Why a keyword count is free to promise
Ask an agency how many keywords they'll target and you'll get a number in about two seconds. Ask how many pages they'll ship and the pause is longer. That gap is the whole answer.
Adding keywords to a campaign costs nothing. It is a filter change in a rank tracker — 50 rows or 500 rows, the same person spends the same afternoon. Adding a page costs a brief, research, a draft, an edit, internal linking and a publish, which is six to nine hours of somebody's real week. So the two numbers move independently, and only one of them has a salary attached.
Which is why the keyword count in a proposal correlates with nothing you care about. Not hours. Not seniority. Not output. Not results. It correlates with how confident the salesperson was feeling. A ₹15,000/mo "unlimited keywords" package and a ₹1,50,000/mo plan tracking 80 terms are not describing different amounts of work — they're describing different amounts of honesty about what a tracker is.
There's a useful precedent here. Google has said plainly that it has no preferred word count for a page, and yet content briefs across the industry still specify 1,500 words as though it were a specification. Keyword counts are the same species of number: precise, checkable, and measuring the wrong thing.
Turning the three standard quotes into a workload
Almost every Indian SEO proposal states scope in one of three formats. None of them tells you what will happen in a month. Each one has a single question that turns it into something comparable.
- Two quotes both saying "100 keywords" can differ by 4x in real hours. The keyword count cannot distinguish them; the page count and the hour split can.
- A quote that answers the middle column in thirty seconds has planned the work. A quote that offers to "share the detailed scope after onboarding" has not.
- If you want the longer version of this exercise, how to compare SEO proposals runs the same conversion across price, hours and deliverables.
| What the proposal says | What it actually tells you | The question that resolves it |
|---|---|---|
| 50 keywords | That 50 rows will appear in a monthly report. Nothing about pages, hours or who writes them. | How many new pages ship each month, and how many hours sit behind each one? |
| 200 keywords | That someone widened a filter. Often the same team, same hours, same output, bigger spreadsheet. | Which of the 200 get a dedicated page, and which are hoping an existing page picks them up? |
| Unlimited keywords | That the work has never been costed. Unlimited is safe to promise precisely because tracking is the only part that scales for free. | What is the monthly output in pages and outreach hours, and which named person produces it? |
One page, one cluster: the unit the work is actually done in
Here's the mechanical reason keyword counts don't describe work. Google does not match strings any more. It matches meaning, which means a single well-built page ranks for a wide spread of phrasings nobody explicitly targeted.
You can verify this on your own site in two minutes. Open Search Console, go to the Performance report, filter to one URL that has been live for six months, and switch to the Queries tab. On a mature page the count is routinely dozens, and on a good pillar page it runs into the hundreds — most of which nobody wrote down in advance.
So the honest planning unit is the topic cluster: one substantial page answering the main question, a handful of supporting pages answering the adjacent ones, and internal links wiring them together. You plan clusters. You count pages. Keywords are what falls out of the arrangement afterwards.
This also explains the failure mode at the other end. An agency that genuinely targets 200 keywords by building 200 thin pages produces cannibalisation, not coverage — two or three of your own URLs competing for the same query and none of them winning. See keyword cannibalization for what that costs and how it shows up in Search Console.
What ₹75,000 a month covers, in units you can count
We publish our prices, so we can be specific rather than comparative. SEO runs from ₹75,000/mo, and smaller sites from ₹40,000/mo, both ex-GST. A steady-state ₹75,000 month is roughly 42 planned hours — a blended rate near ₹1,800 an hour across a strategist, a technical SEO, writers and an outreach lead.
Those 42 hours split roughly 4 strategy, 6 technical, 18 content, 9 links, 5 reporting. Translated into the units you'd actually check:
- Two to three substantial new pages a month, or three to four shorter ones, plus refreshes on two or three pages already slipping. Eighteen content hours does not stretch to eight researched pages, and any quote implying it is costing writing at under two hours a page.
- Six technical hours — crawl and indexation review, Core Web Vitals, redirects, schema, and catching whatever shipped on Friday.
- Nine outreach hours — prospecting, one angle worth pitching, the emails, the follow-ups. This is the slowest-compounding hour in the plan and the first one a cheap retainer deletes.
- No keyword ceiling at all. We'll track whatever set you want to be judged on. Tracking is not the constraint and never was — the 42 hours are.
The one place a fixed keyword list belongs
None of this makes keyword lists useless. It makes them the wrong thing to price. There is exactly one place a fixed list earns its keep, and it is at the other end of the engagement.
Agree a tracking scope: a named set of queries, frozen on day one, that both sides accept as the honest sample of whether the work is landing. Fifty is plenty. It should be weighted toward commercial intent, should exclude your brand name, and should be written down before anyone touches the site.
The distinction matters because the two lists behave in opposite directions. A work scope should grow as clusters get built. A tracking scope should stay still, because the moment an agency can add and remove tracked queries mid-engagement, the average position chart becomes an editorial decision.
And a tracking scope is still not the number that pays salaries. Rankings are a leading indicator; qualified leads from organic are the outcome. We freeze your trailing-90-day organic lead count on day one and beat it inside 90 days or keep working free until we do — a commitment against your own baseline, never a promised position for a keyword, because nobody controls Google's index. How the guarantee works sets out the mechanics.
- Freeze the list before work starts. Additions mid-quarter are how a flat campaign starts reporting improvements.
- Weight it commercially. Twenty queries that mean buying intent beat 200 that mean idle curiosity.
- Exclude branded terms. Brand queries rise with every other marketing channel you run and prove nothing about SEO.
- Keep a losers column. A report that only shows what improved is a highlight reel. Ask for movement on every tracked query, both directions.
Five questions that turn a keyword count into work
Run these on every proposal on your desk, including ours. Each one takes an agency that has planned the work about thirty seconds, and an agency that hasn't considerably longer.
- How many new pages ship in month three? Month one is audit-heavy everywhere, so month three is the fair question. A number you can hold them to.
- Which clusters, and what pages sit inside each? If the answer is a flat keyword list rather than a page map, the research hasn't happened yet.
- How many hours, split by workstream? Divide the fee by the hours for a blended rate. Under roughly ₹1,200 an hour in the Indian market means juniors are doing nearly all of it.
- What is the tracking scope, and can it change mid-engagement? The right answer is: fixed, agreed up front, changeable only by mutual agreement in writing.
- What number are we measured on, and what is it today? Not keywords. A business number, with a baseline, and a date. What questions to ask an SEO agency has the rest of the list.