Three thousand keywords is data. Twelve decisions is research.

Because they end at export. A spreadsheet of 3,000 keywords with volume and difficulty columns is raw material, not research: it contains no decision about what to build, in what order, or why, and it quietly encourages chasing volume over value. Useful research ends in a prioritised build plan: these pages, in this order, targeting these query clusters, because of this evidence, with the SERP reality for each honestly assessed.

By expected value rather than volume: estimated traffic multiplied by conversion likelihood multiplied by deal or transaction value, discounted by realistic difficulty against your current authority. A 40-search-a-month term for a B2B service can outrank a 5,000-search informational term on every metric that matters. We also weight strategic fit, terms that build topical authority for the cluster you must eventually own get priority over isolated wins.

Intent mapping classifies each query by the job behind it (learn, compare, decide, buy, troubleshoot) and matches it to the page type that can satisfy it. It matters because targeting the wrong intent guarantees failure regardless of effort: a service page cannot rank for an informational query where Google has decided guides belong, and a blog post cannot rank where the SERP is full of product pages. We check the live SERP for every priority term, because Google's output is the definitive statement of what it thinks the query means.

Partly: conversational and question-shaped phrasings matter more in an AI-mediated landscape, and the prompts people type into ChatGPT or Perplexity are longer and more contextual than classic keywords. We include prompt-pattern research for the questions your buyers ask assistants, and note where being the cited source matters more than ranking blue-link position, which changes the content shape (answer-first, quotable, structured) rather than the target itself.

Included, with judgement applied: gap tools show every term competitors rank for and you do not, most of which you should never pursue. The value is pattern recognition, which clusters they systematically own, which content types earn those positions, and which gaps represent genuine commercial opportunity for your business model rather than theirs. Raw gap exports are the classic example of data volume substituting for thought.

A build plan: prioritised query clusters with a recommended page for each, intent classification, SERP notes (what currently ranks and why it will be hard or easy), value and difficulty reasoning, internal-link mapping into your existing site, and phasing across quarters. Plus the underlying data, because you own it, but the document that matters is the one with decisions in it.

Standalone research and build plans from £750 for focused businesses, £1,500-£3,000 for multi-service or multi-market scope. Included as the opening phase of any retainer, because starting execution without it is how budgets get spent on the wrong pages. Refresh reviews quarterly inside retainers as SERPs and business priorities move.