Size Total Search Demand For A Business Case

The usual answer sums the volume column of a keyword export, which counts the same demand several times across near-duplicate phrasings and then values clicks that no longer exist because the result page answers the query in place. The number is always too large and it collapses the first time somebody samples it. A defensible size needs deduplication, a click-through assumption taken from your own data, and an explicit split between total demand and the share you could actually take.

The skill file

---
name: search-demand-sizing-for-a-business-case
description: Use when someone asks how large the search opportunity is for a category or market before budget is committed.
---

# Size Total Search Demand For A Business Case

The usual answer sums the volume column of a keyword export, which counts the same demand several times across near-duplicate phrasings and then values clicks that no longer exist because the result page answers the query in place. The number is always too large and it collapses the first time somebody samples it. A defensible size needs deduplication, a click-through assumption taken from your own data, and an explicit split between total demand and the share you could actually take.

## What you need first

- a keyword export covering the whole category, including terms you cannot win
- your own Search Console click-through rate by position band, for this site
- live SERPs for the twenty largest clusters, checked in the target country
- the tool name, country setting and data date for the export

## Method

1. Deduplicate into clusters by shared ranking URLs, then count each cluster once at the volume of its largest member rather than summing the members, because tools routinely report the same underlying demand against several close variants.
2. Remove competitor brand terms. That demand is real but no content plan addresses it, and it inflates a category total more than any other single factor.
3. Open the SERP for the twenty largest clusters and mark the ones where an AI answer, a shopping block or a large feature set pushes organic below the fold. Discount those clusters visibly rather than deleting the rows, so the assumption can be argued with.
4. Apply your own click-through curve by position band instead of a published one. Published curves average across query types and run optimistic for commercial intent, which is where the value in the export sits.
5. State the addressable subset as its own number: the clusters inside your realistic difficulty range that current production capacity could reach within twelve months.
6. Give the answer as a range and name the two assumptions driving its width, normally the click-through rate and the share of clusters won. A single point estimate invites the reader to treat a sizing as a commitment.
7. Record the tool, country and data date beside the figure, so it can be reproduced or challenged rather than becoming a number the business repeats for years with no provenance.

## What this produces

A deduplicated category demand figure of one page, with a separate addressable subset, an explicit click-through assumption and a stated range.

## Where this goes wrong

- Summing the volume column, which commonly overstates real demand by two to five times once duplicates and brand terms come out.
- Presenting the total as a traffic target, which turns a sizing exercise into a number you will be measured against and miss.
- Sizing a market with a tool whose data for that country is thin without saying so, which puts a confident figure on a weak sample.

---

From the QuQi skill library - https://www.quqi.io/skills/search-demand-sizing-for-a-business-case

What you need first

• a keyword export covering the whole category, including terms you cannot win
• your own Search Console click-through rate by position band, for this site
• live SERPs for the twenty largest clusters, checked in the target country
• the tool name, country setting and data date for the export

Method

1. 01 Deduplicate into clusters by shared ranking URLs, then count each cluster once at the volume of its largest member rather than summing the members, because tools routinely report the same underlying demand against several close variants.
2. 02 Remove competitor brand terms. That demand is real but no content plan addresses it, and it inflates a category total more than any other single factor.
3. 03 Open the SERP for the twenty largest clusters and mark the ones where an AI answer, a shopping block or a large feature set pushes organic below the fold. Discount those clusters visibly rather than deleting the rows, so the assumption can be argued with.
4. 04 Apply your own click-through curve by position band instead of a published one. Published curves average across query types and run optimistic for commercial intent, which is where the value in the export sits.
5. 05 State the addressable subset as its own number: the clusters inside your realistic difficulty range that current production capacity could reach within twelve months.
6. 06 Give the answer as a range and name the two assumptions driving its width, normally the click-through rate and the share of clusters won. A single point estimate invites the reader to treat a sizing as a commitment.
7. 07 Record the tool, country and data date beside the figure, so it can be reproduced or challenged rather than becoming a number the business repeats for years with no provenance.

What this produces

A deduplicated category demand figure of one page, with a separate addressable subset, an explicit click-through assumption and a stated range.

Where this goes wrong

• Summing the volume column, which commonly overstates real demand by two to five times once duplicates and brand terms come out.
• Presenting the total as a traffic target, which turns a sizing exercise into a number you will be measured against and miss.
• Sizing a market with a tool whose data for that country is thin without saying so, which puts a confident figure on a weak sample.

Use this skill in your own AI

The download is a plain markdown file with the name and trigger in its frontmatter. Where an assistant supports skills it can load itself, that frontmatter is what it reads to decide this one applies.

Questions about this skill

More in Keyword research