How to Run a Marketing Retrospective That Actually Improves Things

A marketing retrospective that works is one where the team leaves with specific things to do differently — not just a cathartic venting session followed by a Notion doc nobody reads.

Here’s how to run one that matters.

The Problem With Most Retros

They’re structured as “what went well / what didn’t go well / what should we do differently” — which sounds right but produces vague outputs.

“Communication could be better.” Okay, how? “We need more budget for paid.” Based on what evidence? “The content calendar process was chaotic.” So what changes next quarter?

The problem is that identifying issues without diagnosing causes doesn’t lead to solutions. And solutions without owners don’t get implemented.

Prep Before the Meeting

The retro leader should do this work before the session, not during:

Pull the data. What did you commit to last quarter? What actually happened? MQLs, pipeline, campaign performance, content output, spend vs. budget. Have the numbers in front of the room.
Identify the anomalies. What significantly overperformed expectations? What significantly underperformed? The extremes are where the learning lives.
Collect written input beforehand. Send a brief form to the team 48 hours before: “What worked? What didn’t? What are you most unsure about?” Written responses before the room conversation prevent the loudest voices from setting the agenda.

The Retro Structure That Produces Decisions

Part 1: What the numbers say (15 minutes)

Walk through the key metrics vs. targets. Not a performance review — a shared understanding of reality. Resist commentary at this stage. Just look at what happened.

Part 2: What worked and why (20 minutes)

Go beyond “the webinar worked.” Ask why it worked. “The webinar worked because we partnered with a complementary brand that brought 40% of registrants, and we had a specific topic rather than a generic one.” The “why” is what you take forward.

For each win, capture: what was the key factor? Can we replicate it?

Part 3: What didn’t work and why (20 minutes)

Same discipline. Not “the email campaign was bad.” “The email campaign underperformed because we sent to a broad list instead of a segmented one, and the CTA wasn’t specific enough for the audience.” What’s the actual root cause?

For each failure, capture: what was the root cause? What would we do differently?

Part 4: What’s changing next quarter (15 minutes)

This is where most retros collapse — no decisions get made. Force three outputs:

What are we stopping? (At least one thing.)
What are we starting? (At most two things.)
What are we changing? (Specific process or approach.)

Every change needs an owner and a by-when.

The Follow-Up That Makes It Stick

Send a retro summary within 24 hours. It should be short:

What we learned (two or three sentences)
What we’re changing (bulleted list with owners)
What we’re tracking next quarter

Add the changes to your operating plan or project tracker immediately. If they’re only in the retro doc, they won’t get done.

Quarterly vs. Campaign Retros

Two retro types serve different purposes:

Campaign retros run immediately after a major initiative (a launch, a big campaign, an event). Focus is tactical — what specific choices worked or didn’t in this campaign. Run within a week while memory is fresh.

Quarterly retros are strategic — what does a quarter of activity tell us about our overall direction, priorities, and capability gaps? Run in the first week of the new quarter.

Both are valuable. Most teams do neither.

The Indicator That Your Retros Are Working

After two or three retros, your team should be able to point to specific things that changed because of a retrospective. A process that got simplified. A channel that got cut. A campaign type that got prioritized.