Your best customers are not who you think
The biggest single order does not tell you who matters most. How RFM segments, cohorts and the guest and registered split show who to look after and who is slipping away.
The customer you remember is not always the one who matters
Ask who the best customers are and the names that come to mind are the biggest orders, or the people who message most. Both are memorable. Neither tells you who will buy again next month.
RFM looks at three things together: how recently a customer bought, how often, and how much they spent in total. A customer who bought last week, for the sixth time, deserves more attention than one who placed a large order a year ago and has not been back.
How RFM Analysis scores customers
RFM Analysis scores each customer from 1 to 5 on recency, frequency and spend, relative to the rest of your customers, using orders in your net sale statuses. The three scores place each customer in one of 11 segments, from Champions and Loyal through Needs Attention, About to Sleep, At Risk and Can’t Lose to Hibernating and Lost. Recompute RFM refreshes the scores after a sync.
Illustrative example: Nile & Thread has two customers who look very different in the admin. Customer A placed one order of EGP 9,000 eleven months ago. Customer B has placed 7 orders of about EGP 900, EGP 6,300 in total, the last one 12 days ago. A spent more, but B is the one still buying. B is likely to score high on recency and frequency and land in Champions or Loyal; A is likely to sit in Hibernating. If B keeps the same pace for another year, that is roughly another EGP 6,300. A adds nothing unless something brings them back.
The Segment Breakdown shows each segment’s size with its average spend, average orders and average recency in days, and Top 20 Customers by RFM Score names the strongest. The page does not list every customer in a segment, so treat the counts as the trend to watch: are Champions growing, and are At Risk and Can’t Lose shrinking?
Are new customers coming back?
Cohort Analysis groups customers by the month of their first purchase and shows what share came back in each month after: M1, M2 and so on. Avg Month-1 Retention sits at the top, and Revenue by Cohort shows what each group spent later.
Big months can hide weak ones here. Say the Ramadan cohort had 1,200 new customers and 6% came back in M1, while a normal month brought 400 with 12% returning. Ramadan produced more returning customers, 72 against 48, but at half the rate. If the rate stays low in M2 and M3, the season brought mostly one-off buyers, and next Ramadan’s offer may need a reason to come back built into it.
Guests and registered customers
If your checkout allows guest orders, some of your best repeat buyers may never have made an account. Registered vs Guest compares the two groups for the selected dates: customers, orders, revenue, average order value, revenue and orders per customer, Repeat Rate, and how many were new in the period. Guest orders are linked by email, and the page warns you when orders could not be linked to any customer.
Top Guest Customers lists the guests who spent most in the period, and the table can be exported as CSV. These are the people to invite to register, or to contact directly, because they already buy without an account.
Who is slipping away
On Customers, Repeat Purchase Interval shows the average and median days between purchases for customers who bought more than once, and Recency Distribution shows how many customers last ordered in each band, from Last 7 days to 90+ days. If your median gap is 35 days, a regular customer at 60 days is overdue. CLV by Segment shows average lifetime value for each RFM segment, which tells you what an At Risk customer is worth keeping.
AI Predictions adds a churn view. Run Churn fills a risk distribution and a High Risk Customers table with up to 10 names, their days since last order and their spending trend. That short list is a practical place to start: a personal message, a reminder or a small offer for the next order.
A monthly customer check
Customer patterns move slowly, so once a month is enough. Recompute RFM first so the segments reflect the latest orders.
- RFM Analysis: note the size of Champions, Loyal, At Risk and Can’t Lose against last month.
- Cohort Analysis: check M1 retention for the newest cohort against the months before it.
- Registered vs Guest: export Top Guest Customers and pick the repeat guests to invite to register.
- Customers: read the median days between purchases and the 31-60 days and 61-90 days recency bands.
- AI Predictions: run Churn and contact the High Risk Customers first.