Detection rules are the core of Revenue Observability. They're the logic that decides: this deal has a blind spot, and here's what it is. But building good detection rules is harder than it sounds.
The most common mistake? Making rules too sensitive. If every deal triggers an alert, nobody pays attention to any of them. That's alert fatigue, and it kills adoption faster than anything.
The anatomy of a good detection rule
Every effective detection rule has four components:
- 1A specific signal (not a vague condition). 'Champion hasn't replied in 5+ days' is specific. 'Deal engagement is low' is vague.
- 2Cross-stack correlation. The signal must come from connecting two or more data sources. If a single tool can detect it, it's not a blind spot.
- 3Business context. A champion going silent on a $50K deal in discovery is different from a champion going silent on a $500K deal in negotiation.
- 4A clear action. The alert must tell the rep or manager exactly what to do. 'Check in with champion' isn't enough. 'Re-engage Lisa Park via her preferred channel (Slack) - last response was 8 days ago on email' is actionable.
See your pipeline's blind spots
Customer City connects to your revenue stack and surfaces the silent failures killing your deals.
Start with these 5 rules
If you're just getting started with detection rules, begin with these five. They catch the highest-impact blind spots with the lowest false-positive rate:
- 1Silent Champion: No reply from primary contact in 5+ days while deal is in Negotiation or later
- 2Unopened Proposal: Contract/quote sent but not opened within 3 business days
- 3Missing Decision-Maker: No VP+ contact on any call for deals over $100K in Proposal stage or later
- 4Stuck Stage: Deal in same stage for 30+ days with activity but no progression
- 5Single-Threaded: Only one contact engaged on deals over $50K
Tuning for your team
Every team is different. Enterprise sales cycles are longer than mid-market. Some industries have naturally slower response times. The key is to start with the defaults, measure false positives for 2 weeks, then adjust thresholds. Customer City makes this easy with per-rule sensitivity controls.