Sales & CRM

How to set up your first crm as Small RevOps Teams

Sales & CRMFor Small RevOps Teams3 apps12 steps~24 min to set up

You're two people holding together the data plumbing for 30 reps. HubSpot or Salesforce is the official source of truth, but the actual source of truth is a Sheets tab someone last updated three weeks ago. Apollo sequences live in one place, LinkedIn touches are invisible, inbound form fills are in a third tool, and nobody agrees on what 'opportunity' means. Every Monday you spend 45 minutes copy-pasting pipeline snapshots into slides so a VP can read numbers you already know are wrong. Setting up a new CRM view or hygiene rule means filing a ticket with yourself and doing it at 7pm.

Build this on Starch → See the apps used

Sales & CRMFor Small RevOps Teams3 apps12 steps~24 min to set up

Outcome

What you'll set up

A single connected dataset pulling from HubSpot (or Salesforce/Pipedrive), Apollo, LinkedIn, and Gmail — no more tab-switching to triangulate where a deal actually stands

A custom CRM surface with hygiene rules, territory logic, and attribution fields your reps actually fill in — described in plain language, not configured by clicking 47 dropdowns

An automated forecast-prep view that assembles the weekly pipeline snapshot so you stop building slides by hand every Monday morning

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

Sales Agent CRM CRM Email Agent

Data sources & config

Starch syncs your HubSpot contacts, companies, deals, and owners on a schedule — that's the core dataset. Salesforce or Pipedrive connect from Starch's integration catalog; the agent queries them live when your CRM app runs. Apollo contacts and sequences connect from Starch's integration catalog the same way. Starch syncs your Gmail messages on a schedule so the CRM can show real email activity per contact. LinkedIn enrichment runs through browser automation — no LinkedIn API needed.

Prompts to copy

Connect our HubSpot deals and contacts. Build a CRM view that shows every open opportunity with: owner, stage, last activity date, next step, and whether we've had a Gmail thread with the contact in the past 14 days. Flag anything where last activity is older than 21 days and the stage hasn't changed.

Add an Apollo sequences column that shows which sequence each contact is currently enrolled in, and a LinkedIn enrichment field that auto-updates job title and company when a contact's profile changes.

Build me a weekly forecast prep view: group open deals by rep and territory, calculate weighted pipeline using our stage probabilities (Disco 10%, Proposal 30%, Negotiation 70%, Verbal 90%), and generate a one-paragraph narrative summary I can paste into the Monday deck.

Set up a hygiene automation: every Sunday night, scan all deals last modified more than 30 days ago and still marked open. Slack me a list sorted by deal value so I can triage before the Monday call.

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

Walkthrough

Step-by-step

1 Connect HubSpot in Starch — Starch syncs your HubSpot contacts, companies, deals, and owners on a schedule and stores them in a typed dataset you can query from any app you build.

2 If your org also runs Salesforce or Pipedrive, connect those from Starch's integration catalog; the agent queries them live so you can pull in any field that lives there without a separate export.

3 Connect Apollo from Starch's integration catalog so sequence enrollment, account data, and contact status are queryable alongside your CRM records.

4 Connect Gmail — Starch syncs your messages on a schedule so the CRM surface can show last-touch email date per contact and flag gaps without you manually checking threads.

5 Start from the Sales Agent CRM template in the Starch App Store and describe your actual pipeline: the stages your team uses, the fields reps actually fill in, and the territory groupings the CRO cares about this quarter.

6 Add a LinkedIn enrichment step by telling Starch: 'For every contact where job title or company hasn't been verified in 90 days, check their LinkedIn profile and update the record.' Starch automates this through your browser — no LinkedIn API required.

7 Build the hygiene rule automation: describe the staleness threshold (no activity in 21 days, stage unchanged), the deal value floor worth surfacing, and where you want the alert — Starch can Slack you a sorted list every Sunday night.

8 Build the territory model by describing it: 'Group deals by rep, assign territory by the state field on the company record, and flag any deal where the assigned rep's territory doesn't match the company location.' Starch generates the logic; you review and adjust in plain language.

9 Build the forecast prep view: give Starch your stage-to-probability mapping and tell it what a useful Monday narrative looks like. It assembles weighted pipeline by rep and territory and writes the summary paragraph.

10 Wire the Email Agent to your inbox so inbound deal-related emails are triaged, threaded to the right contact record, and surfaced as follow-up tasks when a rep hasn't responded in 48 hours.

11 Add any attribution logic you need — for example, 'tag each deal with the first Apollo sequence the contact entered, the last LinkedIn touch if one exists, and the inbound form source from HubSpot' — and Starch builds the attribution view across all three data sources.

12 Publish the forecast view and hygiene alert as internal apps your reps and CRO can access directly, so the Monday-morning data run stops living in your personal Starch workspace and becomes a team resource.

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 Close-Month Hygiene Sprint — Mid-Market Segment

Sample numbers from a real run

Deals flagged stale (no activity 21+ days, open stage)	14
Total pipeline at risk (weighted)	287,000
Deals with no Gmail thread in past 14 days	9
Contacts with outdated LinkedIn job title (90+ days)	31
Territory mismatches surfaced by hygiene rule	4

It's March 3rd and the CRO wants a clean pipeline number before the board call on the 7th. Normally this is a four-hour manual exercise: export HubSpot, cross-reference Apollo, check Gmail threads, rebuild the weighted calc in Sheets, write the narrative in a doc. This time, Starch already ran the Sunday hygiene scan. You wake up to a Slack message listing 14 stale deals sorted by weighted value — the top one is a $62,000 Negotiation-stage deal where the last HubSpot activity was February 6th and there's been no Gmail thread since January 29th. You open the forecast view, which has already calculated $287,000 in weighted pipeline across the mid-market segment using the stage probabilities you set in December. The LinkedIn enrichment pass flagged 31 contacts whose titles changed in the past 90 days — three of them are your champions at accounts in active deals, two of whom left their companies. The territory check surfaced four deals where the assigned rep's territory label doesn't match the company's state, which explains the quota attainment discrepancy the CRO flagged last week. You spend 40 minutes actually fixing problems instead of finding them.

Measurement

How you'll know it's working

Stale opportunity rate: percentage of open deals with no activity in 21+ days, tracked weekly

Data completeness score: percentage of open deals with all required fields populated (next step, close date, stage-appropriate fields)

Attribution coverage: percentage of closed-won deals with a traceable first-touch source across HubSpot, Apollo, LinkedIn, and inbound

Forecast accuracy: actual close-month revenue vs. weighted pipeline as of the 1st of that month

Time to Monday packet: hours spent assembling the weekly forecast call materials (the KPI you're trying to drive toward zero)

Comparison

What this replaces

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

HubSpot + native reporting

HubSpot's reporting is solid for what it's designed to report, but building a hygiene rule that spans Apollo sequences, Gmail activity, and LinkedIn profile staleness requires either Operations Hub at enterprise pricing or a lot of manual joins — Starch lets you describe that cross-source logic in plain language without a workflow builder.

Salesforce + custom objects + Reports & Dashboards

Salesforce can store and report on almost anything, but configuring custom objects, roll-up summaries, and cross-object reports for a 2-person RevOps team is a multi-week admin project; Starch builds the equivalent surface from a natural-language description in minutes.

Google Sheets + manual export cadence

Sheets is free and flexible, but it has no memory — every Monday someone has to export, paste, recalculate, and re-check; Starch replaces that recurring manual run with a scheduled app that does it automatically.

Apollo + Apollo dashboards

Apollo's native dashboards are useful for sequence analytics, but they don't know about your HubSpot stage data or your Gmail threads, so you're still triangulating across tabs — Starch puts Apollo sequence data alongside your CRM and email data in one queryable surface.

On Starch RECOMMENDED

One platform — sales agent crm, crm, email agent 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

We're already on HubSpot. Does Starch replace it?

No. Starch syncs from HubSpot on a schedule and lets you build views, hygiene rules, and automations on top of your HubSpot data. Your reps keep working in HubSpot. You use Starch to do the ops work that HubSpot's native tools make you do manually or pay for at a higher tier.

What about Salesforce? We have some reps on SFDC, some on HubSpot.

Connect Salesforce from Starch's integration catalog — the agent queries it live when your apps run. You can build a unified CRM view that pulls from both HubSpot (synced on a schedule) and Salesforce (queried live), so you finally have one place to look instead of two.

Can Starch write back to HubSpot — update a field, change a stage?

The HubSpot scheduled sync is read-only today. Starch surfaces what's in HubSpot and lets you build ops tooling on top of it. Write-back is on the roadmap; for now, Starch tells you what needs to change and you or a rep makes the update in HubSpot directly.

How does the LinkedIn enrichment actually work? LinkedIn locks down their API.

Starch automates LinkedIn through your browser — no LinkedIn API required. It opens a browser session, checks the contact's profile, and pulls the current job title and company back into your CRM record. It's the same thing a human would do; Starch just does it for 31 contacts while you're asleep.

Is Starch SOC 2 certified? We're going to have deal data in here.

Not yet — Starch is not SOC 2 Type II certified today. That's worth knowing before you put anything that triggers a vendor security review. If your deal data is sensitive enough that your security team will ask about it, that's a real conversation to have before you connect production systems.

The CRO keeps changing the territory model. How hard is it to rebuild the view?

Tell Starch what changed — 'move these three states from the East rep to the Central rep, recalculate all open deals' — and it rebuilds the view. You're not editing a formula or re-mapping a custom object. This is the reason a 2-person team can run this without a full-time admin.

We use QuickBooks for billing. Can we see revenue vs. pipeline in one view?

Yes. Starch syncs your QuickBooks invoices, payments, and billing data on a schedule. Note: QuickBooks report views like P&L are temporarily disabled while an upstream fix is in progress, but invoice- and payment-level data syncs normally — which is usually what you want for a pipeline-to-revenue reconciliation anyway.