Usage-Based Health Scoring
Roll product usage from Snowflake into a Gainsight health score on a schedule and alert the account team in Slack when an account trends down.
The problem
Usage data sits in the warehouse while the customer team works from anecdotes and quarterly check-ins. By the time a drop in adoption is noticed, the renewal conversation is already defensive. Manual exports and one-off dashboards do not scale across the whole book of business.
How it runs on Neblex
A cron schedule runs the scoring flow on your cadence. An arbitrary SQL step queries Snowflake for the usage signals you care about, such as active users, feature adoption, and login recency, and transforms turn them into a health score per account. Built-in Data Tables keep each account and its previous score as state between runs.
The flow updates scores in Gainsight in chunked batches, fanning out across accounts with bounded concurrency. Conditional branching compares each new score with the stored prior value, and when an account trends downward it posts an alert to the owning team in Slack with the drivers behind the change.
Per-step retries and replay keep a flaky API from costing you a scoring cycle. When the scoring logic changes, promotion approvals gate the rollout from dev to prod.
Step by step
Query the warehouse
A scheduled arbitrary SQL step pulls usage signals from Snowflake.
Compute health scores
Transforms weight the signals into a score per account.
Compare against history
A Data Table holds each prior score, so the flow knows direction, not just level.
Update Gainsight
Scores load in chunked batches with bounded concurrency across the account base.
Alert on downward trends
Conditional branching posts a Slack alert to the account owner when a score falls.
Platform capabilities used
- Cron schedules
- Arbitrary SQL on database connectors
- Data Tables for state
- Conditional branching
- Chunked batch processing
- Parallel fan-out with bounded concurrency
Common questions
Can the flow run inside our network?
Yes. On-Prem Workers run flows inside your own network over an outbound-only connection, and for those flows business data is processed inside your network. That includes the warehouse queries this use case depends on.
How do we tune the scoring model safely?
Scoring lives in transforms and Data Tables, so changes are versioned flow changes, not scripts on a laptop. Environments with promotion approvals gate the path to production, and what-if replay shows how a change would have scored a past run.
What if the Gainsight update fails midway?
Chunked batch processing records checkpoints, and per-step retries absorb transient errors. A failed run can be replayed from the exact step that broke, so one API error does not skip a scoring cycle.
Related use cases
Support-to-CRM Escalation
Escalate high-priority Zendesk tickets into linked Salesforce cases automatically, with full conversation history attached and statuses kept in sync.
MarketingCampaign-to-Pipeline Attribution
Sync Google Ads spend and HubSpot campaign engagement into BigQuery on a schedule so marketing and sales report from the same pipeline numbers.
AgenticCustomer Feedback Synthesis Agent
An agent gathers feedback from Zendesk tickets, app store reviews, and G2 reviews, finds recurring themes, and sends a cited digest to leadership.
Want this running on your stack?
Neblex Integration Fabric is in beta: full platform, free while in beta. Bring this workflow and we will map it to your systems.