Content strategy for a business that has never published

The sentence comes up in almost every first conversation about content: "We don't really have anything to write about." It is said with a kind of apologetic certainty, as though the business has been checked and found empty.

It is nearly always false. What is true is that the business has never written anything down. Those are different problems, and only one of them is hard.

Consider what happens in a week at a firm that fits windows, or does bookkeeping, or repairs commercial fridges. Someone rings and asks whether the job needs scaffolding. Someone else asks what happens if the part is discontinued. A third asks why the quote from down the road is half the price. Each of those is a question a stranger typed into a search box an hour earlier and got a bad answer to. The business answered it, out loud, for free, to one person, and then let it evaporate.

That is not a business with nothing to say. That is a business with no filing system.

Start by transcribing, not by inventing

The instinct when starting a blog is to sit down and think of topics. This is the worst possible first move, because thinking of topics puts you in the position of guessing what people want to know. You do not have to guess. You already have the answers; they are just stored in people's heads and in your sent folder.

So the first exercise is not creative. It is clerical.

For two weeks, keep a single document open — a notes app, a page in a notebook by the phone, anything. Every time a customer or enquirer asks something, write the question down in their words, not yours. Not "boiler service intervals" but "do I actually have to get it done every year or is that just to sell me something". The second version is the useful one, because it carries the doubt, the suspicion, and the phrasing of the person asking.

Add to that document from four other places:

1. Your sent email. Search for the phrases you type most often when explaining something. Every long reply you have written twice is a page you should only have to write once.
2. Quotes that went quiet. The questions that came just before a prospect stopped replying are worth more than the ones that came before they said yes. They mark the point where confidence broke.
3. Whoever answers the phone. If that is not you, they know the repeat questions better than you do. Ask them for the five things people always ask and the two things people always get wrong.
4. Complaints and refunds. Uncomfortable, but the misunderstanding behind a complaint is usually something you never explained anywhere public.

Two weeks of this in an averagely busy business produces more raw material than most people expect. The point is that none of it was invented. Every line came from someone who cared enough to ask.

Turning the raw list into twenty topics

Now the clerical work becomes editorial. Take the list and run it through four passes.

Pass one: merge duplicates. Ten variations of "how much does it cost" collapse into one topic, but keep all ten phrasings in a note underneath — they tell you the shape the answer needs to take.

Pass two: sort into three piles. Questions asked *before* someone knows they need you. Questions asked *while* they are choosing between suppliers. Questions asked *after* they have decided but before they commit. These map onto genuinely different pages and you will need all three.

Pass three: cut anything you cannot answer better than a generic article. If your answer to "what is a heat pump" is the same paragraph anyone could write, drop it. Your version of that topic is "what a heat pump costs to run in a solid-wall Victorian terrace" — narrower, and answerable only by someone who has done it.

Pass four: write the title as a promise. "Insulation options" is a category. "Which insulation is worth it if you are only staying five years" is a promise. The first attracts nobody in particular. The second is aimed at a person who exists.

If the list still comes up short of twenty after all four passes, add these three generators, which reliably produce topics for any business:

• The price question, answered properly. Almost nobody in your sector publishes real numbers or explains what moves them. A page that explains what makes a job cheap or expensive — with honest ranges and the reasons behind them — tends to be the most-read page on a small business site for years.
• The comparison you are asked to make. Whatever the alternative is — a cheaper competitor, doing nothing, doing it yourself — write the honest comparison including the cases where the alternative wins. This is uncomfortable and it is exactly why it works.
• The thing customers get wrong. Every trade has one. The homeowner who paints over damp. The shop owner who lists the same product forty times. Write the diagnosis and the fix.

None of this required a keyword tool. If you want to check demand afterwards, you can — there are free ways to do keyword research without a paid subscription — but do it as a sanity check on a list you already have, not as a substitute for having one.

Two jobs, two kinds of page

Content that attracts and content that converts are not the same work, and the most common failure is publishing only one kind.

| | Attracting content | Converting content | |---|---|---| | Reader's state | Doesn't know you exist; may not know they need you | Deciding between you and two others | | Typical topic | The problem, explained | Your process, prices, guarantees, edge cases | | Where it sits | Blog, guides, explainers | Service pages, pricing, FAQs, case detail | | How it is judged | Did anyone find it? | Did the people who found it get in touch? | | Failure mode | Traffic that never enquires | Nobody ever reaches the page |

A site with only attracting content collects readers who leave. A site with only converting content is a brochure nobody arrives at. You need enough of the first to bring people in and enough of the second that arriving means something.

The join between them matters more than either. A guide about damp that never mentions what a survey costs, how long it takes, or what happens next has done half a job. Not a sales paragraph bolted on — one honest line about what the reader would do next if they wanted this handled, placed where the question naturally arises.

Clusters, or why a pile of posts is not a strategy

Twenty unrelated posts perform like twenty small websites. Twenty posts arranged around a subject perform like one site that clearly covers that subject.

The structure is simple: one substantial page covering a subject broadly, and a set of narrower pages each handling one part of it in depth. The broad page links down to each of the narrow ones. Each narrow page links back up. Where two narrow pages genuinely relate, they link across.

An illustrative example — a two-van damp proofing firm in Leeds:

• Hub: *Damp in older Yorkshire housing: what it costs to fix and what it doesn't*
• Spokes: what rising damp actually is and how often it is misdiagnosed; condensation versus penetrating damp, told apart in five minutes; what a damp survey should include; why chemical injection is oversold; what to do if you are buying a house with a damp report

That is a business that looks, to a reader and to a search engine, like it knows about damp in Yorkshire housing. The same five posts scattered across unrelated topics do not.

Two practical rules. Do not build a cluster around a subject you have only five posts' worth of material for and no experience in — the shape is not the point, the substance is. And write the hub last if you can; you will understand the subject's real structure better after writing the parts.

The internal linking is not decoration. It is how a reader gets from a general question to the specific answer, and it is one of the few ranking-relevant things entirely within your control. If you want the broader mechanics, we have written plainly about how Google ranking actually works.

How much to publish

Less than you have been told, more consistently than you expect to manage.

The honest answer for a business where nobody's job title contains the word "marketing" is one solid piece a fortnight, or even one a month, held for two years. That beats eight posts in January and silence until the following autumn, every time.

Two reasons. First, momentum: search engines and readers both respond to sites that are visibly still alive, and abandoned blogs advertise abandonment. Second, and more practically, a schedule you can hold is a schedule you will still be holding when the results arrive — and they arrive later than anyone wants. Our summary of how long SEO actually takes based on what the evidence supports is not encouraging reading for the impatient, and it is the honest picture.

On length: Google's documentation states there is no preferred word count and no minimum. Write until the question is properly answered and then stop. A 600-word page that fully settles "do I need scaffolding for this" is worth more than 2,000 words circling it.

Why most business blogs fail

They almost never fail because the writing was bad. Three causes account for most of them.

They were published inconsistently. Six posts in the first month, two in the third, none after that. The pattern is so consistent it has a shape: enthusiasm, then a busy period, then the blog becomes the thing you feel guilty about rather than the thing you do. The fix is unglamorous — commit to a frequency you would still hit in your worst month, not your best.

They were written for nobody in particular. "Top 5 tips for choosing a builder" is aimed at an average of all possible readers, which is to say at no one. Content that works is usually written to a single person you can picture: the landlord with a tenant complaining about mould, the buyer whose survey came back with a red flag. Narrow feels risky. It is the opposite.

They were abandoned at month four. This is the cruellest one, because month four is roughly where nothing has happened yet but the work has stopped being novel. Anyone who has done this properly will tell you the curve is flat for an uncomfortably long time and then is not. Stopping at month four means paying the entire cost and collecting none of the return.

A fourth, less common but more expensive failure: publishing at volume purely to have pages, particularly bulk-generated text with no first-hand knowledge in it. Google's spam policies address scaled content abuse directly, and beyond the risk, it does not work for the simple reason that nothing in it is worth reading.

When to stop adding and start improving

There is a point — usually somewhere between twenty and forty published pages — where the best next post is not a post at all.

Work through these in order:

1. Is anything ranking on page two? Those pages are close. Improving one of them — better answer, more specifics, a section addressing what the current top results omit — is worth more than three new posts.
2. Is anything getting read but never leading anywhere? Find the page with visitors and no enquiries. It is usually missing the join described earlier: what to do next.
3. Is anything now wrong? Prices, regulations, product names. A page confidently stating something outdated does active harm.
4. Are two thin pages fighting each other? Merge them into one good page and redirect.
5. Only then, write something new.

Most sites carry a handful of pages that would double in value with an afternoon's work, while their owner writes post twenty-one. If nothing is moving at all, the problem may not be the content — a diagnostic checklist for why a site is not ranking will tell you faster than another post would.

The trade-off worth naming

Everything above takes eighteen months to two years to matter, and costs mostly time. That is the deal, and for a business that intends to exist in five years it is a good one — the pages you write keep working without further payment.

But it is a poor answer to "I need enquiries this quarter". Content does not do urgency. Paid channels do, at a cost per click that stops the day you stop paying, and paid placements — including a position on our own public rankings, where every figure is visible — buy visibility, not credibility. A listing puts your name in front of people. What happens when they click through depends entirely on whether the pages they land on answer their questions. We have been explicit about what paying for rank cannot do for you, and this is the largest item on that list.

If you have to choose one, and the business is not on fire, choose the writing. If you are choosing paid visibility, write anyway — otherwise you are paying to send strangers to a website that does not tell them anything.

Questions people ask before starting

We are a two-person business. Who actually writes this? Whoever knows the answers. Being able to explain something to a customer is the whole qualification. If writing is genuinely painful, talk the answer into a voice recorder as though a customer had rung, then transcribe and tidy it. The result reads better than most deliberately written business content.

Should we use AI to write it? As a drafting and editing assistant, working from your knowledge, it is fine. As a replacement for having knowledge, it produces exactly the interchangeable text nobody needed more of. The test: does this page contain something only someone who has done the work could have written? If not, it is not worth publishing regardless of what produced it. There is a longer treatment of what AI actually changed about search and what it did not, including what Google's documentation says about how content is produced.

How do we know whether it is working before it works? Watch for early signals rather than rankings: are people finding pages you did not promote, are enquiries mentioning things they read, does the sales conversation start further along than it used to. Those move before the numbers do.

Do we need to promote each post? The best promotion for most small businesses is telling the people who ask the question. Send the page to the next customer who asks it. Over time, some of those pages get linked to naturally, which is the only durable version of link acquisition — the mechanics are covered in our piece on how backlinks actually work.

Start with the phone. Two weeks of writing down what people ask, and the objection you started with will have quietly disappeared.