Strategy & Planning

How to run a retrospective or post-mortem as Chief of Staff and Founder's Office

Strategy & PlanningFor Chief of Staff and Founder's Office3 apps11 steps~22 min to set up

You're the one who ends up running every post-mortem and retrospective at a 150-person company, which means you're also the one chasing five functional leads for their notes the day before, copy-pasting Slack threads into a Notion doc at 11pm, and manually pulling HubSpot deal data and QuickBooks actuals to figure out what actually happened on the project that went sideways. By the time you've assembled enough context for a useful conversation, half the attendees have already moved on mentally. Action items from the session get buried in Notion pages nobody revisits. Three months later, the same failure mode repeats itself and someone says 'I feel like we've been here before.' You know exactly when — you wrote that retro doc.

Build this on Starch → See the apps used

Strategy & PlanningFor Chief of Staff and Founder's Office3 apps11 steps~22 min to set up

Outcome

What you'll set up

A pre-built retrospective brief that auto-pulls relevant data from Slack, Notion, HubSpot, Gmail, and Google Calendar before the meeting even starts — so you walk in with facts, not vibes

A structured action-item tracker tied directly to the Project Management app, with owners, due dates, and status visible across every retro in one place

A searchable archive of every retrospective your company has run, so when history repeats itself you can prove it — and show exactly what was decided last time

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

Project Management Knowledge Management Meeting Notes

Data sources & config

Starch syncs your Slack, Notion, Gmail, and Google Calendar data on a schedule — so the pre-brief pulls from live, up-to-date sources without you manually exporting anything. HubSpot (for deal or pipeline context on GTM retros) and QuickBooks (for budget actuals on finance retros) are also synced directly by Starch. The Project Management and Knowledge Management apps are built into Starch, so action items and retro archives live in the same connected workspace — no separate tool to log into.

Prompts to copy

Build me a retrospective brief app that pulls the last 30 days of Slack messages from #q2-launch, any Notion pages tagged 'Q2 Launch', Gmail threads where I was CC'd with subject lines containing 'launch', and all calendar events with 'launch' in the title. Summarize what happened, surface any recurring complaints or blockers, and format it as a one-page brief I can share with the leadership team 24 hours before our retro.

Create a post-mortem action tracker: every action item we log during a retro should become a task in Project Management with an owner, priority (P1–P4), and due date. After 2 weeks, Slack me a digest of which items are still open and who owns them.

Build a retrospective archive in Knowledge Management — every retro we run should auto-save a structured entry: what worked, what didn't, root causes identified, decisions made, and action items. Tag each entry by team and quarter so I can search across all of them.

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

Walkthrough

Step-by-step

1 Connect Slack, Notion, Gmail, Google Calendar, HubSpot, and QuickBooks — Starch syncs all six on a schedule, so they're ready to query before the first retro you run.

2 Tell Starch what you're retrospecting on: a product launch, a missed quarter, a specific initiative. Give it a date range, a project name, and the Slack channels or Notion pages most relevant — Starch assembles a draft brief from that context automatically.

3 Review the auto-generated brief: timeline of key events, a summary of blockers and complaints surfaced across Slack and email, relevant HubSpot pipeline movement or QuickBooks actuals if this is a revenue or budget retro.

4 Share the brief with attendees 24 hours before the session so everyone walks in with the same facts — no more 'wait, what actually happened?' consuming the first 20 minutes.

5 Run the retro session with Meeting Notes active — it transcribes in real time and pulls out decisions and action items as the conversation happens, so you're not typing while also facilitating.

6 After the session ends, Meeting Notes generates a summary with key decisions, root causes identified, and action items extracted and attributed to specific people.

7 Review the action items and push them to Project Management with a single prompt: 'Create tasks for each action item from today's Q2 launch retro, assign owners as discussed, mark anything with a 2-week deadline as P2.' Starch creates the tasks, assigns them, and sets due dates.

8 Save the full retro — brief, transcript summary, decisions, and action items — as a structured entry in Knowledge Management, tagged by team and quarter.

9 Two weeks out, Starch Slacks you a digest: which action items are still open, which are overdue, who owns them. You don't have to chase anyone — the data does it.

10 Before the next retro on a similar topic, search the Knowledge Management archive: 'Show me every retro we've run on GTM execution issues in the last 12 months.' If the same root cause is showing up twice, you have the receipts.

11 Share the action-item completion rate with the CEO or exec team as part of your monthly operating review — it's a concrete signal that retrospectives are actually producing change, not just conversation.

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 GTM Launch Post-Mortem — June 2026

Sample numbers from a real run

Action items from retro	14
Action items closed within 2 weeks	9
Recurring root causes identified vs prior retros	3
Hours saved on pre-brief assembly	6
Retro sessions archived and searchable	11

After a Q2 launch that slipped by three weeks, you need a post-mortem that actually produces something other than a long Notion page nobody reads again. Two days before the session, you tell Starch: 'Pull everything relevant to the Q2 enterprise launch — Slack from #q2-launch and #gtm-team from April 1 to June 15, Notion pages tagged Q2 Launch, Gmail threads I was on with subject lines containing launch or delay, calendar events for launch syncs, and HubSpot pipeline movement for enterprise deals in Q2.' Starch synthesizes a five-section brief: timeline of key events, top three blockers mentioned repeatedly across Slack (engineering dependency on a third-party API, unclear handoff between product and sales engineering, two-week delay in final pricing approval), deal-level HubSpot data showing 7 enterprise deals pushed from Q2 to Q3, and a QuickBooks actuals summary showing $180K in unplanned contractor spend to compensate for the delay. You share it with the six attendees at 8am the day before. The session runs 90 minutes instead of the usual 2.5 hours — people arrive knowing the facts. Meeting Notes captures the discussion; 14 action items come out, assigned to six owners. You tell Starch: 'Create tasks for each of the 14 action items from today's Q2 launch retro, assign owners as we discussed, set P1 for anything blocking Q3 launch readiness.' Done in 30 seconds. You save the full retro to Knowledge Management under 'GTM / Q2 2026.' When the Q3 retro prep rolls around, you search: 'Show me all GTM retros from the last 18 months.' Two of the three root causes from Q2 appeared in the Q4 2025 retro too. You bring that to the CEO before the Q3 retro even starts.

Measurement

How you'll know it's working

Action item closure rate within 30 days of each retro (target: >75%)

Recurring root causes identified across consecutive retros (a rising number is a leadership problem, not a process problem)

Pre-brief assembly time (hours spent manually pulling data before a retro session)

Time from retro session end to action items assigned and visible to owners

Number of retros with searchable archives vs. retros that exist only in someone's memory or a buried Notion page

Comparison

What this replaces

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

Notion (manually assembled retro docs)

Notion is fine for storing the output, but you still have to pull every data point yourself — Starch automates the brief assembly and writes the action items back into a tracker without a second tool.

Confluence + Jira

Jira handles the action item tracking well, but the brief-building and cross-system data pull still falls on you; Starch handles the whole loop from data assembly to task creation in one workspace.

Miro or FigJam retro templates

Useful for the in-session facilitation experience, but they produce outputs that live nowhere useful — no searchable archive, no action item tracking, no connection to your actual data.

Dedicated retro tools (Retrium, EasyRetro)

Built for the facilitation ritual, not for a CoS who needs to pull cross-functional data, track action items over time, and connect retro outputs to the rest of the company's operating cadence.

On Starch RECOMMENDED

One platform — project management, knowledge management, meeting notes 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

Can Starch pull from all of our Slack channels, or just specific ones?

Starch syncs Slack on a schedule and gives the agent access to channels and users. When you build your retro brief app, you specify which channels are relevant — you're not dumping the entire company's Slack history into every brief, just the channels where the relevant work happened.

We use Linear for engineering task tracking. Can Starch push action items there instead of its built-in Project Management app?

Linear is reachable from Starch's integration catalog of 3,000+ apps — the agent queries it live when your automation runs. You can tell Starch to create Linear issues for engineering-owned action items while logging everything else in the Starch Project Management app. Describe the routing logic and Starch builds it.

What if the retro is about something that doesn't have a clean data trail — like a culture issue or a team dynamic problem?

The brief will only be as strong as the data that exists. For a culture or team retro, Starch can still surface relevant Slack signals, meeting frequency data from Google Calendar, and any Notion notes from 1:1s or team check-ins — but it won't manufacture insight where the data doesn't exist. In those cases, the Meeting Notes transcription and action item extraction are often the more valuable parts of the Starch workflow.

Is my data stored in Starch, and should I be worried about sensitive retro content sitting there?

Starch is not SOC 2 Type II certified yet. If your company has strict data residency or compliance requirements, that's worth flagging before you connect Slack and Gmail. Starch doesn't offer an on-prem or self-hosted option today.

We've tried retro action item tracking before and it always falls apart. How is this different?

The usual failure mode is that action items live in a different place from where people do their work — so nobody checks them. Starch closes that loop in two ways: first, it creates the tasks in a tracker that's visible in the same workspace as your other operating tools; second, it Slacks you a digest of open and overdue items on whatever cadence you set. The follow-up doesn't require you to remember to do it.

Can I use this for board-level post-mortems, not just internal team retros?

Yes. The same workflow applies — you'd pull from QuickBooks, Stripe, and HubSpot for financial and pipeline context, and the resulting brief and archive are formatted for a more exec-facing audience. You'd describe to Starch what level of detail and what framing you need, and the app outputs accordingly.