The churn radar nobody builds

If you own renewals at Datadog, this is the board you open on a Monday morning before you decide who to call. It lists 59 customer accounts grouped by the quarter their contract comes up. Each is scored red, amber or green, and each score opens into the evidence behind it. I watch four things per account: the champion or exec sponsor leaving, hiring that implies someone is building or buying a replacement for us, layoffs, restructuring or acquisition, and a new leader arriving over the function we sit in. Where a new leader is in the seat, the card names them, shows their verified work email, and opens a brief with their dated career history and what their company is dealing with this quarter, so the first message isn’t cold.

What it replaces is the routine of checking each account by hand. That means LinkedIn for who left, a job board for what they’re hiring, a news search for layoffs or a deal, and then a spreadsheet you have to remember to update. Working out who now owns the function, finding an email for them and reading up on them is a separate errand per account. Here the trail sits under each score, and the person, the address and the brief sit on the card.

The workspace held none of this, so I started with trayo_import_accounts, which created the 59 accounts from the customer list. trayo_create_signal defined the four checks, and five trayo_run_discovery runs covered all 59 accounts. Discovery returned zero events, which I cover in the last paragraph. To get the leadership side onto real data, I ran trayo_search_job_changes with the account as destination and seniority set to director. It returned dated arrivals at 13 accounts, including John Hunt starting as Chief Digital and Technology Officer at Marks and Spencer, Pranathi Pulimamidy as Sr Director of Platform Engineering at Autodesk, and Amit Thawani as CIO Consumer Bank at Lloyds. Six trayo_search_stakeholders calls gave me the people who own our function at those accounts, and six trayo_research_company calls and six trayo_research_person calls produced the briefs. trayo_enrich_emails returned 14 work emails, all marked verified. Every person, photo, logo, title, date and address in the app came from one of those calls.

The part you should know first: discovery found nothing. All five runs settled clean, with no blocked signals, no errors, and every account finished, and each wrote eventsFound: 0. To rule out my own signal wording, I re-ran with a 365-day lookback and two deliberately wide-open probe signals (“any news about this company” and “any engineering job posting”) against Adobe, PayPal and ServiceNow, plus the workspace’s own pre-built signals. All came back zero. trayo_whoami reports monitoring.enabled: false on this workspace, and Trayo’s guidance is that a workspace the standing scan doesn’t cover looks exactly like a quiet week. So the champion-departure and replacement-hiring checks are wired up, ran, and found nothing, and the board says that, with run receipts on the /method route. Because you asked for the champion-left check to go through the job_change signal in discovery, I kept it there and didn’t swap in job-changes/search. That means the “who replaced our champion” flow is shown on the leadership-arrival side instead: who owns it now, their email, the brief. It’s the same three steps, on real data. One acquisition, Autodesk/MaintainX, reached the board from company research rather than the news signal, and the card labels it as the weaker source. Renewal date, ARR, seats and account owner aren’t in the Trayo API, so they come from a stand-in customer list built from the real accounts, which you can download at /customer-list.csv. The draft intro email, the Salesforce push and the #renewals post are previews. They show what would have been sent and say so.