All updates

Campaign and signal reports, grounded Capability suggestions, clearer feedback evidence

In this edition

Review campaign performance in Reporting

The Campaigns page has a Reporting tab with a sticky switcher for active campaigns. The diagnostic report view includes:

Read more
  • Origin filters for Outbound, Inbound, Warm, and All.
  • Eight KPI tiles with 30-day sparklines, a stage-by-stage funnel, diagnostic callouts, and an origin-by-ICP-fit matrix.

Engaged Back and conversion rates use an outreach denominator that includes members who advanced directly to later stages. Replies count for open and converted members; silent connection accepts do not count as inbound dialogue.

See a contact’s activity across campaigns

Kanban cards and pipeline rows show the total campaign count when a contact is enrolled in two or more active campaigns, plus an origin badge for Inbound or Warm members. The member detail panel links to the contact’s other active campaigns and shows touchpoints across campaigns in the workspace, with badges marking interactions from other campaigns.

Read more

The pipeline table populates the first 1,000 members immediately upon the first page response and renders them in 75-row increments as you scroll or open a deep-linked member. Search, filters, and column sorting still evaluate the full campaign member list.

Review better-grounded Capability suggestions

When reviewing suggestions for a Change, direct source evidence can retain relevant Memory, Chat, or Sitemap matches even when domain or repository heuristics conflict. Existing equivalents are compared by behavior, including their descriptions. Under current scope and eligibility checks, a clearly evidenced new ability can be suggested alongside existing ones.

Read more
  • Suggestions distinguish changed behavior from maintenance, ownership, and incidental dependency use. A nearby new workflow alone does not justify linking an existing Capability; the source must establish a change to that specific behavior.
  • Suggestions distinguish a new workflow from a change to an existing Capability. Business actions can be recognized as product workflows when supported by evidence; suggested parents follow the workflow owner.
  • Specific workflows retain their normal steps and result views rather than generating duplicate child Capabilities. Grounded short Sitemap quotes and dashboard date-filter evidence can retain the correct existing Capabilities.

New suggestions still require review. When the Capability tree is nonempty, a new suggestion needs a current parent; an empty workspace can receive its first root.

Explore signal reports and customer context

Signals Reporting has dedicated reports for Customer Jobs (JTBD), Product Areas & Demand, Problems & Target Gains, and Inbox Triage & Rot.

Read more
  • Customer Job details include a filtered Capability view and related Problems with their Done When target gains.
  • Problem details let product managers define or edit a Done When criterion, link or unlink Customer Jobs, and inspect Classification & Synthesis status in the technical tab. In the Signals Problems workspace, product managers can filter Problems by JTBD synthesis coverage.

Contacts and Signals Reporting classify and filter speakers on two separate axes: Voice Relationship (Customer, Trial / Evaluator, Prospect (ICP), Market Voice, Peer / Advisor, Internal (Founder / Team), or Bot / AI Agent) and Buying Role (Economic Buyer, Champion, Practitioner, or Non-Buyer / N/A). Scope presets include Market Only, Customers & Trials, Prospects & Market, Peers, Internal, and Bots.

Navigate reports from the Bridge tab

The Bridge tab brings Risk & Product Health, Signals & Demand, and active Campaigns into one navigation rail. The risk group includes Competitor Risks and Product Health & Quality reports.

Find Granola notes moved into selected folders

When Granola grants access to or updates a note, that note is imported into Feedback if it belongs to the selected folders for its connection. Manual sync can also find older notes moved into those folders.

Follow interview evidence extraction and review

Feedback details show the active evidence-extraction progress while classification is processing, rather than displaying “No Evidence Identified.” Multi-turn interview transcripts are split into chronological verbatim evidence items of 1–3 sentences, up to 380 characters each. Speaker profiles are deduplicated when an email-derived attendee name lacks spaces compared with the spaced transcript speaker name.

See how Outcomet works.

Follow a customer problem from feedback to a product decision.

See how it works