Tip 5: Cluster Up

A cluster is a set of different questions

Clustering is the most repeated advice in content SEO and one of the most misapplied. Stores hear “build topic clusters”, pick a subject they sell, and publish eight posts about it. Six of them turn out to say the same thing in slightly different words.

That’s a pile. A cluster is a group of pages where each page answers a different question a buyer has on the way to a purchase, and the pages link together and to the range that sells the answer. The difference sounds pedantic until you look at Search Console and find two of your own posts taking turns at position 14 for the same query.

This tip is the practical version: how to plan a cluster for a store, how to keep its pages from competing, and how to tell when a new page belongs in it at all. We’ve written separately about why ecommerce needs a content architecture rather than a hub-and-spoke diagram, and about why topical authority isn’t a score you can measure. Those explain the thinking. This is the method.

Start from the collection, work backwards

For a store, every cluster should have a commercial anchor: the collection page or product range the content ultimately serves. Choosing it first keeps the cluster honest. If you can’t name the collection a set of articles supports, you’re planning a magazine, which is a fine thing to run but a different project.

Take a fictional store selling cast iron cookware. Its best-margin range is skillets. That collection page targets the commercial query, something like “cast iron skillet”, and it should stay the page that ranks for it. Your blog shouldn’t try to take that query away from it.

Next, list what people need to know before they’re confident buying a skillet, and after they’ve bought one. Pull these from real sources rather than brainstorming:

• Search Console queries already landing on the collection and product pages.
• Customer service emails and live chat logs.
• Returns reasons and negative reviews (“too heavy”, “rusted after a week”).
• The “People also ask” questions for your main commercial query.
• Questions your staff answer at markets, trade shows or on the phone.

For the skillet store, that produces a list like this: seasoning a new pan, whether cast iron works on induction, cast iron versus carbon steel, how to fix rust, what size skillet a household of two needs, whether you can cook tomatoes in cast iron, and why enamelled cast iron costs more.

One question per page, one page per question

Here’s the rule that stops a cluster turning into a pile. Each page gets exactly one primary question, and no other page on the site targets it. Write the question down before the page exists, then check every new brief against the list.

The failure mode is subtle because the titles look different. “How to season a cast iron skillet”, “The complete guide to cast iron care” and “5 mistakes that ruin your cast iron pan” can all end up answering “how do I season and look after cast iron”. Google has to pick one to show. It may alternate between them, rank neither well, or pick the weaker one. Each page also splits the links and attention the topic would otherwise concentrate on a single strong URL.

Two pages targeting the same query compete with each other. It’s one of the most common problems we see on ecommerce blogs, and it’s easy to create by accident when content is planned title by title. Sprite’s own blog had it too, with several posts covering overlapping ground, and we’re consolidating those now. It happens to people who know better.

For the skillet store, the finished question map might look like this:

1. Collection page: the skillet range (the commercial query, with a short buying section at the foot).
2. Seasoning guide: how to season a new cast iron pan, including which oils work and oven temperatures.
3. Rust repair: how to remove rust and reseason, for the customer whose pan came out of the dishwasher.
4. Induction question: does cast iron work on induction hobs, with notes on warped bases and scratching glass.
5. Comparison: cast iron versus carbon steel, for the buyer still deciding between materials.
6. Sizing guide: what size skillet to buy, by household and hob size.
7. Acidic foods: can you cook tomatoes in cast iron, and for how long.
8. Enamelled explainer: why enamelled cast iron costs more and who it suits.

Notice that “cast iron care” as a broad topic doesn’t get its own page. Seasoning and rust repair are the questions people actually ask, and a vague care page would compete with both.

Checking for overlap before you publish

A quick three-step check catches most competition before it goes live.

1. Search your own site. Run a site search in Google for the core phrase of the planned article. If an existing URL already answers the question, improve that page instead.
2. Compare the results pages. Search the planned article’s target query and the nearest existing page’s query. If Google shows largely the same results for both, it treats them as one intent, and they need one page.
3. Write the answer in a sentence. If the first-paragraph answer of the new page would be the same sentence as an existing page’s, you’re about to publish a duplicate.

For pages already live, Search Console does the detective work. In the Performance report, filter to a query and switch to the Pages tab. If two or more of your URLs collect impressions for one query over a few months, that’s competition worth investigating. Our guide to Search Console and content strategy goes further into reading these patterns.

Fixing pages that already compete

When you find two pages fighting over one query, whether through Search Console or an ecommerce content audit, you have three options. Pick based on what each page is for.

• Merge and redirect when both pages answer the same question. Keep the stronger URL, fold the useful parts of the weaker one into it, and 301 redirect the weaker URL. This is the most common fix.
• Refocus when the pages drifted together but could answer different questions. Rewrite one around a narrower question (rust repair rather than general care) and update its title, headings and internal links to match.
• Retire a page with no traffic, no links and nothing worth keeping. A redirect to the nearest relevant page is usually kinder than a 404.

Blog posts that compete with a collection page deserve special attention. If a guide titled “The best cast iron skillets” outranks your skillet collection for “cast iron skillet”, the blog has taken a query the collection should own. Refocus the guide on a comparison question, and link it clearly to the skillet range. Why collection page SEO is usually broken covers the collection side of that fix.

How the pages connect

A cluster’s links follow the buyer’s path. Every supporting page links to the collection where it’s natural to shop, using an anchor that describes it (“shop cast iron skillets”, “our carbon steel pans”). Pages that sit side by side in a decision link sideways: the seasoning guide links to rust repair, and the induction page links to the sizing guide. The collection page links back to the two or three guides that most help someone choose.

The detail of anchors, placement and link audits belongs to Tip 8: Link the Dots. For clustering, the point is that links should reflect how the questions relate, which is easy once each page has a single, distinct job.

Sizing a cluster to the questions

A cluster should have as many pages as there are distinct questions, and no more. The skillet cluster above has seven supporting pages because buyers ask roughly seven different things. A narrow product like replacement skillet handles might justify two. Padding a cluster to hit a target count is exactly how overlapping pages get written.

Build one cluster properly before starting the next. Pick the collection with the best margin and the clearest search demand, map its questions, fix any competition among existing pages, then fill the gaps. When a new question turns up in customer service or Search Console, add it to the map, check it against the existing pages, and only then brief it. Each page still has to be good in its own right, which is the subject of Tip 6: Go Deep.

Keeping the map alive

Clusters drift. Ranges change, a new material launches, someone writes a seasonal gift guide that quietly repeats the sizing page. Keep the question map somewhere the whole team can see it, with one row per question and the URL that owns it. Before anything new is commissioned, it gets a row. If the row already has a URL, the work is a content refresh.

This is how we plan content at Sprite: every article is briefed against the questions a store’s pages already answer, so new work fills a gap instead of competing with something already published. Applying the same discipline to our own blog is what surfaced our overlapping posts in the first place.

This is Tip 5 in Sprite’s 11-part series on ecommerce content that compounds. Next: Tip 6: Go Deep.

Frequently asked questions