Customer Support

How to watch for churn risk accounts as Small Customer Success Teams

Customer SupportFor Small Customer Success Teams2 apps12 steps~24 min to set up

Your usage dashboard updates once a week, which means by the time you see a red flag — a customer who logged in twice last month instead of twenty times — they've already had three bad experiences you didn't know about. You're managing 250 accounts across three people. HubSpot has the contact records. Intercom has the support tickets. Your product analytics tool has the usage data. None of them talk to each other. So churn risk lives in your head: a gut feeling from the last call, a sticky note about the account that's been 'quiet lately.' Gainsight would fix this, but it costs more than your combined CS salaries and requires a CS-ops person to set it up. You don't have either.

Build this on Starch → See the apps used

Customer SupportFor Small Customer Success Teams2 apps12 steps~24 min to set up

Outcome

What you'll set up

A live account health dashboard that pulls HubSpot contact data, Intercom ticket history, and Gmail thread recency into a single view — scored by risk level, not sorted by whoever emailed last

A weekly automated churn-risk digest delivered to your Slack, flagging accounts with declining engagement, open tickets older than 7 days, or no contact in 30+ days — before your Monday standup

A trigger-based alert system that fires when a specific account hits a risk threshold, so you can intervene with a personal outreach the same day instead of at the quarterly review

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

Data sources & config

Starch syncs your HubSpot data (contacts, companies, deals, owners) on a schedule. Starch syncs your Gmail on a schedule for thread history and last-contact-date signals. Connect Intercom from Starch's integration catalog — the agent queries it live when your health dashboard or automations run to pull open ticket counts and ticket age. Connect Slack from Starch's integration catalog for digest delivery.

Prompts to copy

Build me an account health dashboard that pulls from HubSpot (contacts, companies, deals) and scores each account by churn risk. Risk signals I care about: no activity in HubSpot in 30+ days, open Intercom tickets older than 7 days, no email exchange in the last 14 days based on my Gmail. Show me a ranked list with a risk score, last contact date, open ticket count, and the last thing we talked about.

Every Monday at 8am, generate a churn-risk digest from my account health dashboard. Surface the 10 highest-risk accounts, summarize why each is flagged, and post it to our #cs-weekly Slack channel with a one-line suggested action for each account.

Set up an alert: if any account's risk score crosses 70, immediately draft a check-in email from my Gmail to the primary contact and flag the account in HubSpot with a 'churn risk' tag. Show me the draft before sending.

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 owner assignments on a schedule, so every account's CRM record is always current in your health dashboard.

2 Connect Gmail — Starch syncs your email threads on a schedule and uses last-send and last-reply dates as a direct signal for 'has anyone actually talked to this customer recently.'

3 Connect Intercom from Starch's integration catalog — the agent queries it live to pull open ticket counts and ticket age for each account whenever your health score runs.

4 Connect Slack from Starch's integration catalog so Starch can post your weekly digest and account alerts to the right channel without you having to log in and copy anything.

5 Start with the CRM app from the Starch App Store, then customize it: tell Starch to add a churn-risk score field, a last-contact-date field pulled from Gmail, and an open-ticket count pulled from Intercom.

6 Define your risk scoring logic in plain language — for example: 'an account is high risk if there's been no HubSpot activity in 30 days AND an open Intercom ticket older than 7 days, or no email exchange in 14 days AND the deal stage hasn't moved in 60 days.'

7 Tell Starch to build a ranked dashboard view sorted by risk score descending, with columns for risk level, account name, last contact date, open tickets, ARR, and the last email subject line.

8 Set up the Monday 8am automation: Starch pulls the current health scores, compiles the top-10 at-risk list, formats a digest with one suggested action per account, and posts it to Slack before your standup.

9 Set up the real-time alert trigger: when any account's risk score crosses your defined threshold, Starch drafts a check-in email using that account's context from HubSpot and Gmail, stages it for your review in the Email Triage app, and tags the account in HubSpot.

10 Review the first week's output with your team and refine the scoring weights — if 40 accounts are flagged high-risk on day one, tell Starch to tighten the criteria. The logic lives in natural language, so adjusting it takes thirty seconds.

11 Once the scoring is stable, assign ownership by segment: tell Starch to add a 'CS owner' column that maps to your three reps based on the HubSpot owner field, so every risk flag lands with the right person.

12 Set a monthly review prompt: 'Show me all accounts that were flagged high-risk in the last 90 days and are now low-risk — what changed?' This gives you a feedback loop on which interventions actually worked.

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 churn watch — week of March 10

Sample numbers from a real run

Meridian Logistics ($42,000 ARR)	42,000
Crestwood HR ($18,500 ARR)	18,500
Talbot & Associates ($9,200 ARR)	9,200

Monday morning, 8:02am. Before your standup you have a Slack message from Starch. Meridian Logistics ($42,000 ARR) is flagged critical: no HubSpot activity in 38 days, two open Intercom tickets aged 11 and 14 days, and no reply to the last email you sent on February 21. Starch has already drafted a check-in email to their primary contact pulling in the subject line of that last thread and a note about the unresolved tickets. Crestwood HR ($18,500 ARR) is flagged high-risk: usage data isn't in Starch directly, but their Intercom ticket volume has doubled over the last 30 days and they haven't had a call logged in HubSpot since January. Talbot & Associates ($9,200 ARR) is flagged medium: last email exchange was 19 days ago, no open tickets, but their deal stage hasn't moved in 45 days and renewal is in 6 weeks. Your team reviews the three flags in four minutes, approves the Meridian draft with one edit, and assigns Crestwood to your mid-market rep for a same-day call. None of this required logging into HubSpot, Intercom, and Gmail separately to build the picture yourself.

Measurement

How you'll know it's working

Churn rate by cohort (monthly, trailing 90 days)

Average days to first response on at-risk account alerts

Number of high-risk accounts resolved to medium or low within 30 days of flag

Open ticket age by account segment (target: zero tickets older than 7 days for accounts over $20k ARR)

Renewal forecast accuracy vs. actual — how many flagged accounts churned that weren't on your risk list

Comparison

What this replaces

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

Gainsight

Purpose-built for exactly this workflow but starts at six figures annually, requires a CS-ops person to configure health scores and playbooks, and assumes a team that has time to run inside it daily — not a 3-person team covering 250 accounts with no admin headcount.

ChurnZero

Strong churn-scoring and automated playbooks, but pricing and implementation overhead are built for a dedicated CS team, and you'll spend months getting your data model right before seeing any output.

Manually pulling HubSpot + Intercom + Gmail

Free and already available, but the synthesis — who's actually at risk right now, ranked — happens in your head once a week if you're lucky, which is how accounts slip through unnoticed until it's too late to do anything.

Totango

Reasonable mid-market option with usage-score templates, but still requires integration work and a segment model that takes weeks to calibrate, and the out-of-box templates won't match how your team actually thinks about risk.

On Starch RECOMMENDED

One platform — 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 actually read our Intercom ticket data, or does it just connect to HubSpot?

Both. Starch syncs your HubSpot data directly on a schedule. For Intercom, you connect it from Starch's integration catalog and the agent queries it live whenever your health dashboard or automation runs — so ticket counts and ticket age are always current at the moment the score is calculated, not cached from a week ago.

We don't have a product analytics tool inside Starch — how do you handle usage data?

Today, Starch uses HubSpot activity, Gmail thread recency, and Intercom ticket signals as churn proxies. If your product analytics tool is web-based, Starch can automate it through your browser — no API needed — to pull usage metrics into the health score. Or if it has an API, connect it from Starch's integration catalog and the agent queries it live. If it's a tool with no web interface and no API, that's a genuine gap worth naming: Starch can't pull data that isn't reachable.

Can Starch actually send the check-in email automatically, or does a human have to approve it?

Either. You can set Starch to draft the email and stage it for your review in the Email Triage app — which is how most teams start. Once you trust the drafts, you can tell Starch to send automatically for certain risk levels (say, medium-risk accounts) while still requiring approval for anything flagged critical. You define the rule in natural language.

Is this SOC 2 certified? We have enterprise customers who will ask.

Starch is not SOC 2 Type II certified today. If that's a hard requirement for your enterprise accounts, it's worth knowing upfront. It's on the roadmap, but we won't claim it until the audit is complete.

We track churn risk differently than the standard 'no login in 30 days' model — can we change the scoring logic?

Yes, and this is where Starch is different from a tool like Gainsight where the health score model lives in configuration screens. You describe your scoring logic in plain language — 'high risk means two or more open tickets older than 7 days AND no call logged in 45 days, OR ARR over $20k AND no QBR scheduled this quarter' — and Starch builds the scoring rules to match. If you want to change a weight or add a new signal next month, you describe the change and Starch updates it.

What happens if HubSpot has bad data — contacts assigned to the wrong owner, deals in the wrong stage?

Your health score will reflect whatever is in HubSpot. Starch can help surface the inconsistencies — you can ask it 'show me all accounts where the owner field is blank or the deal stage hasn't been updated in 60 days' — but it won't fix underlying CRM hygiene problems automatically. Garbage in, garbage out is still true here.