Content Marketing · September 23, 2026 · 8 min read
A Year of Blog Topics Without a Keyword Tool
Your sales calls, support tickets and invoices already hold a year of blog topics — here is how to pull 50 candidates out for free and cut them to 30.
By FluxWriter Team
Blog topics are the one input a keyword tool is worst at supplying, and paying roughly $1,200 a year for one before you have published 20 posts is a mistake most owners make early. The tools size demand you already suspect exists — they cannot tell you what your buyers say out loud. This covers where a year of topics comes from, how to rank them, and when a paid tool earns its fee.
Why Starting With a Keyword Tool Stalls
The standard first move is to buy a subscription, export a few thousand keywords and sort the column by search volume. What comes back is a spreadsheet, not a plan.
Volume describes the whole internet. It does not describe your customers. A term with 2,400 monthly searches looks better than one with 40 — until the 40 is your exact service in your exact city, and the 2,400 turns out to be students writing coursework. Only one of those two topics closes a deal.
The second problem is the difficulty score sitting in the next column. Difficulty is calculated against every site competing for that term, including the ones with 15 years of history and 5,000 referring domains. A new site that filters for high volume and ignores that number will spend a year producing pages that never reach page two. That is not a writing problem. It is a selection problem.
None of this makes the tools bad. It makes them mispriced for your stage — they answer a sizing question you cannot act on until enough pages are live to show which topics brought money in.
The Five Sources You Already Own
Every business generates topic supply as a byproduct of trading, and almost none of it gets written down. The questions arrive daily. They leave in the same conversation.
Start with the phone. Sales and enquiry calls produce the exact phrasing people use before they trust you, including the objections you answer so often you have stopped hearing yourself say them. Support email and tickets produce the other half, the problems that surface after purchase.
Then look at your own records. Quotes, invoices and order lines tell you what people buy rather than what they browse, and the line items that repeat are topics with a proven wallet behind them.
Here is what each source realistically yields over twelve months:
| Source | What it gives you | Topics per year |
|---|---|---|
| Sales and enquiry calls | Objection language, pricing questions | 12–20 |
| Support tickets and email | Post-purchase problems worth solving | 10–15 |
| Quotes, invoices, order lines | The services people pay for, not browse | 6–10 |
| Free search surfaces | Phrasing variants and adjacent questions | 10–20 |
| Gaps you spotted in competitors | Questions nobody answered properly | 4–8 |
Take the middle of each range and the total lands in the mid-50s — more than double what a twice-monthly schedule needs, drawn entirely from records you already keep.
The hard part is capture, not creativity. Keep one running list in whatever you open every day, and add the question the moment it is asked rather than at the end of the week. One rule keeps the list honest: if two different customers ask the same thing, it becomes a post.
What a Year of Topics Actually Requires
Count before you collect. Twice a month is 24 posts. Weekly is 52, and owners running the business at the same time rarely hold that pace past the first few months.
Build the list at roughly 1.5 times what you plan to publish, because a quarter to a third of any list dies on contact — the question turns out too thin to carry 1,200 words, or two entries collapse into one page, or the honest answer is a paragraph on a service page.
Leave room as well. Four or five of the 24 slots should stay unassigned: a price change, a rule change in your industry, the question that suddenly arrives six times in a fortnight.
So the arithmetic is 24 posts, 5 held back, 19 planned, and 30 candidates collected to fill them. With the sources above in front of you, that is one afternoon of work, not a quarterly research project.
Sorting Topics by Money, Not Volume
Once you have 30 candidates the ranking question replaces the finding question, and search volume is the wrong sort order for a small library. Sort by wallet distance instead.
Buying questions come from people comparing options with a budget in hand. Pricing, cost ranges, whether something is worth it, one approach against another. Conversion on that kind of page tends to run several times what a general explainer does, which is how a 30-visit-a-month post out-earns a 900-visit one. Confirm the multiple in your own analytics before betting a schedule on it.
Problem questions come from people who have the problem but have not decided to pay anyone yet. Symptoms, and the why-is-this-happening searches that come before any decision to hire. These carry most of the traffic and feed your list.
Curiosity questions are the trap. Broad and high in volume, satisfying to rank for. Tying one to revenue is close to impossible.
For the first year, weight the schedule hard toward the first two — roughly 6 buying and 13 problem posts out of the 19 you plan, which leaves the held-back slots for whatever the year throws up. Reverse that ratio and you get a blog that looks busy and converts nothing.
The Free Surfaces That Fill the Gaps
Some gaps still need outside input, and the free surfaces cover more ground than their reputation suggests. Type a service into the search box and read the suggestions that appear underneath. That list is built from real queries and takes about ten seconds per seed term.
Below the results themselves, the People Also Ask box and the related searches strip give you the questions Google associates with your topic. Expand two levels deep and a single seed routinely produces 15 to 25 phrasings — enough to shape both a post and its subheadings.
Search Console is the strongest free option, and it only works once you have published something. It holds 16 months of the queries that already earned your site impressions. Any query above roughly 50 impressions with no page dedicated to it is demand you have proven and not yet supplied. Sort by impressions descending. Read the first 100 rows.
Two other places pay off consistently. Reviews, yours and your competitors', carry the complaint in the customer's own words. So do the forums and subreddits your buyers actually use — usually sharper than the tidied-up phrasing a research tool hands back.
When a Paid Tool Earns Its Fee
Paid research platforms are good, and this is an argument about sequence rather than quality. At some point you will need them.
Entry plans at the major suites run roughly $100 to $140 a month, with mid tiers closer to $200 to $250. Lighter options in the $15 to $50 range cover volume lookups adequately for one site.
Three conditions mean it is time to pay. The first is scale — past roughly 40 published posts you are choosing between similar topics every week, and guessing gets expensive. The second is delegation. A writer or an agency needs a brief with numbers on it, because "my customers ask this a lot" does not survive being handed to someone who has never met them. The third is a genuinely competitive niche where you need to see who links to whom before committing three weeks of production.
Until one of those is true, that $1,200 a year buys 8 to 12 more posts instead.
FAQ
How do I know anyone is searching for these topics?
You do not, precisely, and no free method gives you an exact number. What you get instead is proof of demand from a different direction — a question asked by three paying customers has shown commercial interest that a 200-a-month volume figure never proves. Search Console shows you afterwards how often that question actually put you in front of someone.
Will I run out of topics after a few months?
No, and the opposite problem is far more common. Every topic splits into three once you write it: the original question, the objection to your answer, and the next question your answer creates. Owners who keep a capture list usually finish the year holding more candidates than they started with.
Is free keyword data good enough to compete?
For a site under 50 posts, yes. The binding constraint at that size is publishing capacity rather than research quality, since you cannot act on more topics than you can write. Paid data starts to matter when you have the output to chase 30 opportunities and must pick 10.
The Practical Takeaway
Open one document today and write down every question a customer asked you in the last 30 days, from calls, from email, from the quotes you sent out. Add the repeat line items from your last 20 invoices. Run each through autocomplete and People Also Ask to collect the phrasings, then sort the list into buying, problem and curiosity and cut the third bucket. Stop at 30 candidates, schedule 19 of them across the year, and hold the last 5 slots open for whatever arrives. Start with the five questions you answered on the phone this week.
If you are working a list like that at two posts a month, tools like FluxWriter can help turn approved topics into drafts on a fixed schedule — but deciding which customer question deserves a page, and which one is a two-line answer on your contact form, is a judgement only you can make.