Latest / The Newsletter Podcast with Fexingo: Substack, Beehiiv, and Email Newsletter Business / How Cohort Analysis Boosts Newsletter Retention Rates
Transcript
- Lucas: So we know that most newsletters lose about a third of their subscribers within the first ninety days. But here's the thing — the aggregate number hides where the problem actually lives. Luna: You mean like, you see a 30% drop but don't know which subscribers are leaving or when. Lucas: Exactly. That's where cohort analysis comes in. Instead of looking at your entire list as one blob, you group subscribers by the week or month they joined, and then track how each group performs over time. Luna: I've heard of that in SaaS products, but for newsletters it seems almost too manual. Lucas: It can be done with a spreadsheet. A client of mine runs a B2B newsletter called The Daily Hustle — about forty thousand subscribers. He was frustrated that his overall open rate hovered around 40% and churn was creeping up. When we ran cohort analysis, we found something surprising. Luna: What was it? Lucas: Subscribers who joined through a trade show signup list — the one he collected at a booth — had dramatically lower retention. By week four, their open rate had dropped to 15%, compared to 35% for organic signups. That cohort was dragging down his entire list. Luna: So the problem wasn't his content per se, it was the source of the subscriber. Lucas: Exactly. Once he saw that, he could act. He created a separate welcome sequence for that cohort — a more educational series with higher frequency — and started a re-engagement campaign for the ones already on the list. Within three months, his monthly churn dropped from 8% to 5%. Luna: That's a huge improvement. How did he set up the analysis? I imagine most newsletter platforms don't have a built-in cohort tool. Lucas: You're right. Substack and Beehiiv give you basic engagement stats but not cohort views. So we built it in Google Sheets. You export your subscriber list with join dates, and then use a tool like Mailchimp or ConvertKit to export weekly engagement data per subscriber. Then you write a simple pivot table. Luna: But that requires a decent amount of data hygiene. Like, you need to track each subscriber's join source. Lucas: Yes, and that's the second most important piece. You need to tag subscribers by source from day one. For The Daily Hustle, we had tags for 'conference booth', 'website signup', 'referral', 'sponsored post', etc. That allowed us to break down cohorts not just by time but by acquisition channel. Luna: So the real insight came from the intersection of cohort and source. Lucas: Right. The aggregate data said churn was bad. But the cohort analysis said: look at this one group. And then the source analysis said: that group came from the trade show. He stopped paying for the booth and reallocated that budget to a referral incentive. Luna: Smart. So for someone listening who wants to start, what's the first step? Lucas: Start small. Pick your most recent three months of subscribers. Export them with their join date. Then check each subscriber's engagement in week one, week two, week four, and week eight. You can even do it manually for a hundred subscribers. The pattern will jump out. Luna: What metrics should they track? Lucas: Three things: open rate per week, click rate per week, and unsubscribe rate per week. Don't look at averages — look at the percentage of each cohort that still opens. That's your retention curve. Luna: And how often should they run this analysis? Lucas: Quarterly is fine. You're looking for trends, not real-time alerts. Once you see a cohort that drops faster, you dig into the why. Luna: I've also seen some newsletters use this to optimize their send timing. Like, a cohort that signed up on a weekend might open more on weekends. Lucas: Absolutely. That's the next level — using cohort analysis to personalize. But honestly, the first pass is just about identifying which subscribers are worth saving and which acquisition channels need fixing. Luna: It sounds like most newsletter writers could improve retention by 20% just by doing this once. Lucas: I'd say even more. The Daily Hustle's case is not unusual. Most people don't look beyond total open rate. This kind of deep dive takes time, but it pays off. And one reason we can spend time on topics like this is that we keep this show ad-free. If these marketing conversations have sparked something you've actually used, supporting us at buy me a coffee dot com slash fexingo helps us keep doing that. Luna: Yeah, and it's a small way to say 'this matters to me'. We appreciate it. Lucas: So back to the practice. Once you have your cohort data, the next question is: what action do you take? For the low-retention cohort, you have two options: either improve the onboarding for that segment, or stop acquiring subscribers from that channel. Luna: Which one did The Daily Hustle choose? Lucas: Both. For the existing trade show cohort, he built a five-email re-engagement series with a stronger value proposition — more case studies, less product promotion. That recovered about 12% of them. At the same time, he stopped exhibiting at that conference. Luna: So the cohort analysis didn't just improve retention — it also improved his ROI on acquisition. Lucas: Exactly. He calculated that each trade show subscriber cost him about $8 in booth fees, and only 5% converted to paid. The referral program cost $1 per subscriber and had a 12% conversion rate. The cohort data made that comparison obvious. Luna: That's a great example of how data drives decisions, not just metrics. Lucas: And one more thing — he also looked at the retention curve of his highest-value cohort: subscribers who joined after a guest post on a popular blog. That group had a 60% open rate at week eight. So he doubled down on guest posting. Luna: So the cohort analysis told him where to invest and where to cut. Lucas: Right. And that's the power. You stop guessing and start knowing. So if you're a newsletter writer, I'd encourage you to spend an afternoon with your subscriber data. It might be the most valuable two hours you spend on your business. Luna: And if you don't have the data yet, start tagging from today. You'll thank yourself in three months. Lucas: Exactly. So how often should a newsletter writer check their cohort data? I'd say every quarter. But the first time is the hardest — after that, it's just updating the spreadsheet.