The problem
Mailmodo is email software: businesses use it to send emails to their own customers. Most of Mailmodo’s customers are small and use it without anyone from our team looking after them. That is called self-serve. A support team cannot talk to everyone, so it has to choose who to spend time on.
We were choosing badly. The groups we used described what we had decided, not what the customer was doing: “we decided they don’t need help”, “they chose to be self-serve”, “they ghosted us”. Nothing in them told us who was about to leave and who was ready to spend more, so the same message went to everyone, it worked for no one, and more customers were leaving.
What I did
I redesigned it. Every self-serve customer is now placed in one of four groups by how they actually use the product, and each group has its own plan: what to do, how often, and who does it. Placement is automatic, from usage data that refreshes every day, so nobody sorts customers by hand.
The goal was simple: keep more of these customers, help the ones who can grow, and stop spending the team’s time on accounts that don’t need it.
Two email terms, in plain words
- Bulk email. One message sent to a whole list at once, like a newsletter or a sale announcement.
- Automatic email. Emails that go out by themselves when something happens, like a welcome message after someone signs up, or a reminder after an abandoned basket. Mailmodo calls these journeys and trigger campaigns.
Almost everything in the strategy comes back to one question: does this customer only blast
emails by hand, or have they set up emails that work for them?
The four groups
Auto PilotAlready good at it. Sends steadily, gets good results, and asks for help only when something breaks. Light-touch care: regular check-ins and product updates.
Needs a nudgeUses it, but results are weak or patchy. Wants ideas and a little help. We coach them toward Auto Pilot.
Stable, not growingJoined for one job, and it is done. Little time and no appetite for more. Light, automatic contact, and make sure payment is on autopay.
Gone quietNothing sent for 30 days or more, though they still pay. We find a use that makes the product stick.
Four groups, sorted by behaviour
The group names are Mailmodo’s own: Auto Pilot, Non-Auto Pilot, No Growth Opportunity and
Dormant.
How we decide who is on Auto Pilot
Auto Pilot is the group everything else is measured against, so the test has to be strict and
steady. My first version looked at one month, and customers kept flipping in and out. The
version that works asks for three months in a row. Each month the customer must have:
- had at least 1 in 10 emails opened, and 2 in 100 clicked;
- sent at least 5 bulk emails;
- used automatic emails, in a journey that month or a trigger campaign at some point in the three;
- used at least 100 email credits (the unit of sending in their plan).
Miss any one month and the customer drops out. Customers who use it but do not pass go to the
“needs a nudge” group, and every group has a written rule for moving in and out, marked
automatic or needs a person.
How it runs day to day
- Usage data arrives every day: emails sent, opens, clicks, credits used, automatic emails running.
- It is combined with account data: who the customer is, their invoices, their onboarding stage and who looks after them.
- The rules place each customer in a group and move them when their behaviour changes.
- The group sets the plan. For example, an Auto Pilot customer gets a check-in every three months. A support message from them is marked high priority. A customer whose results slip gets a goal, a playbook, and a return to Auto Pilot once they reach it.
What came of it
Mailmodo’s small customers now get the help that suits them, and the team’s time goes where it counts.
- The team knows who to call and why. Time goes to customers who are slipping or ready to grow, and the steady ones get light, automatic care.
- Churn gets an early warning. If a customer’s last 30 days make up less than a quarter of their last 90 days of activity, their activity has fallen off, and they move to a group that gets help before they leave.
- A clear way to grow revenue. The data showed that customers who use automatic emails have higher open and click rates and send much more. The plan is to move the others across, and see whether it helps them as much.
- Every customer gets a message that fits, instead of the same one as everybody else.
It is live and running. How much it has reduced the number of customers leaving is still being measured, and I will add the figure here.
What the data showed
The first look at 88 customers on Auto Pilot found the biggest opportunity in the whole group:
- Most only send by hand. 36 of the 88 (41%) send bulk email and nothing else, and 57% had no automatic emails running in the past month.
- Automatic emails go with better results. Customers using them average about 35% opened and 10% clicked. Five who used only trigger emails reached 46% opened and 27% clicked. Whether automation causes this, or just marks the more skilled customers, is still to be tested.
- They also use the product more. Customers using bulk, trigger and journey emails together used about 75% more of their sending allowance (17.6k credits in 30 days) than bulk-only customers. That is where growth in revenue comes from.
What I learned
- Group people by what they do, not by what you decided. “We decided they don’t need help” says nothing about the customer.
- One month misleads. A rule on a single month moves customers around for no reason. Three in a row holds.
- Write the way out as carefully as the way in. Every group says how a customer leaves it and whether a system or a person makes the call.
- The data pipe is half the job. The groups only work when the right numbers arrive every day, next to the customer details.
Tools
Vitally (customer data), Amplitude (usage data), Chargebee (billing), HubSpot, Google Sheets and Apps Script
