Product-led SEO: how to plan content from your product, not from traffic data
How a new or low-traffic site can decide what to write by starting from the product itself, with the checks that keep topics relevant and non-overlapping.
What product-led SEO means
Product-led SEO starts from the thing you sell. You decide which problems your product solves, who has those problems, and which alternatives people already compare. Content topics come from that list, not from a report of what the site already ranks for.
More
The opposite approach is query-led: you export Search Console data and optimise pages that already get impressions. That works for an established site. It gives a new site nothing, because there are no impressions to read.
The product-led approach has one honest weakness. Without traffic data you are guessing at demand. Everything below is about making those guesses cheap, checkable and easy to reverse.
Step 1: write down a product profile
A product profile is a short, editable record of what you sell. It should be based on your own public pages, not on what you wish the product did. Keep it to facts you could point to on the site.
More
| Field | What goes in it | Why it matters |
|---|---|---|
| Seed topics | Three to ten plain-language phrases for what the product does | Starting points for discovery |
| Audience | Who uses it and in what role | Decides tone and which problems count |
| Problems | What people are trying to fix when they search | Source of how-to and troubleshooting topics |
| Rivals | Products people compare yours with | Source of alternative and comparison topics |
| Call to action | One honest next step, such as a trial or a demo | Keeps articles tied to the product |
Review the profile by hand. A wrong rival or an invented feature will flow into every topic that follows.
Step 2: discover topics without paid data
Search engines suggest completions as people type. Those suggestions are public signals of real phrasing. Take each seed topic and expand it three ways: add question words in front, add common prefixes, and add common suffixes such as 'for small teams' or 'pricing'.
More
- Take one seed topic.
- Collect the autocomplete suggestions for the bare phrase.
- Repeat with question words (how, what, why, can) placed before it.
- Repeat with short suffixes and prefixes.
- Merge duplicates and drop anything unrelated to the product.
Step 3: check each topic against what you already have
Two pages that answer the same question compete with each other. Before a topic goes on the plan, compare it with your existing posts, planned items and indexed pages. If a close match exists, either refresh the old page or drop the new topic.
More
- Compare titles and the main phrase, not only exact keywords.
- A one-word variant is usually the same topic.
- If the overlap is partial, make the new page clearly narrower and say so in its outline.
- Write the decision down, so the next person does not re-add the topic.
Step 4: plan a month, then review it
Pick a small batch, not a year. A month of topics is enough to learn whether the approach produces pages people read. Spread them across themes so you are not writing four articles on one phrase, and keep a mix of buying-intent topics (alternatives, comparisons, use cases) with a smaller share of general how-to guides.
More
After each article is live, look at what it does in Search Console. That is the point where the plan becomes query-led, with real data to adjust against.
Common mistakes, and how to judge the first month
- Writing about the category instead of the product. An article that never connects to what you sell brings readers who have no reason to stay.
- Planning only comparison pages. People who are not yet shopping will never reach them, and thin comparisons read as sales copy.
- Treating autocomplete order as demand. It is a convenient list, not a measurement.
- Adding many topics at once and drafting them all. Start with a few, read the results, then widen.
- Making claims the product cannot support. A comparison page that is wrong about a rival or about your own product costs trust.
More
Search results take time, so do not judge a page a week after it is live. Check whether pages are indexed, whether they begin to collect impressions in Search Console, and whether the queries match the intent you planned for. A page that is indexed but shows impressions for unrelated queries needs a clearer focus.
Keep a simple log with the topic, the date it went live, the page address and one line on what you expected. When you review, compare expectation with what happened. That log is the start of the data you did not have on day one.
Where BlogFactory fits
BlogFactory does the mechanical parts and leaves the decisions to a person. Given a public product URL it builds an editable product profile with seed topics, rivals, audience, problems and a suggested call to action. It then discovers topics through Google autocomplete with question, prefix and suffix expansion, and flags topics that overlap content the site already has.
More
The 'Fill my month' action can add up to 12 topics to the plan. It only adds plan items; nothing is drafted until you start a draft. Drafts stop at a CMS draft and a person approves the exact version before delivery. Agents never publish live.
- Needs: a public product URL. No Search Console data and no keys for the profile and autocomplete discovery.
- Optional: an operator DataForSEO key adds search volume and ad competition to topic discovery.
- Cannot do: predict traffic, guarantee rankings, or replace your judgement about which rivals and problems are real.
Try it in BlogFactory
- Product-led SEO
- Keyword research
- Topical map
- Plan content from a product URL in the documentation
- Brave, DataForSEO and OpenRouter keys in the documentation
- Next guide: How to build a topical map from a list of keywords
Questions
Is product-led SEO only for new sites?
No, but it helps most there. An established site can use Search Console data. A new site has none, so a profile of the product is the only reliable starting point.
How many seed topics do I need?
Three to ten is enough to start. More seeds produce more noise, and each one still needs a manual relevance check.
Does autocomplete show search volume?
No. It shows that people type a phrase, not how often. Volume needs a keyword data provider, and the numbers are estimates.
Should every topic be a buying-intent topic?
Not necessarily. A balance works better: comparison and alternatives pages for people close to a decision, and some how-to content for people earlier in the process.
When do I switch to a query-led approach?
Once pages have been indexed long enough to collect Search Console impressions. Use that data to refresh winners and fill the gaps around them.
Plan it, draft it, approve it.
BlogFactory does the research and drafting steps in this guide. You approve every draft, and nothing publishes itself.