end up rewriting pages that were fine and breaking pages that were working.

We would look at a rolling twenty eight day average in Search Console and nothing shorter. Rank trackers make this harder, not easier, because everyone sees a different result page now. We dug into that in our piece on rank tracking with personalised results.

What Should You Change First?

Change the opening, then the format, then the evidence, in that order. The opening is cheapest and has the largest effect on whether a passage can be lifted. Put the direct answer in the first forty to sixty words under the heading and stop making people scroll for it.

Format comes second because it is the expensive one. Turning an explainer into a comparison means new research and a new structure, so you only do it once the intent diagnosis says you must. Do not do it on a hunch.

Evidence comes third and is the one most teams skip. Named tools, real figures, a source and a year. Google's original content systems exist to "help ensure we are showing original content prominently in search results", and original means you had something to add, not that you paraphrased differently.

When Should You Give Up on the Query?

When the result page is owned by a format you cannot credibly produce. If the top ten are all free tools and you have no tool, or all community threads and you have no community, you are not one draft away. You are in the wrong competition.

The second case is when the query is not worth winning. Plenty of high volume terms bring traffic that never buys. We would rather rank first for a query that two hundred qualified buyers type than fifth for one that twenty thousand students type.

Retreating well is a real skill. Point your effort at the adjacent queries where your actual expertise is the advantage, and let the unwinnable one go. Coverage across a subject is what compounds, and one lost query rarely costs you the subject.

Does Any of This Change for AI Answers?

The diagnosis is the same and the stakes are slightly different. Answer engines lift passages, so a page whose answer sits in the first paragraph under a clear heading is easier to quote than one that buries it. The fix for being outranked and the fix for being uncited are largely the same fix.

What does change is the value of being wrong in public. When a retrieval system quotes you, an out of date figure travels further than it used to. That raises the case for dating your claims and naming your sources, which is good practice anyway.

It also raises the cost of the thin page that happens to rank. If a competitor's weak page is being quoted as the answer, correcting the record with a better sourced page is worth more than the position alone suggests.

What Does a Sensible Response Plan Look Like?

Pick one query, not twenty. Run the intent check, the eligibility check, and the reputation check, and write down which one failed. Then make exactly one change and wait a month. Teams that change five things at once never learn which one worked.

Keep a short note of what you changed and when. Google is running hundreds of thousands of experiments a year, so your own dated record is the only way to tell your change from theirs. A spreadsheet with three columns beats a dashboard here.

If you are staring at a competitor page that makes no sense to you and you want a second opinion on whether it is beatable, we are happy to look. Find us at phoenix.studio and send the query, not the page.