Crawl Foundry
Domain intelligence

Read Domain Overview as a sourced profile and dated evidence

Domain Overview combines immutable paid analyses with a profile that selects the latest successful source per module. Use it to investigate one domain and market without confusing provider estimates, loaded top sets, last-known evidence, or completed collection with verified business outcomes.

Two valid views

Use an immutable run for one acquisition event and the Domain profile for the latest successful evidence per module.

Visible provenance

Carry domain, market, source run, provider date, module state, and loaded coverage into every comparison.

Human-owned decisions

Use the dossier to qualify an investigation; let the owning workflow approve action and verify a later outcome.

Start with the distinction between a domain profile and an analysis run

Domain Overview keeps paid analysis runs as dated evidence. Each run belongs to one workspace, normalized domain, country, and language. It is not overwritten when you analyze the same domain again. The run picker remains the place to reopen that exact result, including completed, partial, failed, and compacted entries that are still retained.

The default Domain profile is a different read. It composes the latest successful evidence available for each eligible module in the same domain and market context; a directed keyword gap remains tied to its explicit pair and run. Rankings can therefore come from one run while backlink details use another. This protects useful older evidence when a newer provider request fails, but it also means the profile can contain several source dates. Quiet source notices identify sections that are last known or unavailable.

Use one reading order from paid request to verified decision

1
Confirm the exact domain, country, language, keyword depth, optional comparison domain, displayed price, balance, and storage capacity before starting a run.
2
Wait for the run to reach completed, partial, or failed, then read its module coverage and provider dates before using a value.
3
Choose whether the question needs the latest-known Domain profile or one immutable analysis run.
4
Move through Overview, Opportunities, Rankings, Competitors, Pages, Backlinks, and History according to the question, without merging their units.
5
Reproduce the relevant row, page, source, date, market, and loaded-set limit in the feature that owns the next action.
6
Record the human decision and compare a later equivalent measurement before claiming an outcome.

Approve a paid on-demand analysis deliberately

New analysis accepts a domain, one of the curated country and language pairs, a top-500 or top-1,000 ranking-keyword scope, and an optional comparison domain. The gap is directed: the analyzed domain ranks for the returned terms while the comparison domain does not. Reversing the two domains changes the question and requires a different run.

The button stays unavailable until the domain is valid, the estimate and storage usage have loaded, the organization balance covers the estimate, and neither the workspace nor the domain context has reached its storage limit. Price and current balance appear beside the action. The estimate is a maximum for the requested bundle; settlement uses the provider work that actually completed. A network retry reuses the same client request ID so the same submit attempt cannot create a second charge.

  • Country-level markets are supported here; city-level rank tracking belongs in Position Tracking.
  • A tracked own domain can prefill the comparison for a competitor, but the value remains visible and editable before purchase.
  • There is no scheduled paid refresh and no automatic paid retry.
  • Existing usable data from a plain domain link opens without forcing a new purchase.

Keep operation state, coverage, and cost separate

A module can be available, empty, truncated, failed, pending, or not requested. Empty is a successful zero for that provider module. Failed is unknown, not zero. Truncated means returned rows are current evidence, but absence outside the stored top set cannot prove that an older keyword, page, competitor, or link source disappeared.

StateWhat it establishes
RunningThe accepted request is still collecting or persisting modules; the page updates but no final coverage is established.
CompleteThe run reached its normal terminal path; individual modules can still be empty or truncated.
PartialAt least one requested source failed while other fetched evidence remains usable and chargeable.
FailedThe run did not produce a usable dossier; the page distinguishes recorded provider spend from a request that never started.
CompactedDetailed row lists have expired under retention, while retained header metrics and historical series remain readable.

Read the Domain profile as a sourced composition

A profile is built from retained runs without buying new provider data. For each module it selects the newest successful source. When a newer attempt fails, the older source can remain visible as last known with its real date. A successful empty result replaces older non-empty evidence because a confirmed zero is different from a failure.

The profile header states its update time, not one universal provider time for every value. A source chip appears when a section diverges from the profile head or is unavailable, and it can open the contributing snapshot. Switch to an individual run whenever a decision requires one coherent acquisition event rather than the profile's best-known composition.

  • Current means confirmed by the selected successful module source.
  • Last known means an older successful source remains after a newer module attempt failed.
  • Unavailable means no retained successful evidence exists for that module.
  • Profile-building failure does not alter or erase the underlying saved runs.

Choose the tab that owns the question

TabQuestion and main boundary
OverviewHow the headline strength, position mix, competitors, keywords, pages, links, and growth fit together; a dossier, not a replacement for each detailed ledger.
OpportunitiesWhich stored keywords in positions 4 to 20 have modelled CTR uplift, brand share, SERP-feature risk, or an authority gap; these are heuristics, not forecasts or instructions.
RankingsWhich provider-observed keywords, URLs, positions, demand, estimated traffic, difficulty, paid signals, and SERP features are stored for this market.
CompetitorsWhich organic search domains overlap, how their estimated reach and ranking profiles compare, and which should be inspected or tracked.
PagesWhich stored URLs carry estimated traffic and rankings, how concentrated the footprint is, and what a matching Site Audit run observes for those pages.
BacklinksHow backlink attributes, anchors, referring domains, target pages, source distributions, and matching Site Audit target health are structured.
HistoryHow provider monthly history and repeated stored analyses describe direction on different clocks.

Know which questions the evidence can answer

  • How broad is the provider-observed ranking footprint in one country and language?
  • How much of that footprint sits in top positions, and which stored keywords and URLs carry it?
  • Which domains share rankings in this market, and how large is their modelled organic reach?
  • Which comparison-specific terms are present for one domain and absent for the other in the provider dataset?
  • How large and diverse does the observed backlink profile appear, and which sources and pages dominate the loaded evidence?
  • Does monthly provider history or a series of comparable analyses show a persistent movement worth investigating?
  • Which selected keyword or domain evidence can be handed deliberately to the Keyword Database, Keyword Research, Competitive Research, Position Tracking, or Site Audit review?

Do not collapse provider metrics into one score

Authority is the provider's backlink rank displayed consistently on a 0 to 100 scale. Organic traffic and traffic value are modelled from rankings, demand, click assumptions, and paid-click values. Organic keywords count provider-observed rankings in the selected market. Backlinks and referring domains describe link volume and source diversity. These values have different sources, units, and regional behavior.

A movement in one metric does not explain another. More ranking keywords can sit in weak positions. Estimated traffic can change with demand or the provider model. Link volume can grow without relevant source diversity. Read the detailed rows and history before forming a hypothesis, then use first-party analytics, Search Console, crawl evidence, release history, and manual review where the claim requires them.

Treat loaded rows and stored totals honestly

Domain Overview stores bounded provider result sets and reads them in pages. The requested ranking-keyword depth is 500 or 1,000. Other modules have their own limits, and panels often load only an initial page. A caption such as loaded, loaded of total, top, or first indicates the denominator currently available to the visual or table.

Filters, intent strips, charts, selections, and top-row observations can describe the loaded or stored set rather than every ranking, page, competitor, anchor, or referring domain on the web. Load more where available and carry the module coverage into any exported conclusion. An absent row inside a truncated top set is not evidence of absence.

Move evidence without inventing automation

Selected ranking or opportunity keywords can be added to the Keyword Database with imported provider metrics and a domain-named manual list. That handoff does not buy enrichment and does not overwrite the workspace's own-domain tracking state. Existing keywords remain deduplicated under the database import contract. A bounded selection can instead be carried into Keyword Research as explicit seeds.

Discovered domains can be remembered as tracked competitors, opened for a directed gap analysis, or continued in Competitive Research. Pages and linked targets can be cross-checked against the latest matching Site Audit run. These links preserve a starting context; they do not approve a keyword, declare a competitor strategically relevant, create a brief, fix a page, or prove an outcome.

State what Domain Overview cannot prove by itself

  • Estimated visits are not Analytics sessions, users, conversions, or revenue.
  • Provider-observed rankings are not universal across every location, device, searcher, or date.
  • Authority is not a Google metric, a quality certificate, or a causal ranking factor.
  • Organic overlap does not make a domain a direct business competitor.
  • A keyword gap does not prove relevance, a missing page, or an editorial priority.
  • An observed ranking URL does not prove cannibalization or correct search intent.
  • Backlink totals, follow status, or provider spam signals do not prove link quality or justify removal or disavowal.
  • A chart movement does not identify the cause, and completing a paid run does not establish business impact.

Understand what remains available over time

Recent retained analyses keep their detailed keyword, competitor, page, anchor, and referring-domain rows. Older eligible runs can be compacted so the selector, headline metrics, monthly history, and trend charts remain while detail lists are removed. The newest usable analysis for a domain context is protected from automatic compaction and deletion. Failed runs and older unprotected headers follow shorter or longer retention tiers described in the snapshot guide.

A user with the required workspace editing capability can delete a terminal run. Deletion removes its stored dossier and can require the profile projection to rebuild around remaining sources. Billing records remain as financial audit evidence. A running analysis cannot be deleted through the same terminal-run action, and removing a snapshot does not reverse provider work that already occurred.

Finish with a defensible review record

1
Write the decision question before opening a metric, for example whether a decline is broad, page-specific, link-related, or only modelled.
2
Record workspace, domain, country, language, profile or run, source snapshot, provider date, comparison direction, and visible coverage.
3
Separate observed values from explanations and mark unknown, empty, failed, truncated, last-known, and compacted evidence correctly.
4
Open the exact keyword, URL, competitor, source, or history row that supports the next investigation.
5
Confirm the action, owner, constraints, and success measure with a person in the owning workflow.
6
Compare a later equivalent measurement and first-party outcome evidence before reporting success or causation.