New

• Keyword intent, topic clusters, and Share of Voice — Tracked keywords now show a search-intent badge (Informational, Navigational, Commercial, or Transactional) so you can see what each query is really after. A new Clusters tab groups your keywords into topics by shared root terms and overlapping search results, each cluster showing its keyword count and combined search volume. The Keywords page and the SEO Intelligence overview also gain Share of Voice — the estimated slice of total search traffic your rankings capture — as a percentage plus estimated monthly clicks. The Strategy Hub and Data Chat factor in intent and Share of Voice too. Intent and clusters are generated during your next comprehensive scan.
• Internal link and orphan-page analysis in site audits — Audits now map how your pages link to each other. A new Internal Links card shows your orphan-page count (pages nothing links to), max and average crawl depth, your most internally linked pages, and any pages buried more than 3 clicks from the homepage. A sortable orphan-page table lists every orphan with a suggested fix. Data Chat can answer questions about orphan pages and crawl depth, and the Strategy Hub factors both into its technical recommendations.
• Auto-suggested competitors when you have none configured — If your project has a website but no competitors yet, the Competitors page now shows a Suggested Competitors card. Suggestions are discovered from how much your keyword rankings overlap with other sites, each with a 0–100 overlap score, and you can add any of them with one click. Once you have competitors, suggestions tuck behind a Discover more toggle so you can keep finding new rivals.
• Spot and disavow toxic backlinks — The Backlinks page now shows a spam and toxicity score from 0 to 100 for every referring domain, with a green/amber/red traffic light so risky links are obvious at a glance. A new Toxic filter tab narrows the list to high-spam domains, each row has a checkbox to mark a domain for disavowal, and an Export disavow.txt button downloads a ready-to-upload Google-format file. Works in agency mode scoped to the selected client.

Improved

• Backlink scans pull far more data by default — Scans and the Data Chat backlink tools now fetch up to 200 referring domains and 100 anchor texts, up from 50 and 20, for a much fuller picture of your link profile.
• A failed competitor scan no longer looks like a real "nobody's talking about them" result — When a competitor sentiment scan failed partway, the brand-versus-competitors comparison used to persist all-zero neutral sentiment, which read like a genuine finding. A failure is now detected and reported honestly, so the UI shows "Scan failed — retry" instead of fake zeros. A genuine zero-mentions result is still treated as a real, empty finding.
• The SEO scan is clear when you have no competitors configured — The competitor sub-tasks (keyword gap, backlink gap, competitor overview) used to silently complete with nothing, which looked like the analysis ran and found nothing. They now say explicitly that no competitors are configured and point you at where to add them.
• The Strategy Hub stops inferring a competitive position from missing data — With no competitors configured, it now explicitly calls out the data gap instead of speculating about your standing. When a competitor sentiment scan has failed, the social intelligence analysis says the data may be stale or incomplete rather than treating zeros as a result.
• The Strategy Hub now knows how old your data is — Each analysis is told how recently every underlying source was refreshed, anything older than 7 days is flagged, and recommendations leaning on stale data are caveated with a suggestion to re-run the relevant scan. A compact "Data as of [date]" line with a hover breakdown of every source's age now sits in the analysis header.

Fixed

• You're no longer charged for competitor scans that failed — A failed competitor sentiment scan used to still deduct a scan from your wallet and report "Scan complete". Both the single-competitor scan and the scan-all action now skip the charge for any scan that didn't succeed and return a clear "Scan failed — retry".