Turn competitor data into a reviewable comparison
A frozen claim
A living catalog
Visible boundaries
Move from a comparison question to reviewed evidence
The route and its actions have separate access gates
Competitive Research sits behind the Keyword Database entitlement and the competitor feature. Viewing comparisons and the catalog requires keyword view access. Creating a paid comparison, curating competitors, or recomputing the catalog requires keyword edit access. Keyword import, list management, and a new Keyword Research run have their own permissions and feature gates.
The current surface has Overview, Competitors, Comparison, and Statistics tabs. A teammate can therefore read saved evidence without necessarily being allowed to spend balance, activate a competitor, import a keyword, or start another research job.
Keep the saved comparison and the live competitor set distinct
A saved comparison freezes one own domain, one market, one depth, and the selected competitor domains. Its snapshot references do not change when the active competitor list changes later.
The operational catalog is a workspace projection. It combines recurring SERP domains, manual evidence, tracking, Domain Overview, discovery, and comparison evidence. Search competitors, marketplaces, publishers, directories, and known business rivals can all appear there for different reasons. The two sets may overlap, but they are not interchangeable.
One comparison is an immutable anchor over domain snapshots
The comparison row stores normalized domains, country, language, depth, analysis plan, snapshot references, quote, creator, time, and request identity. The domain measurements live in ordinary Domain Overview snapshots rather than copied result rows inside the comparison.
That design lets a new comparison reuse matching evidence without rewriting an older claim. Preparing Deep from Summary or retrying missing domains creates another comparison request. It does not mutate the saved comparison that was already reviewed.
The catalog is a local workbench, not another provider crawl
Catalog recomputation reads evidence already stored in the workspace. It can update SERP overlap counts, scores, current summaries, and free active competitor slots without reserving balance or contacting a provider.
A catalog row can be a rich profile, a tracked-domain stub, or an evidence-only candidate with no measured metrics yet. Zero-looking placeholders on an evidence-only row must not be read as measured zero. Open the row and check its source badges, dates, and coverage first.
One screen can contain several evidence clocks
Paid comparison snapshots carry their own provider and fetch times. Monthly history belongs to the stored base snapshot. Deep keyword and page rows belong to detail snapshots. The catalog can also join current Keyword Database, tracking, SERP, Domain Overview, and manual evidence.
A newer catalog metric does not refresh an older comparison, and a current tracked rank does not rewrite provider history. Record source, market, module, and observation time beside any number used in a decision.
Comparison status is an aggregate, not a blanket quality claim
The comparison state is derived from all referenced snapshots. Running wins while any required snapshot is still running. All failures produce failed. A mix of completed, partial, or failed references produces partial. A missing referenced snapshot counts as failed.
Each module has separate coverage and freshness. Rank overview can be available while keyword gap, anchors, referring domains, or indexed pages are pending, truncated, unsupported, or failed. Read the coverage ledger before treating completed as complete evidence for every question.
Paid acquisition and local actions are different events
| Action | Current boundary |
|---|---|
| Create a saved comparison | May reserve and settle balance for new matching snapshots. Reused snapshots are shown as zero-cost lines. |
| Open history, signals, gaps, or pages | Reads the saved snapshot references within server limits and does not purchase additional data. |
| Recompute the competitor catalog | Rebuilds local projections from stored evidence and makes no provider request. |
| Run full competitor keyword analysis | A separate paid job with its own depth, quote, reservation, status, cache, and settlement. |
Saved comparisons remain dated claims
Comparison history is paginated newest first and currently retains comparison anchors for up to 365 days. The interface reads a bounded page rather than loading the whole workspace history at once.
Reuse windows are shorter than retention. Matching competitor snapshots are normally reusable for 7 days, the own Summary snapshot for 30 days, and matching in-flight work for 30 minutes. These are current operating limits, not a promise that a historical comparison will always remain available or current.
Competitive evidence does not choose the strategy
Contextter can show scale, movement, overlap, gaps, pages, intent samples, link signals, and tracked positions. It does not know by itself whether a domain serves the same buyers, whether a query can convert, whether a page fits the product, or whether the organization can credibly publish the answer.
Treat the workbench as an investigation ledger. Relevance, priority, risk, page ownership, content direction, and publication remain human decisions.
A defensible review leaves a recoverable trail
- The own domain, competitors, country, language, depth, and decision question match.
- Every cited metric includes its source, date, and coverage state.
- Saved comparison domains are not confused with the current active competitor list.
- Provider estimates, local projections, and tracked rankings are interpreted separately.
- Partial, truncated, unsupported, and missing modules remain visible in the conclusion.
- Any import, tracking change, research run, page brief, or content task is approved as a separate action.