AIOuditJournal
← All reports

Search Intent

Rankings went up but sales didn't? Look at search intent first

The average position line in Search Console has been climbing for weeks, and the order notifications are as quiet as ever. Before you change the page again, there's something to check: what the people searching those queries came to do. Today we'll pick one page, pull its queries, and sort them by intent together.

November 12, 2026CHROMETI Research7 min read

A ranking only means something once you know "for which query"

Say you sell a camping chair. The product page's average position moved from 14 to 6. Good news. But what if the climb came from "how to fold a camping chair"? Those are people who already own a chair and want to fold it. A "Buy now" button is not what they need.

One position number blends many queries. A rising number means the page got more visible somewhere. Whether that somewhere is where buyers stand, you don't know yet.

Same product, different search intent Example queries

Learning"how to choose a camping chair"Hasn't decided what to buy. Wants criteria and differences.
Comparing"best light camping chair under 1kg"Has conditions now. Narrowing down candidates.
Ready to buy"[brand] camping chair best price"Nearly decided. Just needs price, shipping and stock.
Already bought"how to fold a camping chair"Already owns one. Looking for how to use or fix it.
Example queries, not real search data. Depending on which box the climbing queries sit in, the same ranking gain is near to or far from a sale.

Google reads intent first, too

Google's "How Search Works" explains that before returning results, it first establishes what you're looking for, the intent behind your query. Use words like "cooking" or "pictures" and it figures recipes or images may fit best. Search "pizza" and you get nearby businesses that deliver. It also says what counts most changes with the nature of the query. For current news, freshness plays a bigger role than it does for dictionary definitions.

Carry that over to a product page. When someone asks "how to choose," Google tries to show pages that explain how to choose well. So if a product page grows long sections of tips or comparisons, it can start ranking for learning queries before buying queries. The page didn't get worse. It started answering a different question.

The climbing query vs. the question the page answers Example copy

Intent mismatch Climbing query: "how to choose a camping chair" First screen: price, options, buy button The choosing criteria are buried in a long section below Visitors looking for criteria move on to another article.
Intent match Climbing query: "light camping chair under 1kg" First screen: 0.9kg, folded size Right below: price, delivery date, exchange terms People who arrive with conditions finish checking on the first screen.
Both pages climbed. The difference is whether the visitor's question and the question the first screen answers are the same.

Pull the intent out of Search Console

You don't have to guess. It's already in the Search Console Performance report. The "Common tasks" help page says to add a URL filter to focus on the page you care about, compare the time frame with a similar earlier one, and sort by the Difference column to see what changed.

The same page has another line worth noting. If a page is seen but has a low click-through rate, users may see it and not think it answers their query. In that case it suggests updating the title and description to better represent the content, or adjusting the content to better match the queries that show the page.

Open a tab and give this three minutes. In the Performance report, add a URL filter for one main product page and compare the last three months with the period before. Sort the Queries tab by Difference and copy the five queries that climbed most into a notes app.

Four steps to sort a ranking gain by intent

  1. 01 · FilterJust one pageAdd a URL filter to the Performance report so only that product page's numbers remain.
  2. 02 · CompareQueries that climbedCompare periods, sort by Difference, and copy the biggest climbers.
  3. 03 · SortName the intentLabel each query: learning, comparing, ready to buy, or already bought.
  4. 04 · Line upNext to your order recordsPut it beside that product's orders for the same period. If "ready to buy" is empty, you can see why rank and sales move separately.
If step 03 has no "ready to buy" queries, it isn't a sales problem. It's an intent problem. The page is answering people who came to learn before people who came to buy.

Once sorted, the work splits. If most visitors came to learn, it's better to write a separate guide for them and link from there to the product page. If buying queries are there but orders aren't, it's time to check whether the price, delivery and exchange terms those buyers wanted are where they look first.

From search to order, what decides each step

  1. 01ImpressionThe page shows up in resultsDecided byHow relevant the page is to the query
  2. 02Click"My answer is probably here"Decided byWhether the title and description fit that intent
  3. 03ReadingDoes the first screen settle the question?Decided byThe visitor's question and the first screen
  4. 04Order"I'll take this one"Decided byWhether they came to buy, plus price and terms
Ranking works on steps 01 and 02. Step 04 leans heavily on whether the visitor came to buy in the first place. That's why intent comes before rank.

Google's "helpful, reliable, people-first content" page asks the same question. It suggests asking whether someone will leave feeling they've learned enough to help achieve their goal, or feeling they need to search again for better information. For a product page, that goal is usually "decide whether to buy this."

One more thing. Break average position down rather than taking it at face value. Search Console Help explains that when several results from the same site appear, only the topmost position is reported. At the property level, the number can look better than it does page by page. That's why you add a page filter.

A 3-minute check for today

Put the five queries you copied next to the product page.

  1. 1 min · SortOne intent word per queryLearning, comparing, ready to buy, or already bought.
  2. 1 min · PickOne "ready to buy" queryNone? That's today's answer. If there are some, circle the highest-ranked one.
  3. 1 min · CheckDoes the first screen answer it?Open the page on your phone. Without scrolling, can you see the price, delivery and the condition in that query?

Where doing this by hand gets stuck

You can do this for one page. With a hundred products, you'd filter every page, copy queries, label intent and open each first screen. And some customers now ask ChatGPT or Gemini the same question instead of a search box. How your page reads in those answers doesn't show up in Search Console.

AIOudit does that repetition for you. Paste a page URL and it looks at how customers and AI read that page in one go, then shows what to fix first, in order.

Dit with a white body and a horizontal scan bar across its visor, reading something
DIT "Years of building your site. For the price of a coffee, Dit shows you how AI reads it."
Dit, the robot that reads signals. Dit checks first whether the signal you sent reached AI.

AIOudit can't see things outside the page, though, such as price competitiveness, stock, ad budgets, seasonal demand or brand reputation. And which page shows up for which query is ultimately up to Google and each AI.

See how customers and AI read your page. Run it through AIOudit.

Sources

  1. Google · How Search Works: Ranking results (establishing the intent behind a query first; the "cooking", "pictures" and "pizza" examples; what counts most changing by query; freshness) · checked 2026-11
  2. Search Console Help · Performance report (Search results): Common tasks and use cases (URL filter, comparing periods and sorting by Difference, what low CTR can mean and what to do) · checked 2026-11
  3. Search Console Help · Performance report (Search results): About the data (only the topmost position is reported; property vs. page level) · checked 2026-11
  4. Google Search Central · Creating helpful, reliable, people-first content (self-assessment questions on learning enough and needing to search again; last updated 2026-10-05)