Strategy & Planning

How to run a retrospective or post-mortem as Small RevOps Teams

Strategy & PlanningFor Small RevOps Teams3 apps12 steps~24 min to set up

When a big deal falls through or Q2 misses number, someone has to run the post-mortem. For a two-person RevOps team, that means manually pulling closed-lost reasons from HubSpot, cross-referencing Apollo sequence data to see which touches happened, digging through Gmail threads to find where the deal stalled, and then assembling a deck nobody will read after the meeting. The retrospective itself is just you talking while someone frantically types notes, and the action items get lost in Slack by Thursday. You do this quarterly at best, which means the same pipeline hygiene mistakes repeat for 90 days before anyone names them.

Build this on Starch → See the apps used

Strategy & PlanningFor Small RevOps Teams3 apps12 steps~24 min to set up

Outcome

What you'll set up

A retrospective dashboard that pulls closed-lost deals from HubSpot, sequence data from Apollo, and email threads from Gmail into one view — so you arrive at the meeting with the data already assembled, not a blank slide deck

An automated meeting notes capture for the retro call itself, with action items extracted and assigned by name so nothing disappears into Slack

A knowledge base entry for each retrospective, searchable by quarter, deal type, or rep name, so next quarter's retro starts with last quarter's findings instead of a blank page

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 Meeting Notes Knowledge Management

Data sources & config

Starch syncs your HubSpot data on a schedule — contacts, companies, deals, and owners — and syncs Apollo contacts, accounts, and sequences on a schedule. Gmail is synced on a schedule for thread-level deal context. Notion can be connected from Starch's integration catalog if your team already documents there; the agent queries it live. Salesforce and Pipedrive are reachable from Starch's integration catalog and queried live if HubSpot is not your primary CRM. The Knowledge Management app stores retrospective outputs natively inside Starch.

Prompts to copy

Build me a closed-lost retrospective dashboard showing all deals lost in the last 90 days, grouped by lost reason, with the assigned rep, deal size, stage at loss, and last activity date. Pull from HubSpot and Apollo.

For each closed-lost deal, show me the Apollo sequence the contact was enrolled in, how many touches were completed before they went dark, and whether there's a Gmail thread in the last 30 days.

Capture today's post-mortem meeting, extract all action items with owner names, and save a summary to Knowledge Management under 'Retrospectives / Q2 2026'.

Search Knowledge Management for all retrospectives tagged 'closed-lost' and summarize the top 3 recurring themes across the last 4 quarters.

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

Walkthrough

Step-by-step

1 Connect HubSpot as a scheduled-sync provider. Starch syncs deals, contacts, companies, and owners on a schedule — this is your deal history for the retro, not a manual export.

2 Connect Apollo so Starch syncs your sequence data. You'll be able to see which sequence each closed-lost contact was in, how many steps they completed, and when they last responded.

3 Connect Gmail so Starch syncs recent threads. When a deal closed-lost, you want to see the last email exchange — was it a ghosting situation, a pricing objection, a competitor win?

4 Tell Starch: 'Build me a closed-lost retrospective dashboard for the last 90 days, grouped by lost reason, with deal size, rep, stage at loss, and last Apollo sequence touch.' This becomes your pre-meeting data view, ready before the call starts.

5 Before the retro meeting, open the dashboard and flag deals worth discussing: outliers by deal size, patterns by lost reason, reps with disproportionate losses at a specific stage.

6 Start your retro call with Meeting Notes running. It transcribes in real time — you talk through each deal or theme, you don't type.

7 After the call, Starch generates a summary with key decisions and extracts action items. You review, confirm the owners, and the items are assigned — not floating in a Slack thread.

8 Tell Starch: 'Save this retrospective summary to Knowledge Management under Retrospectives / Q2 2026, tagged with closed-lost, pipeline-hygiene, and the rep names discussed.' It's searchable immediately.

9 For recurring issues — a specific stage where deals stall, a competitor showing up repeatedly, a rep whose deals close-lost on pricing — tell Starch: 'Create a task to update our qualification criteria doc and assign it to [name], due next Friday.' This goes into Task Manager with a due date.

10 At the start of next quarter's retro, tell Starch: 'Summarize all retrospectives from the last 4 quarters and surface the top recurring themes.' You're not starting from a blank page — you're working from a pattern library.

11 If your team documents playbooks or process changes in Notion, connect Notion from Starch's integration catalog. The agent queries it live and can cross-reference retro findings with existing process documentation.

12 Set a recurring automation: 'On the last Friday of each quarter, pull all closed-lost deals from the past 90 days, group by lost reason, and Slack me a summary with the top 3 patterns so I can prep the retro agenda.' The retro prep writes itself.

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

Q2 2026 Revenue Retrospective — Mid-Market Segment

Sample numbers from a real run

Deals reviewed (closed-lost, Q2)	34
Total closed-lost ARR	412,000
Lost at Proposal / Pricing stage	18
Lost to identified competitor	9
Ghosted after demo (no follow-up logged)	7
Average deal size at loss	12,118
Action items generated from retro	6

Going into the Q2 retro, you'd normally spend two hours pulling a closed-lost export from HubSpot, cross-referencing it against Apollo to see which sequences ran, and manually scanning Gmail for threads on the big deals. This time, the dashboard is already built. You can see immediately that 18 of 34 closed-lost deals died at the Proposal stage — and when you cross-reference Apollo, 11 of those 18 had a sequence that ended before the proposal was sent, meaning there was no nurture touch between demo and proposal. That's not a pricing problem, that's a sequence gap. The retro call runs 45 minutes instead of 90 because the data is on screen before anyone says a word. Meeting Notes captures the whole conversation. After the call, three action items are extracted: update the mid-market sequence to include a touch between demo and proposal (owner: you), audit the 7 ghosted deals to see if any are re-engageable this quarter (owner: the other RevOps person), and add a required field to HubSpot for 'competitor identified' on closed-lost deals (owner: you, two-week deadline). All three go into Task Manager with due dates. The full summary — including the sequence gap finding — is saved to Knowledge Management under Retrospectives / Q2 2026. When Q3 rolls around, you pull up Q2 findings in 30 seconds instead of trying to remember what you decided three months ago.

Measurement

How you'll know it's working

Closed-lost ARR by lost reason and stage — tracked quarterly to see if the same stage keeps appearing

Sequence completion rate on closed-lost deals — how many steps were completed before the deal died, pulled from Apollo

Time from retro to action item completion — are the things you said you'd fix actually getting fixed before the next retro

Recurring theme recurrence rate — are themes from last quarter's retro showing up again, which means the fix didn't stick

Deals with no activity logged in final 30 days — a pipeline hygiene signal that surfaces during retro data prep

Comparison

What this replaces

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

HubSpot + manual export + Google Slides

This is what most two-person RevOps teams do today — it works but takes 3-4 hours of data assembly before the meeting even starts, and the findings live in a slide deck nobody searches later.

Notion for meeting notes + HubSpot reports

Notion stores the output well and you can connect it from Starch's integration catalog, but it doesn't pull CRM data or sequence data into the retro view — you're still assembling the picture manually.

Gong or Chorus for call recording

Great for deal review and rep coaching, but overkill for a quarterly retrospective workflow, doesn't connect to pipeline data, and adds another per-seat cost your 2-person team will justify to nobody.

Salesforce + Einstein Analytics

Powerful reporting if Salesforce is your CRM, but requires admin configuration time your team doesn't have, and Einstein doesn't capture or store retro meeting outputs in a searchable format.

On Starch RECOMMENDED

One platform — sales agent crm, meeting notes, knowledge management 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 use Salesforce, not HubSpot. Does this work?

Yes. Connect Salesforce from Starch's integration catalog — the agent queries it live when your retrospective dashboard runs. You won't have the scheduled-sync depth you get with HubSpot, but deal data, stage history, and closed-lost reasons are all queryable. If you also use Apollo, Starch syncs Apollo data on a schedule regardless of your CRM.

Can Starch actually read the content of our Gmail threads, or just metadata?

Starch syncs Gmail messages on a schedule, including message content — so you can surface the actual email exchange on a closed-lost deal, not just the subject line. There's a practical limit: Gmail sync is capped at 30 messages per page to avoid errors on long HTML threads, which is more than enough for deal-level retro work.

What if we already store our retrospective notes in Notion?

Connect Notion from Starch's integration catalog; the agent queries it live. You can continue writing in Notion and have Starch pull from it when building the retro dashboard. Alternatively, use Starch's Knowledge Management app to store retro outputs natively — it's searchable, AI-indexed, and you don't need a separate tool for it.

Will the action items from the retro actually get tracked, or will they disappear like they do now?

Meeting Notes extracts action items and assigns them by name. You can tell Starch to push those into Task Manager with due dates. Task Manager has P1–P4 priority levels, overdue alerts, and multiple views — it's a focused list, not a full project suite. If your team already uses Linear or Asana, connect either from Starch's integration catalog and the agent can create tasks there instead.

Is Starch SOC 2 certified? We have to ask our security team.

Not yet — Starch is not SOC 2 Type II certified today. If your company has a hard requirement for SOC 2 Type II before connecting CRM or email data, that's worth knowing upfront. It's on the roadmap.

How is this different from just running a HubSpot report before the retro?

A HubSpot report shows you deal data. Starch combines deal data from HubSpot with sequence data from Apollo and email thread context from Gmail into one view you describe in plain language — no report builder, no manual joins. And after the retro, it captures the meeting, extracts action items, and stores findings in a searchable knowledge base. The HubSpot report is the input; Starch handles the whole workflow around it.