13 actionable tips to increase organic traffic, starting with pages you already have

The top-query column holds the diagnosis. The reviews post appears for people hoping to delete reviews strangers wrote, which Google does not let a business do, so a removal promise wins the glance and loses the reader. The checklist post appeared for a technical checklist and offered a general one.

Source: Search Console for pixelcrayons.com, web search, read through Ahrefs on 10 October 2026. The decisions are this month’s; we have not yet measured what they change, and we will not report a result until the comparison period has run.

The content-refresh method we run

1. Export one full quarter. One row per URL, web results only, so image and news traffic do not blur it.
2. Flag seen-but-not-chosen pages. Thousands of appearances, a handful of visits, sitting somewhere between the bottom of page one and page three. Google already half trusts them.
3. Read what people typed. Open the query list for each flagged URL. When the wording people used asks for something the article never delivers, rewriting the title will not help.
4. Decide one of four. Refresh (same subject, better answer), merge (two pages on one question), rewrite (the subject is right, the post is not), or leave alone (the traffic is not worth having).
5. File a baseline before touching it. A year of the URL’s figures, saved where whoever checks the result next can find them.
6. Change, then wait. Judge it on an equal stretch of time afterwards. For a merge, watch the combined queries of both old pages for four weeks, and roll back if they fall by 30 per cent.

Step 4 is where most refresh programmes go wrong: everything gets “refreshed”, including pages that should be merged or left alone. Deciding that before anyone writes is what a content strategy is for.

Thirteen tips, each with its test

A tip nobody can verify is hard to defend when budgets are set, so each one carries the test we would apply three months on.

1. Write for the buyer’s problem, not your keyword: later, the page’s query list should read like the questions in its headings.
2. Group posts into hubs on one subject, linked both ways: posts inside a hub should collect impressions sooner than strays did.
3. Publish on a schedule you can keep: no quarter in the archive should be empty.
4. Put a two-sentence answer under each question heading: that sentence should start turning up as your snippet.
5. Say what the page is about in the title, the opening and one heading: fewer of its queries should be the shown-but-never-clicked kind.
6. Cover what a competitor answers and you skip: look for impressions on queries that had none before.
7. Answer the questions Google already lists for your topic: those questions should join the page’s query list.
8. Rewrite titles and descriptions on seen-but-not-chosen pages: click-through should rise while average position holds still.
9. Fix slow templates where money is made: those URLs should move to “good” in the Core Web Vitals report.
10. Clear whatever stops crawling and indexing: the pages you care about should sit in the indexed column.
11. Share new pages where your buyers already read: branded searches tend to follow a good share within weeks.
12. Earn a few links from sites your buyers would recognise: new referring sites should be ones you would name in a pitch.
13. Refresh and merge before you write new: re-run the table at the top of this post.

For tip 7, the list is already on the results page: search your main topic, expand the questions box and copy each one down. Any your page leaves unanswered becomes a section.

Tip 9 has its own guide, what actually moves Core Web Vitals, and tip 10 is the first layer of a proper SEO audit.

When a refresh will not help

• The searcher wanted something else. A list of tools shown to people looking for a method will not win the click with a new date; it needs a different post.
• Two of your URLs chase one question. Polishing both leaves them fighting each other; combine them.
• The audience is wrong. Thousands of glances from people who will never buy are not worth chasing, so leave it alone or retire it.
• Nobody saved the before figures. Then nobody can say whether the work paid.