How many blog posts do you actually need? Fewer than you're told.

Publishing cadence is the wrong unit

Ask how much content you need and almost every answer comes back as a rate. Two posts a week. Four a month. Sixteen for the quarter. Rates are easy to sell, easy to invoice against and easy to deliver, which explains their popularity and nothing else.

Search doesn't work in rates. A query either has a page that answers it well or it doesn't. Google isn't tracking your publishing frequency and rewarding consistency — it's matching queries to documents. Publishing four times a month for three years produces 144 documents, and if 100 of them answer questions nobody types, you've bought 100 maintenance obligations.

The unit that actually matters is coverage of a decision. Your buyer is working something out. That process contains a specific, countable set of questions. Answer each one properly, once, and you are done — not done with SEO, but done with the part of it that content solves. Everything after that is refreshing, deepening, linking and converting.

"Done" is a word the industry has quietly removed from the conversation, and its absence is worth about a lakh a month to somebody.

Count the questions, not the keywords

A keyword export is not a question list. It's a list of strings, deduplicated by wording rather than by meaning, ranked by an estimate. Work from it directly and you'll commission four pages that answer the same thing in four registers, which is how cannibalization gets built.

Build the question list from five sources, in this order. The first one is free, takes an afternoon, and is better than anything a tool will give you.

• Write each item as a question in the buyer's words, not as a keyword. "How long before this starts working?" not "seo timeline india". The phrasing decides the page you write.
• Tag each question with the stage it belongs to: doesn't know the category, sizing it, comparing approaches, comparing providers, checking the risk. Stage decides the format.
• Mark which existing page already answers it. Usually a chunk are covered badly. Those are refreshes, and refreshes beat new pages roughly nine times out of ten at this stage.

1. Your last 20 sales calls. Write down every question asked before the price came up. If calls are recorded, this is two hours of skimming. Nothing else on this list has a higher hit rate, because these are questions asked by people who were actually considering buying.
2. Your support and WhatsApp inbox. Post-purchase questions are pre-purchase questions that nobody answered in time. Each one is a page.
3. Search Console, queries you already earn impressions for but rank 15–40 on. Google has already decided you're relevant to these. You're just not good enough at them yet. This bucket is usually the cheapest returns on the whole list.
4. People Also Ask and autocomplete around your five head terms. Two levels deep, then stop — past that you're collecting queries typed by students and competitors, not buyers.
5. Competitor coverage that earns impressions — not their whole sitemap. Look at what actually ranks for them, subtract what you have, and treat the remainder as candidates rather than commissions.

The dedupe rule that halves the list

A raw question list from those five sources usually comes back somewhere between 80 and 200 items. That's not your page count. Most of them are the same question wearing different clothes.

The rule: two questions are one page when a single answer, on a single page, would leave both searchers completely satisfied. Not when the keywords look similar. Not when a tool put them in the same cluster. When one honest answer finishes both jobs.

It sounds obvious and it's routinely got wrong in both directions. "How much does SEO cost" and "what do SEO companies charge" are one page — same person, same need, same answer, different phrasing. "How much does SEO cost" and "is SEO worth the money" are two pages, even though they sit next to each other in every keyword tool, because the first wants a number in a table and the second wants an argument with the counter-case in it. Merge those two and you write a page that half-satisfies both.

Run the list twice. First pass merges by answer. Second pass deletes anything with no commercial relevance — the questions asked by people who will never buy from you, however high the volume. Volume estimates in India are rounded and modelled anyway, so treat them as a rough ordering rather than a measurement, and never as a reason to write a page your buyer wouldn't read.

What the count usually comes to

These are the ranges we plan against, from counting question sets in the categories we work in. They aren't survey data and we're not going to dress them up as such — they're planning numbers, and yours will land somewhere inside them or tell you something interesting about your category.

The last column is the part that ends most arguments about content budgets, because it turns a page count into a calendar without anybody having to sell a cadence.

• The count scales with the complexity of the purchase, not the size of the company. A two-person consultancy selling a considered ₹5,00,000 engagement needs more pages than a 200-person company selling a ₹499 subscription.
• It scales with the number of things you sell. Each additional product line brings its own decision, and decisions don't share pages as neatly as anyone hopes.
• It does not scale with your ambition. Wanting to be the biggest player in your category doesn't add questions to your buyer's list.

Why cadence advice sells retainers, not rankings

Follow the money and the advice makes sense. A monthly retainer needs a monthly deliverable, and pages are the easiest deliverable in marketing to count. Nobody can see a link being earned or a crawl budget improving, but everybody can count four posts. So the industry sells pages per month, and the content plan gets shaped by the invoice rather than by the buyer.

Here's the arithmetic, ours, out loud. A ₹75,000/month SEO retainer covers four disciplines — technical, content, links and reporting. Content takes 40–50% of the hours. Done properly, meaning research, an interview with someone who actually knows the subject, drafting, editing, internal linking, publishing and an index check, that buys four to six new pages a month plus around two refreshes. If someone offers twenty pages at the same price, one of the research, the expert input or the editing has been dropped. Usually all three.

At four to six a month, a 50-page set is finished in nine to twelve months. Which raises the question the cadence model exists to avoid: what does the retainer do in month 14? The honest answer is that the content line goes down and the technical, link and conversion lines go up, because that's where the remaining returns are. The dishonest answer is that you keep publishing, and the site slowly fills with pages that were commissioned to fill a slot.

We run month-to-month after the first quarter for exactly this reason. If the work stops being worth ₹75,000, you should be able to leave in 30 days with every asset. And what we're measured on isn't page count — it's your trailing-90-day qualified organic leads, frozen on day one. Miss it in 90 days and we keep working free until we beat it. That commitment is quite hard to hit by publishing filler, which is the point of writing it into the contract. Our pricing has the rest of the numbers.

The quality floor: what a page has to add

A smaller page count only works if the pages are better, and "better" needs a definition you can apply to a brief before anybody writes anything.

Ours is a floor, not a ceiling. A page ships if it does at least one of these four things. If it does none, it's the fourth-best version of something that already exists, and you'll maintain it for a decade in exchange for nothing.

• States a number nobody else has published. A real price, a range with the reasoning attached, a count, a timeline you'll stand behind. Specificity is also what AI assistants can quote — hedged prose gives a model nothing to attribute.
• Answers the question the top three results dodge. Read them first. There's almost always a part everyone skips because it's uncomfortable or because it makes the writer's employer look bad.
• Records first-hand experience. What happened when you actually did the thing. This is the one nobody can copy and the one generated text cannot produce.
• Makes a decision for the reader, including who it's wrong for. "It depends on your goals" is the sound of a page adding nothing.

What happens after the core set is published

This is the part nobody plans for, and it's where a content programme either starts compounding or starts costing. Six months after the set is complete, the work looks nothing like it did during the build.

In rough order of returns.

1. Refresh everything ranking 8–20. These pages already have relevance and some links. They need depth, a better title and a few internal links. Rewriting one of these usually beats writing three new pages, and it's the single most reliable move in mature SEO.
2. Prune the tail. Any page with zero impressions over 12 months and no referring domains is costing you crawl attention and adding nothing. Merge it into a sibling or remove it. This is uncomfortable and it works.
3. Deepen the three or four pages nearest the money. Now depth pays, because these pages have a cluster pointing at them and a position worth improving.
4. Spend on links to the informational set. Nobody links to a pricing page. They link to the page that explained something clearly, and internal linking moves some of that where it's needed.
5. Fix conversion. The pages rank, the traffic arrives, and the form has six fields and no phone number. At this stage a conversion fix is often worth more than another quarter of publishing.
6. Add new pages only when the category changes — a new service line, a regulation, a competitor category that didn't exist. Not because it's the first of the month.

Where volume genuinely is the strategy

There are categories where the page count really is in the thousands, and pretending otherwise would be its own kind of dishonesty. The distinction isn't the industry — it's whether each additional page carries information that exists only on that page.

Ecommerce is the clearest case. Category, filter and comparison pages aren't blog posts; they're inventory presented in a way search engines can read, and each one genuinely differs because the products differ. The editorial blog on the same site is still a 15–30 page job. Marketplaces and directories are similar: the page count is the product. Publishers run on volume because news is perishable and the business model is impressions.

Then there's the templated middle ground — routes, pincodes, price comparisons, integration pages. Those can be entirely legitimate at scale, and they can also be the fastest way to get a domain classified as spam. The test is in where the scaled content line sits, and it comes down to whether the page carries data nobody else has.

If you're not in one of those categories — and most Indian B2B service businesses aren't — the number is 30 to 60, the work takes about a year at an honest cadence, and the second year is about making those pages better rather than making more of them. What content actually does for SEO makes the wider case; this page is just the arithmetic underneath it.