Customer Support

How to score customer health as Small Customer Success Teams

Customer SupportFor Small Customer Success Teams3 apps12 steps~24 min to set up

Your team of three covers 250 accounts. You know roughly who's healthy and who isn't, but your actual health score lives in a spreadsheet you updated two weeks ago — if you updated it at all. HubSpot tracks deals and contacts but doesn't surface churn signals. Your product analytics tool (PostHog, Mixpanel, or something your eng team built) shows usage data but it's a separate login nobody checks between fires. Intercom has support ticket volume but it's not connected to the renewal date. Every QBR you build takes half a day of copy-pasting. The tools that solve this properly — Gainsight, ChurnZero — start at $30k and require a CS-ops hire to configure. You're scoring health in your head.

Build this on Starch → See the apps used

Customer SupportFor Small Customer Success Teams3 apps12 steps~24 min to set up

Outcome

What you'll set up

A live customer health dashboard that pulls from HubSpot, your support tool, and Gmail — scoring each account on engagement, ticket volume, and deal stage without you manually assembling it

An automated weekly digest that flags accounts whose health score dropped, surfacing the specific signals (no login in 14 days, three open tickets, no email reply) so you can act before the churn conversation happens

A QBR-ready account summary generator that pulls each account's health score, recent activity, open issues, and expansion signals into a structured output you can actually send

The Starch recipe

Apps, data, and prompts

The combination of Starch apps, the data sources they pull from, and the prompts you use to drive them.

Apps used

CRM Sales Agent CRM

Data sources & config

Starch syncs your HubSpot data on a schedule (contacts, companies, deals, owners) and syncs your Gmail on a schedule (messages, threads). Connect Intercom from Starch's integration catalog — the agent queries it live when the health score app runs. PostHog connects from Starch's integration catalog as a live-query source for usage events. Slack connects from the integration catalog to deliver weekly digests.

Prompts to copy

Build me a customer health scoring dashboard. Each row is one of my 250 accounts. Score each account 0–100 based on: days since last HubSpot activity, number of open Intercom tickets in the last 30 days, days since last inbound email in Gmail, and whether there's an active deal in HubSpot. Color-code the score: green above 70, yellow 40–70, red below 40.

Every Monday at 8am, check which accounts dropped more than 15 points in their health score since last week. For each one, pull the three biggest contributing signals and send me a Slack message with the list and a one-sentence summary of why each account moved.

When I give you an account name, generate a QBR summary: health score trend over the last 90 days, last 5 touchpoints from HubSpot and Gmail, open Intercom tickets, current ARR, renewal date, and one expansion signal if there is one.

Run these in Starch → or paste them into your favorite agent

Walkthrough

Step-by-step

1 Connect HubSpot — Starch syncs your contacts, companies, deals, and deal owners on a schedule, so every account's pipeline status is current when your health dashboard refreshes.

2 Connect Gmail — Starch syncs your inbox on a schedule and reads thread history per account domain, so 'days since last email reply' is a real data point, not a guess.

3 Connect Intercom from Starch's integration catalog — the agent queries ticket volume and open issues live each time a health score is calculated, so a spike in support contacts shows up immediately.

4 Connect your product analytics tool (PostHog from the integration catalog, or if you're on something custom, Starch automates the export through your browser — no API needed) to pull login frequency and feature usage per account.

5 Describe the health score model in plain language: 'Score each account based on HubSpot deal activity, Gmail thread recency, open Intercom ticket count, and weekly active usage. Weight them 25% each.' Starch builds the scoring logic and the dashboard.

6 Review the first run — spot-check three accounts you know well (one healthy, one at-risk, one churned last quarter) to validate the scores reflect reality. If a weight feels wrong, tell Starch to adjust it.

7 Set up the weekly Monday digest: tell Starch to flag any account whose score dropped more than 15 points, list the contributing signals, and post it to your team's Slack channel before standup.

8 Build the QBR generator: describe the output format you want (one page per account, health trend chart, last 5 touchpoints, open issues, renewal date, expansion signal), and Starch assembles it from the connected data on demand.

9 Add a 'risk reason' column to the dashboard — for any account below 40, Starch writes one sentence explaining the primary driver (e.g., 'No HubSpot activity in 22 days, 4 open Intercom tickets'). This is what you paste into your weekly standup notes.

10 Each quarter, run the QBR generator for your top 20 accounts by ARR. Review the output, add your own color, and send. Time per QBR drops from 2 hours to 20 minutes.

11 When a health score drops to red, trigger a HubSpot task automatically — Starch writes it as 'Outreach: [Account Name] health score dropped to [X], primary signal: [reason]' and assigns it to the account owner.

12 Review the model quarterly — add new signals as your product evolves (e.g., a new core feature's adoption rate), and update the weights based on which signals actually predicted churn in the previous quarter.

See this running on Starch

Connect your tools, describe what you want, and the agent builds it. Closed beta is free.

Try it on Starch →

Worked example

Q1 2026 Health Scoring Run — 250-Account Book

Sample numbers from a real run

Meridian Logistics (ARR $42,000, renewal April 2026)	31
Trellis Software (ARR $18,500, renewal June 2026)	78
Copperview Manufacturing (ARR $27,000, renewal March 2026)	44
Bellwether Analytics (ARR $61,000, renewal May 2026)	19

On Monday morning your digest arrives in Slack. Four accounts dropped more than 15 points since last week. Bellwether Analytics is the urgent one: score fell from 58 to 19. Starch surfaces the signals — no HubSpot activity in 31 days, 6 open Intercom tickets (up from 1 last month), and no Gmail reply to your CSM's check-in sent 11 days ago. That's $61k ARR with a May renewal. You run the QBR generator for Bellwether and get a structured summary in two minutes: ARR, health trend chart showing the drop, last five touchpoints (two from Intercom, two from Gmail, one HubSpot note from December), and a flag that their main champion changed roles in January. You book a call the same morning with the right context already in hand — not because you dug through three tools, but because Starch connected them and surfaced what mattered. Meridian, at a score of 31 with an April renewal, gets a HubSpot task auto-created: 'Outreach: Meridian health dropped to 31, primary signal: no login in 18 days, 3 open tickets.' Your team works the list in priority order. The spreadsheet you used to maintain manually — the one that was always two weeks out of date — is gone.

Measurement

How you'll know it's working

Accounts at risk (health score below 40) as a share of total ARR — tracked weekly

Average days to first outreach after a health score drop — target under 3 business days

QBR prep time per account — target under 25 minutes from request to send-ready draft

Renewal forecast accuracy — predicted vs. actual renewal rate for accounts flagged as at-risk 90 days out

Ticket-to-churn correlation — percentage of churned accounts that had 4+ open Intercom tickets in the 60 days prior

Comparison

What this replaces

The other ways teams handle this today, and how the Starch version compares.

Gainsight

Purpose-built for customer health scoring but starts around $30k/year, requires a CS-ops person to configure the data model, and assumes you have a dedicated CS platform budget — not realistic for a three-person team.

ChurnZero

Strong churn prediction features but similar pricing floor to Gainsight, a multi-month implementation cycle, and a sales process that's not designed for teams under 10 CSMs.

HubSpot + manual spreadsheet

Free with your existing HubSpot subscription but the health score is only as current as the last time someone updated the spreadsheet, which for a three-person team is never often enough.

Catalyst

Lighter and more SMB-friendly than Gainsight, but still a standalone CS platform with a per-seat cost and an integration setup overhead that lands it in the same category of 'tools that assume CS-ops resources you don't have.'

PostHog + HubSpot + Intercom (no glue layer)

You already have all the data, but it lives in three separate tools with no unified health view — which is exactly the problem Starch solves without requiring you to buy a fourth platform.

On Starch RECOMMENDED

One platform — crm, sales agent crm, founder inbox all running on connected data. Setup in plain English; numbers stay current via scheduled syncs and live agent queries.

Try it on Starch →

FAQ

Frequently asked questions

Does Starch store my customer data or just query it live?

It depends on the source. HubSpot and Gmail are synced on a schedule — that data lives in Starch's database and is what powers the health dashboard refreshing automatically. Intercom and PostHog are queried live from Starch's integration catalog each time your app runs. You're not archiving a historical data warehouse here; Starch is a live operational surface, not a long-horizon analytics store. If you need years of historical ticket data for trend analysis, that's a limit worth knowing upfront.

What if my product analytics tool isn't PostHog?

If it's in Starch's integration catalog (Mixpanel, Amplitude, and similar tools are reachable from the 3,000+ app catalog), the agent queries it live when your health score runs. If your product analytics is a custom internal dashboard, Starch can automate your browser to pull the data — no API needed. Either way, you're not blocked.

Can this replace Gainsight entirely?

For a three-person team managing 250 accounts, probably yes for the core use cases: health scoring, QBR generation, and churn-risk alerting. What Starch doesn't do is the enterprise CS workflow layer — playbook libraries, success plan templates, automated lifecycle campaigns — that Gainsight ships for large CS orgs. If you're at 10 CSMs and 2,000 accounts, you'll eventually want Gainsight. At your current scale, you'd be paying for configuration overhead you'd never use.

Is Starch SOC 2 certified? This matters because I'm connecting my CRM and email.

Starch is not SOC 2 Type II certified today. That's worth knowing before you connect Gmail and HubSpot. If your customers are in regulated industries and your security review requires SOC 2, that's a real constraint. It's on the roadmap; it's not there yet.

How does the health score update — is it real-time or do I need to manually refresh it?

It depends on how you configure it. HubSpot and Gmail data refreshes on Starch's sync schedule, so the dashboard reflects that cadence automatically. Intercom and product usage data are queried live each time the app runs. You can set the dashboard to refresh on a schedule (e.g., every morning at 7am) or run it on demand. It's not real-time second-by-second, but for a weekly health review cadence, the data will always be current when you look at it.

What if I want to add a new signal to the health score later — like NPS survey responses?

Tell Starch to update the scoring model and add the new data source. If your NPS tool is in the integration catalog, connect it and describe the new weight. If it's a tool Starch doesn't formally integrate with, browser automation can pull the data from whatever dashboard you log into. The model isn't locked — you describe changes and Starch rebuilds the logic.