Keyword tools give every term a set of numbers: search volume, keyword difficulty, cost per click, sometimes "traffic potential" or "opportunity" scores. These numbers are useful, but each one is an estimate with limits. If you treat them as facts, you'll make poor decisions, such as ignoring valuable niche terms or chasing huge terms you can't win. This lesson explains what each metric really means and gives you a simple way to prioritise.
Search volume
Search volume is an estimate of how many times a keyword is searched in a month, usually averaged over 12 months, for a given country.
Where the numbers come from
- Google Ads Keyword Planner uses Google's own data, but advertisers without active spend often see broad ranges (for example 1K–10K) rather than exact figures. Google also groups close variants, so "running shoe" and "running shoes" may show the same number.
- Bing / Microsoft Advertising provides Bing's data, which is smaller but real.
- Commercial SEO tools estimate volumes from a mix of sources, such as clickstream panels, Keyword Planner data and their own models. Each tool's database differs, so tools disagree, sometimes by a factor of two or more.
What volume hides
- Seasonality. A 12-month average of 1,000 might be 4,000 in December and 200 in March. Check Google Trends or the monthly breakdown in Keyword Planner. In Australia, think of "tax return" peaking in July, "air conditioner" in summer, and "eofy sale" in June. In the UK, "boiler service" peaks in autumn.
- Location. National volume says nothing about one city. "Plumber" nationally is huge; your share in one suburb is small. Some tools allow regional or city filters; otherwise, use relative population and Google Trends' regional view as a guide.
- Low and zero volume. Tools often show zero for specific long-tail, local or B2B terms that do get searched. "Commercial kitchen exhaust cleaning Newcastle" may show little volume yet bring very valuable leads. Search Console data on your own site is the best proof that real demand exists.
- Clicks versus searches. Volume counts searches, not clicks to organic results. If a query is answered directly on the results page or crowded with ads and a local pack, organic clicks may be far fewer.
Keyword difficulty
Keyword difficulty (KD) scores, usually from 0 to 100, estimate how hard it is to rank organically. Most tools base them on the number and strength of links pointing to the pages currently in the top ten; some also consider content and domain metrics.
Use difficulty carefully:
- Compare terms within one tool, never across tools. A KD of 30 in one tool is not the same as 30 in another.
- It's a starting point, not a verdict. Always look at the real results. If the top ten are national brands, government sites and huge publishers, it's hard. If they include small business sites, thin pages, old forum threads or pages that miss the intent, there's room even if the score looks high.
- It ignores your own strengths. A local business with a strong Google Business Profile and good reviews can win local results that a generic difficulty score rates as hard. A site already known for a topic may rank more easily for related terms.
Cost per click (CPC)
CPC estimates what advertisers pay for a click on a search ad for that keyword. Even if you never run ads, it's a useful proxy for commercial value: advertisers bid more on terms that bring customers. "Personal injury lawyer" has high CPCs because a single client is very valuable; "what is a tort" has low CPCs because few searchers will hire anyone.
High CPC can also mean heavy ad presence on the results page, pushing organic results further down.
Other metrics you'll see
- Traffic potential: some tools estimate total traffic the top-ranking page receives across all its keywords, which helps when a topic has many variations.
- SERP features: whether a query shows a local pack, shopping results, videos or AI overviews, which affects how many organic clicks are available.
- Click metrics: some tools estimate the share of searches that result in any click.
- Parent topic: grouping keywords under the broader term a top page ranks for.
Treat all of these as estimates that help you compare, not as forecasts.
A simple prioritisation model
To decide what to work on first, score each keyword cluster (not each individual keyword) on four factors. A 1–3 scale is enough.
- Relevance (essential): does it match something you actually offer, in a place you serve? If not, drop it, however large the volume.
- Value: how valuable is a visitor or conversion from this topic? Use intent, CPC and your own margins and conversion rates.
- Opportunity: can you realistically compete? Consider your current position (from Search Console), SERP difficulty and competitors.
- Effort: can you improve an existing page, or do you need a new page, new research, photos or development work?
Then prioritise clusters with high relevance and value, good opportunity and reasonable effort. The quickest wins are often terms where you already rank on page two (roughly positions 11–20) for relevant, valuable queries, because improving an existing page is less work than starting from scratch.
Worked example: a Melbourne solar installer
| Cluster | Volume (est.) | KD | CPC | Current position | Priority |
|---|---|---|---|---|---|
| solar panels melbourne | High | High | High | 18 | High: relevant, valuable, already near page one |
| solar battery rebate victoria | Medium | Medium | Medium | Not ranking | High: valuable, timely, needs a new guide |
| how do solar panels work | Very high | High | Low | Not ranking | Low: informational, national competition, low value |
| solar panel cleaning melbourne | Low | Low | Medium | 9 | Medium: small but relevant add-on service |
The broad informational term has the biggest volume and the lowest priority. The installer focuses first on its main service page and a rebate guide.
Forecasting carefully
Stakeholders often ask "how much traffic will we get?" Any forecast should show its assumptions: estimated volume, expected position, an estimated click-through rate for that position, and a time frame. Present ranges (low, expected, high) rather than a single number, and update the forecast as real Search Console data arrives. Never promise a figure.
Reading metrics for local and B2B businesses
Most keyword metrics are built for national, high-volume markets. If you run a local or B2B business, adjust how you read them:
- Tiny volumes can still matter. A Perth industrial electrician might see "switchboard upgrade perth" at 50 searches a month. If one job is worth several thousand dollars, a handful of enquiries a month is an excellent return.
- Intent outweighs volume. "commercial refrigeration repair" with low volume is a better target than "how does a fridge work" with high volume, because the searcher needs a service now.
- Use your own data. Search Console queries, call tracking, enquiry forms and sales conversations ("what did you search for?") reveal the real terms customers use. In New Zealand, the US or Canada, local spelling and terminology differ, such as "tradie" versus "contractor", "lift" versus "elevator", so validate wording per market.
- Check Google Business Profile insights for the search terms people used to find your listing; they often reveal local phrasing that keyword tools miss.
Combining tool estimates with first-party data gives a much more accurate picture than either alone.
Key takeaways
- Search volume is a modelled estimate; it hides seasonality, location and niche demand.
- Keyword difficulty is tool-specific; compare within one tool and always check the real SERP.
- CPC is a handy proxy for commercial value.
- Prioritise by relevance, value, opportunity and effort, not by volume alone.
- Page-two rankings for relevant terms are often the fastest wins.
- Forecasts need stated assumptions and ranges, never guarantees.