HubSpot portal audit, cleanup and ongoing admin for a SaaS company on Salesforce that catches usage drops before churn

What The Early-Warning Portal Rebuild delivers

A 200+ person B2B SaaS company runs Salesforce as its CRM of record and HubSpot for marketing, lifecycle email and some customer communication, connected by the native sync. Customer success keeps learning about churn risk too late: an account's active users and key feature usage slide for weeks, nobody sees it, and the first signal is a cancellation notice or a renewal call that goes badly. The root cause is not a missing dashboard. Usage data lives in the product database or warehouse and never reaches the systems CSMs work in. Workspaces are not reliably tied to accounts. Duplicate companies split activity. Lifecycle stages stop at 'Customer'. Routing rules send nothing when a customer goes quiet. This build gives the RevOps team a clean, governed HubSpot portal and Salesforce sync. Product usage and health fields are mapped to the right account. Customer lifecycle stages run past the sale. Alerts and tasks reach the account owner when usage drops. Reporting covers retention and expansion as well as pipeline. After launch, a HubSpot solutions partner keeps it running: sync errors, new properties, workflow changes and quarterly re-audits. Once the data is trustworthy, an AI agent can sit on top. A custom AI agent can watch usage trends and draft at-risk account briefs for CSMs, and HubSpot Breeze can summarize records and support customer-facing follow-up.

Custom build

Salesforce stays the CRM of record for accounts, contracts and renewals, so the build does not try to move CS into HubSpot. It fixes HubSpot as the engagement and lifecycle layer and makes the sync between the two reliable. HubSpot does lifecycle automation, customer email, list-based suppression and Breeze record summaries well, so those stay there. Usage data comes from the warehouse through n8n or a reverse ETL sync, because neither CRM should calculate product metrics. Ownership, tasks and renewal pipelines live in Salesforce, where CSMs already work. A custom AI build is the right fit for the audit and the usage-monitoring agent, because it has to read both systems and the warehouse together. No single native tool sees all three.

Signs it is time

• CSMs keep finding out about declining accounts from the renewal conversation or a cancellation, not from the CRM
• Product usage sits in the warehouse or a BI tool, but neither Salesforce nor HubSpot shows it on the account record
• The same customer appears as several companies or accounts, so engagement, tickets and usage cannot be read together
• Lifecycle stages and workflows end at closed-won, and there is no owner or process for post-sale risk or expansion
• Nobody trusts retention or expansion numbers because HubSpot and Salesforce disagree on who is a customer and what they pay

Systems and data

• HubSpot portal: settings, users, teams, permissions, properties, workflows, lifecycle stages, lists, forms, reports and connected apps
• Salesforce: accounts, contacts, opportunities, renewal and contract fields, account ownership, plus the HubSpot-Salesforce sync settings and error log
• Product usage data from the product database or data warehouse (workspace ID, active users, key feature events, seat utilization, trial status)
• Billing or subscription system for plan, MRR or ARR, renewal date and payment status
• Support and feedback data (help desk tickets, NPS or CSAT from a tool such as Zonka Feedback)
• Enrichment sources for firmographics and contact refresh, such as FullEnrich or ZoomInfo

Where your team weighs in

• RevOps and CS leadership decide the usage drop threshold and the health score inputs (which features count as core, how many weeks of decline triggers At Risk). The build should not guess what 'healthy' means for this product.
• RevOps owns the sync direction decision for every disputed field, and finance signs off on which system is the source of truth for ARR, plan and renewal date.
• A named RevOps admin reviews ambiguous merges (parent and subsidiary accounts, shared domains, accounts with multiple workspaces) before they run. Bad merges on customer records are hard to undo.
• CS managers decide who owns self-serve and long-tail accounts that drop in usage (pooled queue, digital-only sequence, or account executive), and approve the re-engagement messaging before it goes live.

How we keep it right

• Treat trial, free and paid workspaces differently. A trial workspace going quiet is a conversion problem, not churn, so At Risk logic and CS alerts only fire for paying accounts with an active contract.
• Every workspace must map to exactly one Salesforce account before its usage feeds health scores. Unmapped workspaces go to a review list and are never silently attached by name matching alone.
• No workflow, property or sync rule changes go live without a test on a sandbox or a small list, and an entry in the change log. Critical audit findings get fixed before cosmetic ones.
• Health scores show their inputs on the record, so a CSM can see why an account is flagged. Any AI-drafted at-risk brief cites the actual usage and ticket data it used, and the CSM decides whether to act.
• Marketing automation is suppressed for At Risk accounts. No upsell emails or webinar invites go to a customer whose usage is collapsing until the owner clears the flag.

How you know it works

• Median days from the start of a sustained usage decline to the first logged CSM touch, tracked before and after launch
• Share of paying accounts with a mapped workspace and populated usage and health fields in Salesforce and HubSpot
• Gross and net revenue retention by segment, reported from one agreed definition
• Open audit findings by severity, with Critical and High at zero after the cleanup phase, and sync error count trending down
• Share of churned accounts that were flagged At Risk before the cancellation or renewal date

From pilot to live

1. Phase 1: Audit and first sliceRun the severity-ranked audit across HubSpot and the Salesforce sync. Fix Critical items. Map workspaces to accounts for one segment (for example, the top tier of accounts by ARR) and push the core usage fields for that segment only. CSMs on that segment start seeing usage on the record and getting drop alerts, and they give feedback on threshold accuracy.
2. Phase 2: Full cleanup and post-sale lifecycleFinish permissions, integrations, dedupe and enrichment across the portal. Extend usage feeds and health scores to all paying accounts. Launch post-sale lifecycle stages, the renewal and expansion pipelines, self-serve re-engagement and the retention reporting. Retire legacy workflows and properties flagged in the audit.
3. Phase 3: Governance, ongoing admin and AI on topHand over governance documentation and move to ongoing admin with a HubSpot solutions partner on a weekly, monthly and quarterly cadence. With the data now trusted, add a custom AI agent that reviews weekly usage changes and drafts at-risk briefs for owners, and use HubSpot Breeze for record summaries and customer follow-up drafts, with CSMs approving every outreach.