Keyword Prioritisation Scorecard
Score a keyword list on relevance, intent, difficulty and opportunity, then defend the order.
Scores every keyword against business relevance, intent value, competitive difficulty and effort, produces a ranked build order with a stated weighting, and lists the keywords worth dropping entirely rather than deferring.
Ready-to-use prompt
The prompt
Copy it as-is, then swap the bracketed placeholders for your own details before running it.
Role: You are an SEO lead building a defensible keyword prioritisation scorecard.
Context:
- Keyword list (with any volume/difficulty data you have): {{keyword_list}}
- What the business sells: {{offer}}
- Site authority and current performance: {{site_context}}
- Team capacity: {{capacity}}
Task:
1. Propose a weighting across business relevance, intent value, difficulty and production effort, and justify it for this business.
2. Score every keyword 1-5 per criterion using only supplied or clearly reasoned inputs.
3. Produce a ranked build order with the composite score.
4. Group the ranking into Now, Next and Later against the stated capacity.
5. List Drop candidates — keywords that should never be built — with the reason.
Rules:
- Never invent search volume or difficulty. Where data is missing, mark the score as an assumption and say what would confirm it.
- Relevance to the offer outranks volume in every tie.
- A keyword may not appear in two groups.
Output: weighting rationale, scorecard table, Now/Next/Later grouping, Drop list.
- Make every criterion favourable at 5; explicitly invert difficulty and effort.
- Show the composite formula, with weights totalling 100%, and calculate before rounding.
- Identify shared page targets and existing URLs before assigning production slots.
- Separate score order from capacity-driven scheduling overrides and explain each override.
Estimated results
Editor's note
Why this prompt matters
Every keyword list is longer than the roadmap that follows it, and the cut is usually made by volume because volume is the only number in the sheet. This prompt makes the weighting explicit first, scores each keyword against it, and then splits the list against real capacity — including an explicit Drop list, which is the decision teams most often avoid making.
Anatomy
Prompt engineering breakdown
Role
Context
Goal
Constraints
Output format
What you'll get
Expected output
Scenario and input limits
A small-business accounts payable platform sells approval routing, supplier records and payment tracking, but does not create invoices. Its site has useful product documentation and limited organic visibility. The team can publish two standard pages per month; finance review for downloadable templates is unavailable this month.
The supplied shortlist contains six representative keywords from separate preliminary clusters. No volume or difficulty measurements were supplied. Scores below are reasoned assumptions, not keyword-tool findings. Relevance and intent reflect the stated offer; ease and effort need validation before scheduling.
Weighting and scoring
Use business relevance at 40%, intent value at 30%, competitive ease at 20% and production ease at 10%. This favours qualified demand without letting an easy article displace a commercially important page.
All criteria run from 1 (least favourable) to 5 (most favourable). Competitive ease reverses difficulty: a higher score means an easier anticipated contest. Production ease similarly rewards lower effort. Composite = 0.4R + 0.3I + 0.2C + 0.1P. These weights total 100%; the composite remains on a five-point scale.
Scorecard
| Keyword | R | I | C | P | Composite | Group | |---|---:|---:|---:|---:|---:|---| | invoice approval software | 5 | 5 | 2 | 3 | 4.2 | Now | | invoice approval workflow | 5 | 4 | 3 | 4 | 4.2 | Now | | small business accounts payable software | 5 | 5 | 1 | 2 | 3.9 | Next | | invoice approval checklist | 4 | 3 | 4 | 5 | 3.7 | Next | | invoice approval policy template | 4 | 3 | 3 | 2 | 3.2 | Later | | free invoice generator | 1 | 1 | 2 | 2 | 1.3 | Drop |
Build order and defence
Now: Build the approval software page first, then the workflow guide. Relevance is tied, so stronger purchase intent breaks the composite tie. Keep their purposes distinct: product evaluation versus implementation guidance. Check existing documentation before commissioning either page; updating an indexed page may beat creating another URL.
Next: Use the following month's two slots for the broader software page and checklist. The broader term has strong commercial value but provisionally faces harder competition. Confirm this through live results, competitor page quality and supplied difficulty data.
Later: Hold the policy template until finance review is available. Its score does not override that dependency.
Drop: Exclude “free invoice generator” under the current offer: the required functionality is absent, so traffic would not justify building it. Reconsider only if the product scope changes.
Under the hood
Why this prompt works
Declaring the weighting before scoring is what makes the output arguable in a good way — a stakeholder disputes one weighting rather than sixty rows. Forbidding invented metrics keeps the scorecard grounded in what you actually supplied, and the capacity grouping turns a ranking into a plan instead of a wish list.
Model fit
Best AI models for this prompt
ChatGPT
Fast and consistent across long lists; good at the composite arithmetic.
Claude
Best at flagging assumptions rather than quietly filling missing data.
Gemini
Useful when you paste current rankings so it can weight quick wins higher.
When to use
- After clustering, when candidate pages exceed production capacity.
- Before locking a quarterly roadmap and review slots.
- When stakeholders disagree about commercial versus informational sequencing.
- After an offer change makes previous priorities questionable.
When not to use
- Before overlapping terms are clustered into page opportunities.
- When live search-result inspection or measured difficulty is required.
- When specialist judgement must establish regulatory review effort.
- For paid-search bids, which need conversion economics and auction data.
Get more from it
Pro tips
- 1
Supply volume and difficulty with source, market and collection date; keep missing values visibly unverified.
- 2
Agree weights before reviewing rows; rerun one alternative weighting to expose unstable priorities.
- 3
Require a Drop reason tied to offer mismatch, not merely high difficulty.
- 4
Rerun quarterly, and sooner after capacity or product changes; preserve the previous scorecard.
- 5
Include existing URLs beside cluster representatives to distinguish updates from new builds.
- 6
Express capacity in page types or working days, including specialist review.
- 7
On a second pass, challenge assumption-heavy Now entries and request the evidence needed to retain them.
Don't ship this
Common mistakes
✗ Letting volume dominate the weighting by default.
Fix — State the weighting explicitly and require relevance to break ties.
✗ Accepting invented difficulty scores.
Fix — Require assumption flags and supply real tool data where you can.
✗ Skipping the Drop list.
Fix — Demand it — deferring bad keywords forever is worse than deleting them.
People also ask
Frequently asked questions
Q.Do I need real difficulty data?
It helps a lot. Without it the prompt marks scores as assumptions, which is honest but weaker than joining in tool data.
Q.Should I prioritise keywords or clusters?
Clusters, wherever you have them. Prioritising raw keywords tends to schedule the same page three times.
Q.How often should the scorecard be re-run?
Quarterly, or whenever capacity or the offer changes. Hand-editing the order loses the weighting logic.