Tip 11: Tinker Constantly

Tinkering works when you can tell what worked

Every store owner tinkers. A product title for a 200gsm merino hiking sock gets tweaked on a Tuesday, a collection intro gets rewritten after a slow month, someone adds an FAQ block to the sizing page. Six weeks later traffic is up, or down, and nobody can say which change did it. That’s fiddling.

Useful tinkering is a routine: small changes, made one at a time, on pages where you can measure the result, with a written record and an honest read of the numbers afterwards. It needs no testing platform and no data team. A spreadsheet and Search Console will do.

If you want the bigger picture of what Search Console reveals about your content strategy as a whole, we cover that in what Search Console tells you about content strategy. This tip is narrower. It’s the weekly and monthly habit of testing changes on individual store pages and reading the results without fooling yourself.

Start a change log before you change anything

The log is the whole method. Without it, every test becomes a memory contest.

Make a sheet with these columns:

• Date the change went live
• URL
• What changed, with the old and new text pasted in full
• Why you changed it, in one line
• What you expect to move (click-through rate, clicks, a specific query’s position)
• Review date, and the result once you have it

Log everything that touches a page, including changes you weren’t testing. A theme update, a price change, a product going out of stock, such as the medium size of a best-selling 250gsm merino crew neck, a new internal link from the homepage. Any of these can move a page’s numbers, and when you review a test you need to know what else happened in the window. The redirects and merges from Tip 10: Recycle Smartly belong in the same log.

One change per page, one page per question

Change one thing on a page, then leave that page alone until the review date. If you are testing a retitle on a 65-litre trekking rucksack, the description, photos and price stay exactly as they are. If you rewrite the title, add 200 words of intro copy and restructure the FAQ that afternoon, a rise in clicks tells you nothing you can reuse on the next page.

You can run several tests at once, as long as they’re on different pages and don’t interfere with each other. Retitling five collection pages in the same week is fine. Retitling a collection page while also adding prominent links to it from the blog is a muddle, because both changes could plausibly lift the same number.

Also leave out pages that are about to be affected by something bigger. A collection going into a seasonal sale, a product line being discontinued, a page in the middle of a site migration. Test on pages whose surroundings will stay quiet for the next month.

Where to test first: titles and meta descriptions on pages with impressions

Title tag and meta description tests are the best starting point for most stores. They’re quick to change, easy to reverse, and their main effect shows up in one clear metric: click-through rate from search results.

The catch is that you can only measure them on pages people already see. A page shown in search a few dozen times a month can’t tell you whether a new title works. So begin in Search Console’s Performance report, on the Pages tab, with a date range of the last three months. Sort by impressions. Your candidates are pages with plenty of impressions, an average position somewhere on page one or high on page two, and a CTR that looks low for that position compared with your other pages.

Collection and product pages often top that list, because their default titles are thin. Take an illustrative outdoor store whose base layer collection ships with a default title that says nothing about who the range is for or why to click. Check which queries the page appears for, then write a title that matches them more closely:

• Before: Base Layers | Store Name
• After: Merino Base Layers for Hiking and Skiing | Store Name

On Shopify you edit this in the search engine listing section of the collection or product. On WooCommerce it’s usually a field from your SEO plugin. Then request indexing in URL Inspection so Google picks the change up sooner, and note the date in your log.

Two things to know. Google sometimes rewrites title links in its results, so search for your page after it’s recrawled and check what’s actually showing. If Google is ignoring your new title, the test hasn’t really started. And meta descriptions affect clicks, not rankings. Google has said it doesn’t use them as a ranking signal, and it often generates its own snippet from page content anyway. Judge a description change on CTR alone.

A simple before and after window

You don’t need statistics software. You need a fair comparison.

1. Wait for the recrawl. URL Inspection shows the last crawl date. Your “after” period starts from when Google saw the change, not from when you made it.
2. Use equal periods with whole weeks. Compare the 28 days either side of the change, so each side has the same number of weekends. Store traffic swings by day of week, and uneven windows invent trends.
3. Filter to the page, then look at the queries. A page’s overall CTR can move because it started appearing for new queries, which has nothing to do with your title. Check whether CTR changed on the queries it already ranked for.
4. Compare CTR at similar positions. If average position moved from eight to five during the test, CTR will rise whatever the title says. A title test is only clean when position held roughly steady.
5. Check a control. Pick two or three similar pages you didn’t touch and see how they moved over the same dates. If they all rose by the same amount, the season or the market is the likelier cause. If you retitled the merino base layer collection, the fleece midlayer and softshell jacket collections make sensible controls.

For on-page changes, such as new collection intro copy or an FAQ section, allow longer. Rankings take time to settle after a content change, so six to eight weeks is a more sensible window, and the control pages matter even more. Changes to collection copy are worth testing because collection pages are so often under-written, but they’re slower to read.

What not to conclude from small numbers

Most store pages have modest traffic, and modest traffic is noisy. A few rules keep you honest.

Small click counts swing on their own. A product page for a pair of carbon trekking poles going from 40 clicks to 52 in a month is well within normal wobble. Treat a change as real when it’s large relative to the page’s usual variation, and when it holds for a second period. Look at its last six months to see how much it moves when nobody touches it.

A CTR jump on a quiet page is a hint. If a page was shown 600 times, moving from 1.8% to 2.6% is about five extra clicks. Note it, keep the new title if it reads better, and look again next month. Don’t roll the pattern out across 200 product pages on the strength of it.

Average position lies when the query mix changes. A page that starts ranking at position 40 for dozens of new long-tail queries can see its average position get worse while its useful rankings improve. Always check query-level data before calling a drop.

Coincidence is common. Google ships core updates several times a year, and seasonal demand for many products shifts sharply from month to month. If a test window overlaps either, say so in the log and weigh the result accordingly.

A null result is still a result. If a new title made no measurable difference after two clean windows, keep whichever version is clearer for shoppers and move on. Not every change needs to win. Some just need to stop being a question.

A monthly routine for a store

Put two hours in the calendar at the start of each month.

1. Review last month’s tests against their review dates. Write the result in the log, including “no clear change”.
2. Keep, revert or extend each test. Revert anything that clearly hurt. Extend anything that’s promising but thin on data.
3. Pull a fresh list of pages with high impressions and weak CTR for their position. Pick three to five for new title or description tests.
4. Pick one page for a slower on-page test, such as an intro rewrite, a new FAQ answer drawn from recent customer questions, or a clearer size guide section for a pair of leather walking boots.
5. Glance at the Pages report for anything that has dropped out of view entirely. That can be an indexing problem rather than a content one, and indexing and visibility are different problems with different fixes.
6. Log every change on the day it goes live.

After six months you’ll have something most stores lack: a record of what works on your own pages, for your own customers. That record is worth more than any list of best practices, including this one. It tells you which title formats earn clicks in your category, how long your pages take to respond, and which changes were never worth the effort.

This is also where the series comes full circle. Every earlier tip, from writing for people first to recycling what you have, is a hypothesis about what will help your store. Tinkering, done with a log and some patience, is how you find out which of them are true for you.

This is Tip 11 in Sprite’s 11-part series on ecommerce content that compounds. Start again at Tip 1: People First, Robots Second.

Frequently asked questions