Content decay or a conversion problem?

A page that used to bring in customers brings in fewer. The usual response is to "refresh the content": rewrite it, add a few sections, update the date. Sometimes that is exactly right.

Often it isn't. If the same number of people still arrive and fewer of them buy, no amount of search work will help, because search isn't what broke. And if the page still ranks in the same place but fewer people click it, the fix is the title in the results, not the body of the page.

This guide shows how to split a drop into its parts using Search Console and your analytics, how to read the patterns, and what to do for each. A worked example compares two pages that lost almost the same number of sales for completely different reasons.

Split the drop into its parts

What a page earns from search comes from four numbers multiplied together:

revenue = impressions × click-through rate × conversion rate × value per conversion

• Impressions: how many times the page was shown in results. From Search Console.
• Click-through rate (CTR): the share of those impressions that became clicks. From Search Console, alongside average position.
• Conversion rate: the share of visitors from search who bought, signed up or enquired. From your analytics.
• Value per conversion: what each of those is worth. From your analytics or sales records.

A fall in revenue has to come from at least one of these. Find which before changing anything.

Compare like with like

1. Pick two periods of the same length. In Search Console's performance report, open the date filter and use the Compare tab. If your business is seasonal, compare with the same period last year, not the months just before.
2. Filter to the page. Add a page filter so the chart and table show only that URL.
3. Look at its main searches. Switch to the Queries tab. A page's totals can hold steady while its most valuable search falls and a less valuable one rises.
4. Pull the other half from analytics. For the same two periods, get organic visits that landed on the page and the conversions they produced.

Reading the patterns

The CTR column needs care. CTR always falls when position falls, because lower results get fewer clicks. Using typical figures (they vary by results page), a result at #4 gets about 7.5% of clicks and one at #7 about 3.1%. So a page that drops from 4 to 7 and whose CTR goes from 7.5% to 3.1% has a ranking problem, not a title problem. A CTR problem is a CTR that falls while the position holds.

A worked example: two pages, two periods

Take a company selling time-tracking software. Both pages below lost trials this summer. The figures are illustrative and are monthly averages for June to August last year and June to August this year. Trial rate is trials from organic visits divided by clicks.

You can check each derived figure: 900 ÷ 12,000 = 7.5%; 360 ÷ 11,600 = 3.1%; 36 ÷ 900 = 4.0%; 14 ÷ 360 = 3.9%. For Page B, 1,040 ÷ 20,000 = 5.2%; 1,061 ÷ 20,400 = 5.2%; 52 ÷ 1,040 = 5.0%; 27 ÷ 1,061 = 2.5%.

On a revenue report the two pages look alike. Page A lost 22 trials a month (36 − 14) and Page B lost 25 (52 − 27). They need opposite fixes.

Page A: lost rankings

Impressions barely moved (down 400, about 3%), so people are still searching. The average position fell from 3.9 to 7.2. CTR fell from 7.5% to 3.1%, which is what you would expect at #4 and #7, so the title is doing its job for the position it now holds. The trial rate held at about 4%: the visitors who arrive still convert.

This is a search problem. Searching "time tracking for agencies" shows why: two competitors now rank above it with pages comparing tools side by side, with prices, and Page A hasn't been updated in two years and doesn't mention pricing at all.

If an update got it back to about #4: 11,600 × 7.5% = 870 clicks, and 870 × 4% = 34.8 trials, about 21 more a month than now (34.8 − 14 = 20.8). That is modelled, not promised.

Page B: a conversion problem

Position, impressions and CTR are all flat, and clicks are slightly up. Search is delivering the same visitors. The trial rate halved, from 5.0% to 2.5%.

This is a conversion problem, and rewriting the page for search would change nothing. The first check was tracking: sign-ups recorded in the product's own database from organic visitors fell by the same amount, so the drop is real, not a broken tag. The second check was what changed on the page. In May a redesign replaced the "Start free trial" button at the top with "Book a demo" and moved the trial link to the footer.

Restoring the trial rate to 5.0% would give 1,061 × 5% = 53 trials, 26 more a month than now (53 − 27).

Which first?

Page B. The fix is one button, it doesn't depend on Google, and the evidence for it is strong. Page A needs real content work, and the result will take weeks to show.

Had Page A held position 3.9 while its CTR fell from 7.5% to 3.1%, the diagnosis would be different again: same place, fewer clicks. That is a CTR problem, covered below.

Fixing a search problem

Lost rankings

1. Rule out technical causes first. Use URL Inspection in Search Console to confirm the page is still indexed. Check that a redesign didn't change its URL, add a noindex tag (an instruction telling Google not to index the page) or point its canonical tag (which tells Google the preferred version of a page) somewhere else.
2. Look at who now ranks above you for the main search. What do they cover that you don't? Are they more specific, more current, or a different kind of page, such as a comparison where yours is a product page?
3. Update the page for the searcher, not the calendar. Answer the main search clearly near the top, add the sections that are missing, and correct anything out of date. Google's guidance on helpful content is a good test of whether the update is substantial. Changing the date alone is not an update.
4. Link to it from your related pages, in sentences where the link makes sense.

If the content is already as good as the pages above and the page is commercial, the gap may be links. The page-level backlink gap guide covers how to measure it.

Lost clicks at the same position

1. Search the main query yourself. New ads, shopping results, an AI-generated answer or a video block can push results down the screen and take clicks at the same position. If the results page changed, a better title will recover only part of the loss.
2. Check what Google shows as your title. Google sometimes displays different text from your page title, and the description shown may come from the page itself rather than your meta description.
3. Rewrite the title and meta description. Lead with what the searcher wants, use the main search phrase naturally, and give a specific reason to click. Keep titles to about 60 characters and descriptions to about 155; longer ones are often cut off. Ship/Scale writes within those limits (60 and 155 characters) when it changes titles and meta descriptions.
4. Change the title on its own first, so you can tell whether it worked.

Demand fell

Compare with the same period last year. If the drop repeats every year, it is seasonal. If searches for the whole topic are falling, no page change will bring them back. Don't rewrite a page because its market went quiet.

Fixing a conversion problem

1. Check the tracking before the page. Compare conversions in your analytics with your own records (orders, sign-ups, enquiries). Consent banners, tag changes and a new checkout or form can all break tracking without anything else changing.
2. List what changed on or around the page: design, calls to action, price, stock, delivery costs, form fields, pop-ups, loading speed.
3. Check whether the searches changed. Compare the page's queries between the two periods in Search Console. A page that starts appearing for learning searches brings visitors who were never going to buy, which lowers the conversion rate without anything being broken.
4. Look at the offer next to the alternatives. If a competitor cut prices, your page can convert worse with nothing wrong on it.

These fixes belong with whoever owns the page's conversion, not with the SEO plan.

Where this goes wrong

• Comparing the wrong periods. A December versus November comparison can make a seasonal business look broken. Use the same period last year where the business is seasonal.
• Two problems at once. A page can lose rankings and conversion together. Split the drop: work out how many sales the lost clicks explain at the old conversion rate, and treat the rest as a conversion question.
• Page totals hide the searches. A steady average position can hide a buying search that fell and a learning search that rose. Look at the main searches one by one.
• Small numbers are noisy. A fall from 14 trials to 9 in one month may be chance. Look at a few months before acting.
• Site-wide drops aren't page problems. If many pages fell on the same date, look for a site-wide cause (a migration, a template change, a Google update) before rewriting any one of them.
• Search measurement stops at the click. Ship/Scale measures each content change against the 28 days before it at 7, 14, 30, 60 and 90 days, on position, clicks and impressions for the searches it targeted. That tells you whether the search side recovered; whether sales did is a question for your analytics. The guide to measuring SEO changes at 30, 60 and 90 days covers reading both.

The short version

Split a page's drop into impressions, CTR, conversion rate and value, comparing the same periods in Search Console and your analytics. Falling position means a search problem: update the content and check the technical basics. Stable position with falling clicks means a CTR problem: check the results page, then rewrite the title and meta description. Stable traffic with falling sales means a conversion problem: check the tracking first, then what changed on the page, and leave the content for search alone.