A website with almost no traffic still has a business behind it. That business has an offer, a process and questions it needs to answer before someone can buy or make contact.
Those are the first topic sources I would use. A lack of search data makes careful observation more important; it does not make a list of speculative keywords more reliable.
Begin with the missing explanation
Read the current product and service pages. Can a visitor understand what is offered, who it suits, where it is available and what happens next?
If those answers are missing, fix the main pages before building a large blog archive. An article cannot compensate for an unclear service description at the point of enquiry.
The business-facts audit provides the starting record. It helps prevent content planning from spreading an inaccurate offer more widely.
Collect questions from real conversations
Ask the owner, sales team or support contact which questions repeat. Preserve the customer's language while removing personal details.
Separate questions asked before purchase from those asked during preparation or use. They may need different pages and different closing actions.
The GOV.UK interview guidance offers useful principles for exploring experience without leading the answer. Even a small set of conversations can identify sensible topics, provided it is not presented as representative market research.
Choose questions you can answer distinctly
Prioritise a topic when it is relevant to the current offer, genuinely unresolved and supported by evidence or a useful example you can provide.
Do not publish a generic guide simply because competitors have one. Ask what the reader would gain from your version.
Google's people-first content guidance is a useful editorial reference. A clear intended audience and a substantive contribution matter more than a calendar filled with broad subjects.
Build a small connected set
Start with a few pages that support a real decision sequence: understand the offer, compare an important option, prepare for the next step and use the result.
Link them where the connection helps. Avoid producing several near-identical pages for slight wording variations of the same question.
As discussed in search intent as a customer decision, the page type should follow the task. Some of the first assignments may be updates rather than new articles.
Treat the first period as observation
Check accessibility, accuracy and whether the pages help in actual conversations. Use available search data as it develops, but do not invent a trend from a handful of visits.
Google's Search Console performance documentation explains what the report can show. A low-volume site may need a longer period before a pattern is interpretable.
The first content plan should make the business understandable and establish a useful learning record. Traffic is an important outcome to pursue, but it is not the only evidence that the initial work is worth doing.