How Many Keywords Should You Target? Per Page, and Across a Website
Target one primary keyword plus roughly 3–10 closely related variations on a single page, with one search intent per URL. Never stretch a single page across several intents. Across an entire website, the keyword count is an output, not an input: count your distinct buyer intents, give each one its own page, and the site total takes care of itself. A focused UAE service site usually maps hundreds of queries across a few dozen pages, not thousands spread over ten. One quick clarification before we move on: we're talking about organic search queries here, not Google Ads keywords. Paid-search keyword counts belong to a separate discipline, covered on our Google Ads management page.
RATH Infotech's team put this guide together, led by founders Sumeet Mehta and Manish Doshi. We've served Dubai and the rest of the UAE since 2014, and mapping keywords to URLs is core to what we do. Further down, you'll find the actual map we use on this very website. Below you'll find the per-page number, the per-site method, a real worked example, and a prioritisation matrix for deciding what to target first.
"One keyword per page" is dead: here's what replaced it
The old rule claimed every page should target exactly one keyword. That advice predates how Google actually works today. Google matches pages to intents, not to strings. A single page that ranks for its primary query almost always ranks for dozens, often hundreds, of related phrasings of the same intent, none of which it explicitly "targeted." Industry research on large ranking datasets shows top-ranking pages appearing for many hundreds of additional related queries, and our own client data backs this up (see the outcomes table below).
Google's spam policies call that out directly as a violation, not a ranking signal, and repeating keywords past natural use counts as stuffing under those rules. That's the real problem with cramming twenty keywords onto one page. But there's a second problem hiding in the same instinct: making a separate page for every phrasing of a question. Doing that just splits your relevance across near-duplicates instead of building anything stronger. So both ideas fall apart at once. The fix is simple: one intent cluster per page. Build the best answer to one search intent, cover every natural phrasing of it, and let the page earn the long tail on its own. We're keeping this single myth here as the canonical treatment; the fuller roundup lives in our SEO myths guide.
How many keywords per page: the 1 + 3–10 structure
Take any single page, whether it's a service page or a blog post; the logic doesn't change. A healthy keyword cluster plays four roles, and that's really the answer to the question of how many keywords you should use or focus on. First comes the primary keyword, the one query the page is built to win. Then there's a band of close variants, a set of question forms, and whatever long-tail phrases show up naturally as you write. The counts below are typical examples, not targets you need to force.
| Role | Typical count per page | What it looks like | Where it appears |
|---|---|---|---|
| Primary keyword | 1 | The single query the page is built to win | Title, H1, opening answer |
| Secondary variants | 3–10 | Synonyms and re-phrasings of the same intent | H2s and body copy |
| Question forms | 2–5 | The same intent phrased as questions | FAQs and H3 headings |
| Long-tail support | Open-ended | Specific, natural phrasings that emerge as you write | Woven through the body |
Primary versus secondary is a real distinction, and it matters. The primary term carries the title, the H1, and the opening answer. Secondaries are simply variants on that same intent: synonyms, reordered phrases, question forms, all placed so the page reads naturally for a human reader first. Repetition is where things break down. Saying the same phrase more often won't push you up the rankings, and once you cross a natural threshold it turns into stuffing. Covering the intent beats repeating the phrase, every single time. As for exactly where each keyword lands on the page (headings, body copy, image alt text), that's placement detail. We've laid it out in our on-page SEO guide, so there's no need to repeat it here.
How many keywords should a whole website target?
For the whole site, flip the question around. Don't pick a keyword total and work backwards. Instead, list your distinct buyer intents, give each one its own page, size the cluster on each page, and the site total is just the sum of those. That settles a related question too: how many pages your website needs for SEO. The answer is exactly as many as you have distinct intents, no more and no fewer. It's an output of your topical map, not a number you choose. The map below comes from the real keyword-to-URL plan we use on this site: one primary query plus a named secondary cluster mapped to every URL.
| Page | Primary query | Secondary cluster | Page's job |
|---|---|---|---|
| /seo-services-dubai/ | "seo services dubai" | 1 primary + 17 secondary variants | Commercial pillar, must win the head term |
| /seo-services-dubai/on-page-seo/ | "on-page SEO services" | 1 primary + a focused handful | Service spoke beneath the pillar |
| This guide | "how many keywords should I target" | 1 primary + 2 secondaries | Question guide capturing a pre-sales query |
Notice the pillar carries seventeen variants while this question guide carries two. That's not an accident: cluster size follows the breadth of the intent, not some fixed quota. Sizing varies by site type, too. A service SME might map a few dozen intents across a compact site. An ecommerce catalogue multiplies that by every category, filter and product line, so its map runs far larger, even though the per-page logic stays identical. We deliberately avoid publishing "the average site targets N keywords" benchmarks, because such numbers mean nothing without your intent map in front of us. As for why clusters beat scattered keywords, that's the deeper topic of our topical authority guide.
Keyword targeting strategy: research → cluster → map → prioritise
A keyword targeting strategy isn't a one-off list. It's a sequence you can run again and again. Six steps carry you from raw queries to a live, defensible map.
- Dig into the full query universe. Pull every phrasing your buyers actually use. That means working the entire industry-standard toolkit: Ahrefs, Semrush, Google Search Console and DataForSEO.
- Cluster by search intent. Group queries that want the same answer; each group is a future page.
- Map one cluster to one URL. Write it down as a keyword-to-URL table so nothing is targeted twice.
- Prioritise by difficulty × volume × business value. Sequence the work (matrix below).
- Publish or optimise the page as the single best answer to its intent.
- Review Search Console quarterly for queries you have started ranking for but never targeted, then fold them in.
The prioritisation matrix decides whether a plan succeeds or fails. Chase head terms first, before you've built any authority, and you'll burn months getting nowhere. Start with low-difficulty long-tail terms instead. That's where the early movement comes from, while the harder terms get time to mature.
| Bucket | Difficulty / volume profile | Business value | When to target |
|---|---|---|---|
| Quick-win long-tail | Low difficulty (around KD 0), modest or unrecorded volume | Often high, specific buyer intent | First, early movement while authority builds |
| Mid-tier commercial | Moderate difficulty, useful volume | High, clear commercial intent | Next, as topical authority accumulates |
| Head terms | High difficulty (KD 32 on our own head term), high volume | High but slow | Ongoing, the long game, backed by the cluster beneath it |
| Question / AEO queries | Often vol 0 in the tools | Citation value in PAA boxes and AI answers | Alongside, cheap to add, compounding upside |
Our own dataset makes the order obvious: the head term sits at KD 32, while the long-tail service queries come in at KD 0. Choose keywords based on difficulty and business value together, not on volume alone.
The UAE wrinkle: bilingual maps and "zero-volume" queries
Two realities specific to the UAE shift the numbers. The first is language. An English cluster often has an Arabic counterpart chasing the same answer, so for a business that's truly bilingual, the map can roughly double in size. That's a strategic call, not something that happens automatically. Which language to prioritise, and when, is what our English vs Arabic SEO guide covers.
Second, volume data. Keyword tools tend to under-report smaller markets, so plenty of real, valuable Dubai queries show a recorded volume of zero. That doesn't mean nobody's searching for them; it means the database is thin. In our own plan, we deliberately go after these recorded-zero-volume questions. The page you're reading right now is one of them, picked because it shows up in pre-sales conversations and AI-engine prompts even though the tools register nothing. As for emirate modifiers like Dubai, Abu Dhabi, and Sharjah, they count as separate intents only when your service or offer actually changes from one emirate to the next. If the answer stays the same everywhere, a single page can cover all three.
Too many keywords, too many pages: cannibalisation
Stuffing one page is one failure. Spreading a single intent across many pages is the opposite one. Get two or more pages chasing the same keyword, and Google has to pick a winner. Both usually lose out, since they end up splitting links, relevance and click signals between them. This kind of cannibalisation, driven by sheer quantity, shows up all the time on sites that spin a near-duplicate page for every phrasing rather than building one strong page per intent.
Detecting it is straightforward in Google Search Console: find a query that returns several of your own URLs, each ranking weakly, and you have found a cannibalisation problem. The fixes are consolidation (merge the pages), canonicalisation (point the weaker at the stronger), or re-mapping the intents so each has one clear owner. If your rankings have already dropped and you are not sure why, diagnose it properly first with our rankings-dropped guide; to check your existing keyword-to-URL mapping end to end, our SEO audit guide walks through it.
What disciplined targeting produces
Targeting clusters on purpose means you rank for a lot more than what you're explicitly chasing after. These keyword outcomes come from real engagements. Treat them as honest windows, not guarantees, and certainly never a promise of specific positions.
| Client (as we describe them) | Keyword outcome | Window |
|---|---|---|
| A Mumbai-based digital weighing-scale manufacturer | Ranking keywords grew from 1,168 to 1,557 | Quarter-on-quarter |
| A UAE roller-skating academy (Dubai, Sharjah & Ajman) | 30+ keywords on page 1 (16 in the top 3) | 6 months |
| An industrial brush manufacturer serving automotive, construction and metals | Category keywords ranking #3–#6 ("industrial brushes" #3, "rotary brush" #4) | Year-on-year |
The pattern here matters: each site went after a defined set of intent clusters and ended up ranking for multiples of what it aimed to win. That's intent-cluster targeting doing exactly what it's supposed to do. It isn't a shortcut, and it isn't a fixed number you can promise ahead of time.
FAQ
How many keywords should one page target?
One primary keyword, then roughly three to ten close variants covering the same search intent, plus two to five question forms for your FAQs. That's a cluster, not a list to stuff, Google's spam policies penalise unnatural repetition. One intent per URL is the rule that matters.
How many keywords should a whole website target?
There's no magic number here, it comes out of your topical map. Count your distinct buyer intents, give each one its own page, then total up the clusters. A focused UAE service site usually maps out to hundreds of queries spread across a few dozen pages, not thousands crammed into ten.
Is one keyword per page still best practice?
No, that rule predates Google's intent matching. Current practice is one intent per page: the primary keyword and every close phrasing of it sit together on the same URL. Split them across separate pages and you just dilute your relevance, inviting cannibalisation, which we'll get to below.
How many keywords can a single page rank for?
Far more than it targets. Research on large ranking datasets shows top pages also rank for many hundreds of related queries they never explicitly chased. Our own client data mirrors it: one manufacturer's site grew from 1,168 to 1,557 ranking keywords while targeting far fewer clusters.
What is the difference between primary and secondary keywords?
The primary is the one query the page has to win, it leads the title, the H1 and the opening answer. Secondaries cover the same intent in different clothes: synonyms, reordered phrases, question forms, worked naturally into H2s, body copy and FAQs. For placement detail, see our on-page SEO guide.
Which keywords should I target first?
Weigh difficulty against business value when you sequence. Go after long-tail, low-difficulty commercial queries first, since they move fast, then chase head terms once authority builds. Our own dataset makes the case plainly: the head term carries a KD of 32, while long-tail service queries sit at KD 0. The order picks itself.
Are zero-volume keywords worth targeting?
Often, yes, especially in the UAE, where volume databases under-report smaller markets. Plenty of real Dubai queries show a recorded volume of zero. Question-shaped zero-volume queries also feed People-Also-Ask boxes and AI-engine answers. That's exactly why we target this query on purpose, despite its recorded zero volume.
What happens if two pages target the same keyword?
Keyword cannibalisation is what happens when Google has to pick one page over another, and neither wins. Both tend to underperform. You can spot it in Search Console: look for a single query pulling up several of your URLs. Once you find it, consolidate the pages or re-map them so each search intent has just one clear owner. Rankings already slipped? Figure out the real cause before you touch anything.
Does repeating a keyword more often improve rankings?
No. Once usage goes past what's natural, repetition turns into keyword stuffing, and that violates Google's spam policies. It's as likely to hurt you as help you. Coverage of the intent wins instead: work in variants, question forms, related entities. That beats raw keyword density every time. Write for the reader first.
How many keywords should my SEO agency track and report?
Track every mapped cluster: the primary term plus its notable variants, not just a vanity top-ten. A useful monthly report shows movement across the whole map. It also flags queries you've started winning without meaning to. For what else belongs in a report, check our reporting guide.
Not sure how many keywords your site should be targeting?
Most sites we audit are either stretching one page across too many intents or spinning too many pages at one. The fix starts with a real keyword-to-URL map. Get a free growth audit and we will show you where your intents overlap, what you already rank for, and what to target next. If you are ready to hand it over, our SEO packages and the wider SEO services page explain how we run it. For where budget fits, see our SEO cost guide and what an SEO report should include. We are honest on terms: SEO needs time, so our minimum engagement is six months, and we never guarantee specific positions.