AHREFS VS SEMRUSH: INTERNATIONAL KEYWORD MARKET-TEST PACKET
Template version: 2026-10-22
Status: Published companion template.

This is a blank editorial evaluation template, not a completed benchmark.
No results, coverage claims, price estimates, or ranking forecasts are supplied.
Duplicate sections for each run and each candidate. Preserve raw exports.

1. PROJECT CONTRACT — COMPLETE BEFORE COLLECTION

Project / offer:
Business owner:
Researcher:
Researcher's familiarity with each provider:
Target country:
Query language:
Intended page task and deliverable:
Included meanings / excluded meanings:
Approved local seeds (exact strings):
Seed source and wording reviewer / approval date:
Acceptance rules:
Discovery row budget per run:
Discovery time budget per provider:
Relevant account access / restrictions:
Unresolved prerequisites and owner:

2. RUN RECORD — ONE PER PROVIDER AND DISCOVERY MODE

Run ID:
Provider:
Account plan / access checked on:
Date / time / timezone:
Country database:
Query language and language filter (or Not available):
Input strings / approved seed-list file:
Fixed-list lookup or discovery:
Report / match mode:
Filters / sort / row limit:
Raw export filename:
Observed truncation / account restriction:
Research minutes (include scope of measured time):
Source-update dates shown / Unknown:
Deviation from declared test and reason:

3. PAIRED CANDIDATE RECORD — PRESERVE PROVIDER VALUES SEPARATELY

Candidate ID:
Exact query string:
Country / language:
Meaning and source / reviewer:
Found by Ahrefs / Semrush / Both / supplied fixed list:
Ahrefs run ID / source file / row:
Ahrefs volume estimate / missing reason:
Ahrefs KD / missing reason:
Ahrefs intent label if available:
Ahrefs source-update date if shown:
Semrush run ID / source file / row:
Semrush volume estimate / missing reason:
Semrush KD / missing reason:
Semrush PKD and specified domain, if used:
Semrush intent label if available:
Semrush source-update date if shown:
Scope / string / date differences requiring resolution:

Keep missing values distinct from observed zeros.
Do not average provider estimates or difficulty scores.
Do not treat PKD as a generic difficulty score.

4. SHARED REVIEW — RESTORE PROVIDER LABELS AFTER INITIAL REVIEW

Initial reviewer / date:
Provider labels hidden during initial review? Yes / No; reason:
Business relevance: Accept / Reject / Unclear
Wording review: Approved / Rejected / Pending
Wording reviewer and date:
Page task / suitable page type:
Duplicate exact string? Candidate ID:
Different phrase for the same page task? Task ID:
SERP check required? Reason:
Exact SERP query:
Search date / time / device:
Country setting / location method / limitations:
Evidence file or screenshot reference:
Result URLs reviewed and observed page tasks:
Decision: Candidate accepted / Rejected / Hold
Reason: Wrong meaning / Wrong task / Scope unresolved / Evidence missing /
        Duplicate task / Other (explain)
Existing target URL or proposed brief:
Next action / owner / due date:

Candidate acceptance enables prioritization; it is not publication approval.

5. RUN SUMMARY — RETAIN DENOMINATORS AND SCOPE

Country-language pair / run IDs:
Fixed input rows supplied:
Fixed input rows with usable evidence, by provider:
Discovery rows exported / reviewed, by provider:
Accepted / rejected / unresolved candidates, by provider:
Distinct accepted page tasks, by provider:
Exclusive accepted candidates and task IDs:
Did any exclusive candidate change a page decision? Explain:
Disagreements resolved / unresolved:
Collection / review / cleanup time (state measurement scope):
Account restrictions that affected the comparison:
Reasons this sample may not represent other tasks or markets:

Do not present counts as population-wide accuracy or database coverage.
Do not generalize one researcher's timing to all users.

6. TOOL DECISION MEMO

Decision: Keep current tool / Choose Ahrefs / Choose Semrush /
          Use both for a named question / Hold
Project scope supported:
Observed evidence supporting decision (run and task IDs):
Editorial inference, clearly separated from observation:
Question a second tool would resolve, if any:
Current access, seats, export and refresh limits checked:
Incremental cost and billing period checked on / source:
Unresolved facts and responsible owner:
Reason / trigger to revisit decision:
Decision owner / date:

7. PUBLICATION HANDOFF, IF CANDIDATES BECOME CONTENT

Brief or keyword-map reference:
Author / editor:
Factual reviewer and actual completion date:
Language reviewer and actual completion date:
Publisher and release checks:
Pending review means Hold; never fill approval from assumption.