E-commerce Analytics: Cohorts and Customer Lifetime Value
What cohort analysis and lifetime value (LTV) show in e-commerce, which acquisition channel is more valuable, and what actually drives repeat purchases.
rabbitclip teamPublished: 5 min read
Short answer
Cohort analysis means tracking a group of customers who made their first purchase in the same period, say the same month, over time, to measure their repeat purchase rate and lifetime value (LTV). Rather than looking at a single sales total, it shows which acquisition channel is more valuable in the long run, which gives a far more reliable basis for deciding where ad spend goes.
The most commonly misread figure on an online store is a single month's sales total: the same total revenue can come from a loyal customer base that keeps returning, or from a one-off crowd of discount hunters. Both look identical on the surface but tell a very different story about next month.
This piece covers what cohort analysis shows, how to read lifetime value, and concrete steps that drive repeat purchases.
What cohort analysis actually shows
A cohort is a group of users sharing a common trait; in e-commerce, the most common definition is 'everyone who made their first purchase in the same month'. Google Analytics' cohort exploration report tracks that group across the following months, showing how many return and how much they spend; instead of a single monthly revenue total, it exposes the behaviour pattern sitting underneath it.
A tea and coffee brand noticed a pattern this way: customers acquired during a November discount campaign returned at a noticeably lower rate than customers acquired in a normal month. Part of that campaign-driven group hadn't attached to the brand, only to the price; that finding pointed to building a dedicated second-purchase incentive for customers acquired during that campaign.
How lifetime value (LTV) should be read
Lifetime value is the total value of everything a customer has bought from a brand; Google Analytics' user lifetime report breaks this down by total users, average lifetime value and average number of transactions. A new cohort's LTV looks low simply because time hasn't passed yet; that doesn't mean the cohort is weak, it means it hasn't matured.
Reading LTV as a single number misses the point; comparing cohorts by acquisition source (search, social ads, email) is where the real insight sits. A pet products brand's data showed that customers acquired through search ads spent less on their first order but more in total over six months than customers acquired through social ads; looking only at the first purchase would have missed that entirely.
Which acquisition channel is actually more valuable
This question is answered not just by acquisition cost, what it costs in ad spend to win a customer, but by that customer's long-term behaviour within their cohort. A channel with low acquisition cost and low LTV can look cheap in the short term and lose money over time; a channel with high acquisition cost and high LTV can tell the opposite story.
This comparison isn't meaningful without at least three to four months of cohort data; deciding on a channel using a new store's first few months of data compares an immature cohort against a mature one and produces a misleading answer.
A business owner running this comparison is usually trying to answer one specific question: which channel should get a bit less budget, and which should get a bit more. Cohort data answers that with what customers from each channel actually spent over the same window, rather than a guess.
What actually drives repeat purchases
The single most concrete lever for repeat purchase rate is a reminder timed to the product's consumption cycle. If a skincare product's average consumption cycle is six weeks, a 'running low?' email sent in week five converts better than a generic campaign email sent at a random time.
A small second-purchase incentive (free shipping, a time-limited discount) works on the same principle; the real difference is timing it a set interval after the first purchase, not sending it at random.
- A reminder email timed to the product's consumption cycle
- A time-limited incentive tied specifically to the second purchase
- Channel-level LTV comparisons tracked monthly
- A new cohort never compared directly against a mature one
Who should read cohort data, and how often
Cohort and LTV data gets read monthly, not weekly; short-term fluctuation is normal, and a weekly read leads to the wrong call. This is usually owned by whoever runs marketing, or the business owner, as a standing item in a monthly review; ad budget decisions get made there based on the cohort trend, not on a single month's sales figure.
Common misreadings to avoid
The most common mistake is comparing an immature cohort directly against a mature one; that's an apples-to-oranges comparison that produces a misleading answer. The second is treating one campaign period's cohort as the general trend; customers acquired during a heavy discount period, such as November, behave differently from a normal month's cohort, and ignoring that difference skews the wider conclusion.
Cohort analysis and lifetime value show the real customer behaviour sitting behind a single month's sales figure; deciding where to put ad spend gets far more reliable once that data is in view. In a discovery call with rabbitclip, we check together whether the current Analytics setup is capturing cohort and LTV data correctly.
FAQ
How many months of data before cohort analysis is meaningful?
At least three to four months; deciding on a channel before that compares an immature cohort against a mature one and misleads.
Does a low LTV on a new cohort mean it's a bad cohort?
No; it looks low simply because time hasn't passed, and the same cohort should be measured again a few months later before any call is made.
Should cohort data be checked weekly?
No; monthly is the right cadence, since weekly fluctuation is normal and leads to the wrong conclusion.
How do you tell which channel brings more valuable customers?
By looking at that channel's long-term cohort LTV, not just what it costs to acquire the customer.
