Write the January retention plan in December

Every December a CRM team acquires more customers than in any other month of the year, and every January it discovers that most of them have gone. The pattern is so consistent that it stops registering as a problem. It becomes weather.

The reason it repeats is scheduling rather than skill. In December the team is running peak. In January the team is running the sale, then the post-sale clean-up, then the planning cycle. The one month when the December intake could still be reached is the month nobody has capacity to think about it.

So the work has to happen now, while peak is running and before the intake is complete. What follows is what we build in the first half of December for a programme that wants a different January.

Your December intake is four cohorts pretending to be one

Treating December signups as a group is the mistake that makes everything downstream look hopeless. The month contains people with almost nothing in common, and their retention differs enormously.

Split the intake at least these four ways before you plan anything:

• Gift buyers. Bought for someone else. Delivery address differs from billing, or the order carries a gift message or gift wrap. Their interest in the category may be zero.
• Deal-led buyers. First order carried the deepest discount you offered, or arrived within hours of a promotional send. They bought a price, not a brand.
• Seasonal category buyers. Bought something that only makes sense once a year, such as an advent product or a party item. They may be loyal and simply annual.
• Normal new customers. Bought at or near full price, in a core category, for themselves. These are the only ones your usual retention assumptions apply to.

Each group needs a different January. The gift buyer needs to be told what the brand is for. The deal-led buyer needs a reason to pay full price or needs to be allowed to lapse cheaply. The seasonal buyer needs nothing until autumn, and mailing them weekly until then is how you produce complaints.

You can build these splits from order data you already have. None of it requires a survey or a new tool.

Score last December before you plan this January

You do not have to guess how the groups behave, because last year's intake has already told you. This is the most useful two hours available in December, and almost nobody spends them.

Run the cohort analysis this way. Take the customers acquired in December of the previous year, split them by the same four definitions, and plot the share of each group still buying at three, six and nine months. Put a normal month's cohort on the same chart as a reference line.

The retention curve for each group tells you where to spend. In most programmes the gift and deal-led curves fall away within the first quarter, while the full-price core group behaves like any other month. That shape is the argument for spending January's budget on two of the four groups rather than all of them.

Work out a rough customer lifetime value for each group while you are there, even a crude one based on twelve-month revenue. If the deal-led group's value over a year is lower than the discount that acquired it, that is a finance conversation, not a CRM one, and December is the moment to raise it.

Decide now what January will and will not discount

January plans get written under pressure, and under pressure every plan becomes a discount. Deciding the rules in December, while trading is still good and nobody is panicking, produces better rules.

Write down the answers to three questions and circulate them before the holidays:

1. What is the floor discount for January, and who is allowed to approve an exception?
2. Which segments are excluded from discount entirely, and on what evidence?
3. What is the non-discount offer for each excluded segment?

That third question is the one that takes work. A customer who bought at full price in a core category should not receive the same twenty per cent reminder as everyone else. You are paying to move an order that was going to happen anyway.

The alternatives are usually earlier access, a service benefit, a bundle, or simply better timing. We set out the same argument for recovery journeys in basket recovery without discounts, and the reasoning carries over directly to a January sale.

Build three programmes before the year ends

By the second week of January, build capacity is gone. Anything that is not in the platform and tested by the end of December will slip to February, which is too late for the cohort it was meant for.

Three are worth having ready:

A second-purchase programme for the gift group. Its job is to introduce the brand to someone who has only ever transacted with it. Content, not offers. Trigger it on delivery, and give it a hard exit at the point your usual repeat window closes.

A winback programme aimed at last December's intake. These people bought a year ago and have not returned. They are more reachable in the same season than at any other point, because the occasion that brought them is repeating. Cap it at three messages and suppress anyone who has bought since.

A sunset programme for the addresses peak just added. Peak inflates the list with people who will never open again. A defined inactivity window, with a final re-permission message and then a stop, protects your sending reputation in Q1 when volumes drop and complaint rates become more visible.

Build all three now. Schedule them, test them with seed addresses, and have them approved before anyone goes on holiday.

Agree the January report in December too

The January post-mortem is usually an argument about numbers, because nobody agreed what would be measured before the month started. Fix the definitions now.

Three measures are enough. The share of the December intake with a second purchase by the end of February, split by the four groups. The contribution margin of January revenue after discount, not the revenue line. And the complaint and unsubscribe rate through January, by week, because that is where the cost of a badly targeted sale shows up first.

Put those three on one page, with last year's figures beside them. The comparison is what makes January a result rather than an impression.

The part that stays hard

None of this fixes the underlying issue, which is that a large share of December buyers were never going to be customers. They came for a gift or a price, and no lifecycle programme converts everybody.

The value of doing the work in December is not that you keep them all. It is that you stop spending January's budget and your sending reputation on the ones you cannot keep, and spend it on the group that was always going to stay.