Sudden pipeline contraction = mass-loss event or stage-mover dropped to closed_lost.
At a glance
24-hour anomaly detector for Total Pipeline Value. Fires when the day-over-day pipeline change exceeds 2 standard deviations from the trailing 30-day baseline. Surfaces the specific deals that caused the contraction so sales leadership can intervene the same day.
Calculation
Calculated automatically from your HubSpot data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.Worked example
The same B2B SaaS portal (Total Pipeline = $1.45M). Trailing 30-day daily baseline:
Reading on 14 Apr 26 morning, the alert table shows:
Five things this picture reveals:
- Single-deal closedlost ($144,000) drove most of the breach. Northwind moved at 80%-probability stage to lost; that is not a procurement delay, it is a competitor win or a budget-cut. The deal owner’s call notes are the next read.
- The threshold is set tight (2σ) on purpose. Most B2B portals need to react same-day to single-deal slips at the late-stage. Loosening to 3σ would silence ~99% of alerts but miss the operationally-actionable 75% of single-deal slips.
- Amount-reduction event (GlobalTech, -$58.5k) is a softer signal. The customer asked for a smaller scope or a longer payment plan. Not a loss yet but a leading indicator. Pair with
Average Deal Cycleto see if cycle is also stretching. - The Acme closedate-push to August adds zero to the unweighted 24-hour delta because it stayed open. But it removes $X from
Deals Closing This Month, where it triggers a separate alert. The two alerts together tell a fuller story. - Pipeline Drop alert recovery is the inverse signal. When a previously-lost deal is re-opened (rep moves Northwind from closedlost back to closedate=June), the 24h delta swings positive. Vortex IQ does not fire an alert on positive moves but does log them in the audit history.
Sibling cards merchants should reference together
Pipeline Drop is a real-time alert; pair with the underlying stock and rate cards:Reconciling against the vendor’s own dashboard
Where to look in HubSpot: The HubSpot-native equivalent is the Pipeline Snapshot report, but it is a daily snapshot rather than a real-time alert:HubSpot → Reports → Sales analytics → Pipeline snapshots (compare today vs yesterday) HubSpot → Sales → Deals → Activity feed (filter to “Stage changed” and “Amount changed” in last 24h)The Activity feed shows the same deal events that drive this alert. Why our number may legitimately differ from HubSpot’s:
Cross-connector reconciliation:
Known limitations / merchant FAQs
The alert fired but I do not see anything wrong, what happened? Check the alert table for the contributing events. Most-common false positives: (a) a rep cleaned up old stale deals (bulk-marked as closedlost); (b) a deal-creation script ran and immediately corrected anamount field by 50%+; (c) a multi-currency portal had an unusually large FX move on the home-currency conversion. (a) and (b) are housekeeping; (c) is rare.
Why is the threshold 2σ and not a flat percentage?
Because absolute pipeline volatility scales with portal size. A 50k+ daily moves; a $50k-pipeline portal does not. A 2σ threshold adapts to each portal’s normal noise band. Flat percentage thresholds either fire constantly on small portals or never on big ones.
Can I tune the threshold?
Yes, configurable per portal in Nerve Centre settings. 1.5σ catches more (noisier); 2.5σ catches fewer (only the largest moves). 2σ is the default for most B2B portals.
Why is my deal pipeline number different from my dashboard?
This is the alert (delta) view, not the stock view. The stock view is Total Deal Pipeline Value. Use both: stock for “what is open”, alert for “what changed last 24h”.
Multi-portal aggregation, can I see drops across portals?
No. One card per portal because each portal has a different pipeline-volatility distribution; aggregating would dilute high-signal alerts on small portals.
Sequence enrolment vs send timing, does it affect this?
No. Sequences are 1-to-1 outreach; deals are independent objects. A sequence-driven outreach that converts to a deal contributes positively to pipeline only at deal-creation, not enrolment.
Today-volatility, why does this sometimes fire mid-day and recover before close-of-business?
A rep mistakenly moves a large deal to closedlost (clicked the wrong stage), the alert fires, the rep notices and reverts the move. The 24h delta swings back. We do not auto-suppress; the audit log records both events.
List-segment lag, can it cause an alert?
Indirectly. A list-trigger workflow that mass-creates deals or mass-stages deals based on list membership can cause a 2σ swing if the membership refresh batches a large group. This is rare and usually follows a list-import event.
Lifecycle-stage backfill, does it touch this card?
Only if the backfill triggers deal-creation workflows. Lifecycle changes alone do not move deals; they move contact properties. The cross-trigger to deals is via dedicated workflow rules.
Why is the alert 24h-rolling and not snapshot-at-midnight?
Because B2B sales teams need same-day visibility. Snapshot-at-midnight introduces a 0-23 hour delay; 24h rolling catches breaches within 5 minutes. The trade-off is slightly higher noise (timezone-boundary effects), which we accept.