Keyword Metrics: The 9 Numbers to Check Before You Write a Page

Most keyword lists fail for the same reason: a single number decides everything. A query with 10,000 searches a month looks attractive until you see that the results page answers it directly, or that the people typing it will never buy. This guide explains the keyword metrics that matter, where each number comes from, what it hides, and how to combine them into one decision. Some of them come straight from Google, such as the historical metrics of the Google Ads Keyword Planner, others are modelled by third-party tools.

What are keyword metrics?

Keyword metrics are measurements attached to a search query that describe its demand, its competition and its value. They fall into two families:

• Selection metrics, used before writing: search volume, difficulty, cost per click, intent, trend, click share, and now prompt volume in AI assistants.
• Performance metrics, read after publishing in Google Search Console: impressions, average position, clicks and click-through rate.

The first family helps you choose; the second tells you whether the choice was right. Keep them apart: a keyword that ranks in position 4 with no clicks is a performance problem, while a keyword nobody searches is a selection problem.

The 9 keyword metrics, one by one

1. Search volume

Search volume is the average number of monthly searches for a query in a given country. In Google Ads, the Keyword Planner shows the average number of times people searched for a keyword and its close variants, rounded, and only for exact matches. Accounts without significant ad spend often see ranges (1K to 10K) rather than precise figures. Third-party tools blend this data with clickstream panels, so two tools rarely agree. Read volume as an order of magnitude, never as a forecast of traffic. Our article on keyword search volume goes deeper into the sources.

2. Keyword difficulty

Keyword difficulty is a score, usually from 0 to 100, that estimates how hard it is to reach the first page. Each tool computes it differently, most often from the number and strength of the backlinks pointing to the pages that already rank. It ignores content quality, topical authority and the fact that a smaller site can outrank a big one on a precise question. Use it to compare keywords inside the same tool, not as an absolute verdict.

3. Cost per click (CPC)

The cost per click is what advertisers pay for a click on the ad shown for that query. Google publishes a low and a high top-of-page bid, roughly the 20th and 80th percentiles of recent bids. A high CPC is the best free signal of commercial value: advertisers do not keep paying 15 dollars a click for queries that never convert. You can turn volume and CPC into a value estimate with our keyword value calculator.

4. Search intent

Search intent is the goal behind the query: learn, compare, buy or reach a specific site. It is not a number, but it overrides all the others. If the first ten results are product pages, a blog article will not rank, whatever its difficulty score. Classify intent by reading the results page yourself, not only from a tool label.

5. Trend and seasonality

A monthly average hides peaks. A query such as “Black Friday email ideas” does most of its volume in a few weeks. Google Trends shows interest on a relative scale: each data point is divided by the total searches of the region and period, then scaled from 0 to 100, as explained in the Google Trends help center. It tells you the shape of demand, not its size. See our guide to Google Trends for SEO.

6. Click share and SERP features

Not every search produces a click. In its 2024 study based on Datos clickstream panels, SparkToro found that 58.5% of US Google searches ended without a click, and that only 374 clicks out of every 1,000 searches went to the open web (360 in the EU) (SparkToro, July 2024). A query answered by an AI Overview, a featured snippet or a calculator sends fewer visits than its volume suggests. This is the logic of zero-click search: always look at what fills the page above the first organic result.

7. Impressions and average position

Once a page is live, Google Search Console reports impressions (how often a link to your site was seen) and average position. Position is the topmost position of your link, averaged across all the queries where it appeared, according to Search Console help. An average position of 8 can hide a page that is 3rd on one query and 40th on another, so filter by query before drawing conclusions.

8. Click-through rate (CTR)

Click-through rate is clicks divided by impressions. Compared with the usual CTR at a given position, it shows whether your title and snippet attract clicks. A page in position 3 with a CTR far below its neighbours usually needs a better title, not more links. Our SERP CTR calculator estimates the clicks to expect from a position.

9. Prompt volume in AI assistants

People now ask ChatGPT, Gemini or Perplexity the questions they used to type into Google, in longer and more conversational form. Prompt volume estimates how often a topic is asked in AI assistants. No platform publishes official figures, so these numbers are panel-based estimates, even rougher than search volume. Use them to spot topics where your brand should be cited, then check the real answers. Sorank's AI prompt research starts from your keywords to find the questions worth tracking.

Keyword metrics compared

How do you combine keyword metrics to choose a keyword?

A simple four-step filter avoids most bad choices:

1. Intent first. Keep only the queries whose results match the page you can build (guide, product page, comparison).
2. Volume and value next. Sort by volume, then use CPC as a tie-breaker: a 300-search query with a 12-dollar CPC often beats a 3,000-search query with no commercial value.
3. Difficulty against your authority. A young site should start with low-competition keywords and precise questions, then climb.
4. Click share last. Remove the queries fully answered on the results page, unless being cited in the answer is your goal.

Our keyword opportunity score calculator combines volume, difficulty and value into a single score to rank a list quickly.

Which keyword metrics are misleading?

Three readings cause most wasted articles. First, volume without intent: “CRM” has huge volume but mixes definitions, software lists and logins. Second, difficulty taken as absolute: a score of 70 in one tool can be 45 in another. Third, average position across all queries: a page can look stuck at 15 while it already ranks 2nd on the query that matters. Keyword density is not a metric to optimize at all: Google does not use a target density, and repeating a term only hurts readability.

Keyword metrics for SEO and GEO

For Google, keyword metrics decide which pages to write. For AI assistants, the question shifts from “how many people search this” to “when people ask about this topic, which sources are cited”. The two overlap: pages that rank well and answer clearly are also the ones assistants tend to quote. A practical setup is to choose keywords with the metrics above, write pages that answer the question in the first lines, and track both Google positions and AI citations on the same topics.

Sorank does this work automatically: its keyword research analyses your site and competitors to select keywords by volume, difficulty and intent, then the software writes and publishes the matching articles and tracks whether ChatGPT, Claude and Gemini cite your brand. The right keyword is the one whose intent, value and competition fit your site today, not the one with the biggest number. You can test it with the 3-day free trial.