Strategy & Planning

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

Strategy & PlanningFor Small Marketing Teams3 apps10 steps~20 min to set up

Your three-person team runs a retrospective maybe once a quarter — if that — and it looks like this: someone pastes a Google Doc link in Slack, people drop bullet points asynchronously over three days, half the team forgets to fill it in, and the doc lives in a folder nobody opens again. Meanwhile the real lessons — why the Q1 webinar pipeline was 40% below forecast, why the November nurture sequence got flagged as spam, why the agency brief took three revision cycles — are scattered across a Notion page, a Gmail thread, a Slack channel, and one person's memory. Nothing feeds back into the next campaign brief. You repeat the same mistakes because the debrief never made it out of the doc.

Build this on Starch → See the apps used

Strategy & PlanningFor Small Marketing Teams3 apps10 steps~20 min to set up

Outcome

What you'll set up

A structured retrospective template that auto-pulls your campaign performance data from HubSpot, GA4, and your ad platforms so you're starting from facts, not memory

A searchable archive of every past retro and post-mortem, indexed so your team can actually find 'what did we learn last time we ran a LinkedIn campaign' in under 30 seconds

A task capture flow that turns retro action items into tracked to-dos with owners and due dates — not another buried bullet in a doc

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

Knowledge Management Task Manager Project Management

Data sources & config

Starch syncs your HubSpot data on a schedule (deals, contacts, pipeline stages) so the retrospective template can pull actual MQL and pipeline numbers without manual export. Connect Google Analytics 4, Meta Ads, Google Ads, and LinkedIn Ads from Starch's integration catalog — the agent queries them live when you run a retro. Notion syncs on a schedule so prior campaign briefs and context are available inside the knowledge base. Slack is connected from Starch's integration catalog for action-item broadcast.

Prompts to copy

Build me a retrospective template for marketing campaigns. It should have sections for: what we were trying to do, what actually happened (with fields for MQL target vs actual, pipeline contribution, cost per MQL, and top channel breakdown), what went well, what didn't, root cause for misses, and action items with owner and due date. Pull deal data from HubSpot and let me paste in GA4 and ad spend numbers manually.

Create a searchable knowledge base where every completed retrospective gets saved automatically. Tag each one by campaign type, quarter, and channel mix. When I search 'email nurture' I want to see every retro where email was a primary channel.

After I finalize a retro, extract all action items, create a task for each one with the assigned owner and due date, and send a Slack message to the team summarizing what we decided and who owns what.

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

Walkthrough

Step-by-step

1 Connect HubSpot to Starch (scheduled sync) and link Google Analytics 4, Meta Ads, Google Ads, and LinkedIn Ads from Starch's integration catalog — this means your retro template can pull real pipeline and spend numbers instead of making you copy-paste from five tabs.

2 Start from the Knowledge Management app in the Starch App Store. Fork the base template and describe the retro structure you want — sections, metrics fields, and how you want completed retros tagged and filed.

3 At campaign close, open your retro app and tell Starch: 'Pull HubSpot pipeline data for the last 30 days filtered to deals sourced by marketing, and pull Google Ads and LinkedIn Ads spend for the same period.' Starch queries live and pre-fills the performance section.

4 Your team fills in the qualitative sections — what went well, what didn't, root causes. Starch does not hallucinate this part; you write it. The data context is already there so the conversation is grounded.

5 When the draft is done, tell Starch: 'Summarize this retro in three bullet points and draft a Slack message to #marketing with the key decisions and action items.' Review and send.

6 Tell Starch to extract every action item from the retro and create a task for each one. Prompt: 'Read the action items section of this retro, create a task for each item, assign it to the person named, and set the due date listed. Flag anything without an owner as unassigned.' Tasks appear in the Task Manager immediately.

7 The completed retro is automatically saved to your Knowledge Management archive, tagged by quarter, campaign type, and channel mix. Future you — or your next hire — can search 'event campaigns Q3' and find every lesson from every past event debrief.

8 Before the next campaign brief, prompt the Knowledge Management app: 'What have we learned from past campaigns that used LinkedIn as a primary channel? Summarize the recurring wins and recurring problems.' Starch scans your retro archive and surfaces the patterns.

9 During campaign planning, open the Project Management app and tell Starch: 'Create a project for the May demand gen campaign. Pull the action items from the March retrospective that are still open and add them as backlog tasks in this project.' Nothing falls through the cracks between cycles.

10 Each quarter, run a meta-retro: 'Across all retros from Q1, what action items were created, how many were completed on time, and which themes appeared in the root cause section more than once?' Starch reads your archive and gives you a pattern summary — useful for the CEO briefing on why MQL volume is trending the way it is.

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

March 2026 Demand Gen Campaign Post-Mortem

Sample numbers from a real run

MQL target	120
MQLs delivered	74
Pipeline target ($)	480,000
Pipeline generated ($)	291,000
LinkedIn Ads spend	18,400
Google Ads spend	9,200
Cost per MQL (actual)	374

You ran a March push: LinkedIn Ads, one nurture sequence in Customer.io, a mid-month webinar, and content syndication through a B2B publisher. Target was 120 MQLs and $480k in influenced pipeline. You landed 74 MQLs at $374 cost per MQL and $291k in pipeline — the webinar alone accounted for 31 of the 74 MQLs, but the nurture sequence had a 12% open rate versus the 28% you expected. When you sat down to run the retro in Starch, you told it to pull HubSpot deal data for March sourced by marketing and query LinkedIn Ads and Google Ads live. The performance section filled in automatically. The team's root cause analysis identified two things: the nurture sequence was sent to a segment that had already been contacted twice in February (list hygiene miss), and the LinkedIn creative hadn't been refreshed since January (frequency cap hit). Both of those problems had appeared in the January retro too — but nobody searched for it. With the Knowledge Management archive in place, Starch surfaced the January note in the pre-retro search: 'We flagged LinkedIn creative fatigue in January. Here's what we said then.' The March retro action items — refresh LinkedIn creative every 6 weeks, audit contact segments before any sequence sends — were captured as tasks with owners and due dates before the meeting ended, and the retro was tagged and filed for the next time someone asks why a nurture sequence underperformed.

Measurement

How you'll know it's working

MQL target vs. actuals by campaign and channel

Cost per MQL across paid channels (LinkedIn Ads, Google Ads, Meta Ads)

Pipeline contribution attributed to marketing (from HubSpot deals)

Retro action item completion rate (created vs. closed by next campaign)

Time from campaign close to completed retro filed (measures whether the process actually runs)

Comparison

What this replaces

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

Notion doc + manual copy-paste

Free and familiar, but your performance data has to be assembled manually from HubSpot, GA4, and ad platforms every time, retros don't connect to task tracking, and the archive is only as findable as your folder structure.

Confluence

Better for structured documentation than Notion, but still requires manual data entry for campaign metrics and has no native connection to your HubSpot or ad platform data.

Miro or FigJam retrospective boards

Good for live facilitation and visual formats, but has no data integration, no persistent searchable archive, and action items still need to be manually moved into whatever task tool you use.

Dedicated retro tools (e.g., EasyRetro, Parabol)

Purpose-built for the retro ceremony format, but they're another point tool with no connection to your campaign data or existing task management — action items still fall into a separate system.

On Starch RECOMMENDED

One platform — knowledge management, task manager, project 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

Does Starch actually pull our GA4 and ad platform data, or do we still have to copy numbers in manually?

GA4, Meta Ads, Google Ads, and LinkedIn Ads are all available through Starch's integration catalog — connect them once and the agent queries them live when your retro app runs. For the narrative and qualitative sections, your team writes those. Starch handles the data retrieval; your judgment handles the analysis.

We already store campaign notes in Notion. Does Starch replace Notion or work alongside it?

Starch syncs your Notion data on a schedule, so your existing Notion pages are available inside the Starch knowledge base as context. You can decide whether to run retros natively in Starch (and have the data integration built in) or keep Notion as your primary doc layer and use Starch for the retrieval and action-item extraction on top of it.

Can Starch write the retrospective for us based on the data?

It can draft the performance summary section from the numbers it pulls — MQL count, pipeline, cost per MQL, channel breakdown — and it can surface relevant notes from past retros as context. The root cause analysis and what-went-well sections are yours to write. That's intentional: the value is in the team's judgment, not in having AI generate conclusions you haven't actually reached.

What if the tool we used for a campaign isn't in your integration catalog?

Starch connects to 3,000+ apps through its integration catalog, plus any website through browser automation. If your ESP or analytics tool isn't in the catalog, Starch can automate it through the browser — no API needed. For most tools a small marketing team runs (Customer.io, Klaviyo, Amplitude), they're reachable through the catalog or browser automation.

Is this useful for a team of three or is it overkill?

It's specifically built for this size. A three-person team has no dedicated ops person to maintain a retro process, no EA to track action items, and no BI tool to pull campaign data together. Those are exactly the gaps Starch fills. If you had a RevOps team and a data warehouse, you'd have different options.

Is Starch SOC 2 certified? We'd be putting campaign performance data and HubSpot deal data in here.

Starch is not SOC 2 Type II certified yet. That's worth knowing before you connect sensitive pipeline data. It's on the roadmap; check with the Starch team for the current status if compliance requirements are a deciding factor for your company.