Customer Cohort Analysis: Track Which Buyers Stay and Which Ones Churn

Customer Cohort Analysis: Track Which Buyers Stay and Which Ones Churn

Most retail analytics focus on aggregate numbers: total customers, average transaction value, monthly revenue. These are useful for tracking performance at a glance — but they mask something critical. Beneath every "new customer" metric is a question your averages can't answer: Are you acquiring customers who actually come back, or are you filling a leaky bucket?

Cohort analysis answers that question. It's one of the most powerful analytical techniques available to small retailers, and one of the least used — primarily because it sounds more complicated than it is.

This guide walks you through exactly how to build a customer cohort analysis using data you already have in your POS system.

What Is a Cohort Analysis (and Why It Changes Your Perspective)

A cohort is a group of customers who share a common characteristic at a specific point in time. In retail, the most useful cohort type is an acquisition cohort — a group of customers who made their first purchase in the same time period.

For example: everyone who made their first purchase in January 2025 is "the January 2025 cohort." You then track that group over the following months to see how many of them come back.

This matters because aggregate retention metrics are misleading. Your overall 30-day repeat rate might hold steady at 25% — but that average could be hiding the fact that customers acquired through a recent campaign are returning at 40%, while customers from a winter promotion are returning at 12%. If you mix them together, you see 25% and think nothing has changed.

Cohort analysis separates the signal from the noise.

The key insight: two businesses with the same revenue growth can have radically different retention health. One is deepening relationships with existing customers. The other is burning acquisition spend to replace customers who don't come back.

The Two Cohort Reports Every Retailer Should Run

You don't need a sophisticated BI platform to run cohort analysis. You need customer transaction data with timestamps. Here are the two most important reports and what each reveals.

1. Acquisition Cohort Retention Table

This report answers: "Of the customers who first bought in Month X, how many came back in Month X+1, X+2, X+3?"

The output is a triangle-shaped table called a retention waterfall. Each row is a cohort month. Each column is the number of months since first purchase. The cell value shows the percentage of the original cohort who returned that month.

A healthy retail retention waterfall looks roughly like this:

• Month 0 (acquisition month): 100% — by definition, this is the cohort's first purchase
• Month 1: 25–45% (varies by category and purchase cycle)
• Month 2: 15–30%
• Month 3: 12–22%
• Month 6: 8–18%
• Month 12: 5–15%

The curve flattens over time. Customers who survive to Month 3 are disproportionately likely to become long-term regulars — which is why the Month 1 to Month 3 drop-off is the most important window to study.

2. Revenue Cohort Table

This report tracks how much revenue each cohort generates over time — not just whether they return, but how much they spend when they do. Some cohorts return at similar rates but with drastically different basket sizes.

Run both. A cohort with 30% Month-1 retention but high average order value is often more valuable than a cohort with 40% retention but low spend. Revenue cohorts prevent you from optimizing for return visits while missing the underlying economics.

How to Build Your First Cohort Table

Here's the step-by-step process using standard POS export data.

Step 1: Pull Your Customer Transaction History

Export at minimum 12 months of data (18–24 months gives you more reliable trend lines). You need three columns at minimum:

• Customer ID (anonymized is fine)
• Transaction date
• Transaction amount

If your POS assigns customer IDs only to loyalty members, your cohort analysis covers that segment only — which is still valuable. Just note the coverage gap when drawing conclusions.

Step 2: Identify Each Customer's First Purchase Date

Group your data by Customer ID and find the minimum transaction date for each customer. This is their cohort assignment date. Label each customer with their cohort month — for example, "2025-01" for anyone whose first purchase was in January 2025.

Step 3: Calculate Months Since Acquisition

For each transaction, calculate how many calendar months elapsed between the transaction date and the customer's first purchase date. A transaction in the same month as first purchase = Month 0. One month later = Month 1. And so on.

Step 4: Build the Retention Table

Create a summary table where:

• Rows represent each cohort month
• Columns represent months since acquisition (0, 1, 2, 3... up to 12 or 18)
• Cell values show the count of unique customers who made at least one purchase that month ÷ total customers in that cohort

Format it as percentages with conditional formatting — green for high retention, red for low. The visual pattern tells you as much as the raw numbers.

Step 5: Build the Revenue Version

Repeat the same process, but instead of counting unique customers, sum their transaction amounts by cohort × month elapsed. This gives you a revenue waterfall showing cumulative spend per cohort over time — which is the number that ties most directly to business growth.

Reading the Waterfall: What the Patterns Mean

Once you have your retention table, interpretation is where the real value lives.

A sharp Month-0 to Month-1 drop (below 20%): Your first-purchase experience isn't converting customers into repeat buyers. The product, service, or expectation gap is too wide. Look at what your lowest-retention cohorts have in common — was there a promotion that attracted bargain shoppers? A category with low replenishment frequency? A particular season or campaign?

A flat or gradually improving curve after Month 2: Customers who return twice have "activated" — they've decided your business is worth revisiting. If Month 2+ retention is stable, your core experience is strong. The fix is getting more people to that second purchase, not overhauling what happens afterward.

Strong cohort-to-cohort variation: This is the most actionable pattern. A January cohort retaining at 35% in Month 1 versus an October cohort at 18% tells you something fundamental differed in the acquisition mix or experience. Was there a seasonal effect? A different product emphasis? A staff change? A promotional approach that attracted different buyer types?

Declining retention across newer cohorts: If your most recent cohorts consistently underperform older ones at the same elapsed months, your recent customer acquisition has degraded in quality. The fix isn't more volume — it's better targeting. Chasing scale on a leaky cohort accelerates the problem.

Cohort Analysis in Cannabis Retail: What's Different

Cannabis retailers face a unique retention dynamic: purchase frequency is highly category-dependent. A flower customer might buy weekly. An edibles customer might buy monthly. A wellness-focused customer might buy quarterly.

This means a single retention table will blend very different purchase cycles, making averages harder to interpret. Consider building separate cohort tables segmented by:

• Primary category at first purchase (flower, concentrates, edibles, topicals, accessories)
• First-purchase channel (in-store walk-in vs. online pre-order pickup)
• Loyalty enrollment status (enrolled at first visit vs. not enrolled)

Dispensaries that run this analysis consistently find a 10–20 percentage point retention gap between loyalty-enrolled and non-enrolled customers at Month 1. That gap alone justifies aggressive first-visit loyalty enrollment — not as a customer experience nicety, but as a measurable operational lever with a clear ROI.

A secondary insight: dispensaries often find that their online pre-order cohorts retain at materially higher rates than walk-in cohorts. The friction of the pre-order process filters for high-intent customers who already know what they want — and they come back.

Turning Cohort Insights into Action

Analytical work is only useful when it changes behavior. Here are the three highest-leverage actions cohort analysis enables.

Target the critical second-purchase window. If your Month-0 to Month-1 retention is 28%, that means 72% of new customers bought once and haven't returned. The 7–30 days after a first purchase is your highest-leverage marketing window. A personalized follow-up — based on what they bought, timed to their likely replenishment cycle — can lift Month-1 retention by 5–12 percentage points. That's a meaningful shift in customer economics, not a rounding error.

Identify your best acquisition channels. If you can tag customers by acquisition source (loyalty enrollment method, referral, event, promotion), you can run separate retention tables by channel. This reveals which sources produce customers who actually stick — and lets you reallocate marketing spend toward higher-LTV channels before you've wasted another quarter on low-retention volume.

Set retention goals grounded in reality. Without cohort data, retention targets are guesses. With three or more months of cohort history, you can set Month-1 retention goals based on your actual baseline, run interventions, and know whether an improvement was real or seasonal noise. That's the difference between a data-informed operation and one that reviews the same aggregate metrics and wonders why nothing changes.

The Bottom Line

Cohort analysis is one of those techniques that permanently changes how you think about your business. Once you see that your January cohort retained at twice the rate of your October cohort, you stop asking "how do we get more customers?" and start asking "how do we get better customers — and keep them longer?"

The data is already in your POS system. It just needs a structured question: of the customers who first bought in Month X, how many came back?

• Start with 12–18 months of transaction history and build acquisition cohort tables by month
• The Month-0 to Month-3 window is where retention is won or lost — focus interventions there
• Segment by category, channel, and loyalty status to find the customer types with the strongest long-term value

At Chapters Data, we help small and mid-sized retailers — including cannabis dispensaries — build the customer analytics infrastructure to run analyses like this without a dedicated data team. If you're ready to stop managing by aggregate metrics and start understanding your customers at the cohort level, we'd love to show you what your existing data can reveal.