Time to Let Your Customer Go: How to Master Cost-effectiveness
Sometimes, giving up on a customer is the smart way to go, as it’s more beneficial than continuing to push through and witness rising costs. Learn what to consider before saying goodbye and good luck
Relationships, all kinds of relationships, require a vast investment of multiple resources. Time, energy, sometimes money. And in any relationship, it is extremely hard to know when and if it’s time to let go. To bluntly say: “it’s not working anymore.”
When a relationship deteriorates, there is inevitably a time when the damage has been done and nothing can save it. When you own a business, the answer should be simpler: weigh your cost and your benefit, play with the equation, know how to identify the sunk costs (that shouldn’t be considered when making the decision to continue investing in an ongoing project, or in a relationship with a customer), and you have a number, no?
Since we’re talking business here, a strategy that will help bring back those who fell out of love with us is crucial. We agree with the common take that every customer counts, and that a business should always aspire to maximize its customer engagement and value by sending personalized messaging customized to his/her behavior. Still, there is an exception to every rule. While a strong retention and reactivation strategy is essential in optimizing the value of your customer base, you should look at the relevant metrics to help and draw the line of cost/benefit.
Accept the fact that some customers simply cannot be served profitably. It may be better to let go, instead of investing more in re-engagement. In this piece, we’ll give you guidelines on how to decide whether to keep your marketing efforts for certain customers, and make sure the cost does not exceed the probability to reactivate.
A Real Churn-Fest
To determine whether a customer is worth reactivating, we conducted research and looked for the cut point between the projected reactivation rate and the expense per customer. This cut point can help the marketer determine the appropriate amount of reactivation investment required, if any. The research was conducted on six ecommerce brands and includes over four million customers. In this blog post, churn customers were defined as customers who didn’t make a purchase for at least a year. The reactivated customers are defined as customers who made a purchase post churn.
Before we dive into the formulas, let’s delve a little more into churn customers’ behavior. In the graph below, we can see what percent of customers churned according to the months following their first order.
From the graph above, it’s easy to understand that most customers churn right after their first purchase. The next graph will help us better grasp this notion; most customers churn on the first day. They make a purchase and never come back.
These two graphs provide a picture of the odds of when a customer will churn according to their last order. This is important information for a marketer when deciding when and what to target a new customer with.
Waking up the sleepers
Always remember that churn customers are about 87% of all customers, and out of churn customers we have on average 15% reactivated customers. The migration between these two segments – churn and reactivated and the probability to reactivate, is one main key that will help us figure out how much we should invest in a customer. We can come up with some sort of formula:
For example, if every SMS we send out costs us $0.04, the probability to reactivate is 15% and we have 100K customers to target – the total cost will be $0.04*100,000= $4,000. The total number of reactivated customers will be about 15%*100K= 15K. And let’s say each reactivated customer will give us $10 after reactivating. Plugging all these numbers into our formula will give us the maximum number of times you should send a customer a campaign: 15,000*10/4,000 = ~38 times. But there’s much more to it than that.
At the speed of your customers
The probability of a customer reactivating is the reason we continue targeting customers after they’ve churned. It’s a known fact that the cost of retaining an existing customer is substantially lower than acquiring a new one. Of course, speed is of the essence. In the graph below, we have the distribution of customers who were reactivated after not purchasing for an entire year, according to the month they made the purchase. Remember that the reactivated customers haven’t been active in the past year so that “month 12” represents their first month in the churn segment.
The graph above shows that the more time a customer spends in the churn segment, the probability of him/her making an order decreases. In other words – as time goes by, customers have a lower chance of reactivating. We might be stating the obvious (you’ll probably say – duh!), but seeing the data helps. We can benefit from this information by deciding when to send campaigns and also identifying when we should stop.
More orders, more…well, orders
The next graph shows how the number of orders effects the churn customers. We see the average reactivation rate according to the number of orders the customer made before reactivating.
This graph stabilizes after about eight orders. The meaning here is that the percent of reactivation is higher as the number of orders a customer made increased. But there is a limit to how much it will go up. If we translate this information to a campaign strategy, a company might want to put more effort into customers who made more orders because the chance of them reactivating is higher, but after a certain point – around the 8th order, they should probably invest the same amount for all customers.
Seeing into the future
To fully understand and determine how much to invest in a certain churn customer, we’d like to know how much money that customer can generate for us in the future. The following graph shows the average number of orders after reactivation, based on how many orders the customer made before churning. The horizontal axis is the number of orders prior to being segmented as churn, and the vertical axis is the number of orders after reactivating.
The more orders a customer makes before churning, the more he will probably make after reactivating. Adding the average order amount will allow the marketing team to put a real number on each customer.
The next table is a bit complicated, but bear with us. It shows the reactivation rate – where the first month in churn (month 12) is the total probability, as we presented in the “Reactivation rate according to the number of orders” graph. As the months go by, the probability decreases by the percent of customers who didn’t reactivate the month before according to the combination of number of orders the customer made pre-churn, and the number of months since the last time he made a purchase.
The trend is, of course, similar to the data we presented earlier – as the number of orders increases the probability to reactivate is higher, and as the months since last purchase go up, the probability to reactivate goes down.
Let the numbers talk
If we take the formula we discussed in the beginning and add parameters like time and number of orders, we will have a better indication of how much we should invest in targeting the customer.
For example, how we should treat customers who bought 15 months ago and made two orders before churning (formula 1), compared to customers who churned 18 months ago, but bought three times (formula 2). This case can be tricky because the first group has a higher chance to reactivate, and the second group has probability to make a higher number of orders after reactivating.
Let’s put these two scenarios to the test with the formula we created and shared above:
So, scenario 1 is: 10%*3.0*10/0.04=75
and scenario 2 is: 8%*3.3*10/0.04=66
To help a bit, here are the numbers again (in scenario 1):
10% – The chance for 15 months and 2 orders scenario, based on the table above.
3 – The predicted orders based on the graph above for customer with 2 recent orders.
$10 – The revenue from an order.
$0.04 – The cost of delivery (SMS/Email)
If we do what the formula recommends, we will be willing to invest more in the first group and will try harder to make them return and buy more.
It is important to say that the number of times the formula gives us is for all kinds of marketing techniques – SMS, Emails, Retargeting CPM etc., so although it may seem like a high number, we should keep in mind it can include all kinds of campaigns and not, for example, 75 text messages.
Every campaign counts
Basically, if we use the good old phrase “every customer counts,” we might eventually find ourselves investing more in customers who have a lower chance to reactivate. The graph below shows the “correct” way to do it, where the number of campaigns we send out is synchronized with the customer’s probability to reactivate. The horizontal axis is the number of months in churn (after 1 year, remember?), the blue line is the probability to reactivate after one order, and the light blue line is the number of campaigns that should be sent according to the calculation we presented above. As the customer’s chance to reactivate decreases, the number of campaigns we should send him will also decrease. Your company can choose and calculate where the stop sign is located, considering many budget and strategy issues.
In conclusion, there are many parameters we can examine when analyzing the complex topic of “when should we let go” of a customer and stop, or at least reduce our marketing efforts. Each brand should look into their numbers and evaluate their costs to make smart decisions about invested money based on actual metrics. And don’t forget, we are always here to consult and advise if needed.
very interesting article. I have a few questions, in the first above case you´ve said that we should send campaign 38 times per customer. What kind of campaign do you mean? Is it a reactivation campaign? In what time period you recommend to send 38 times?
The first formula presented in the blog is a simplified equation that assumes a consistent reactivation rate across all churn customers, regardless of additional substantial factors, such as their past order history or time since last order.
The figure 38 is, ostensibly, the total number of campaigns that you should send to each individual churn customer as of the time of calculation (again, in a simple world, where all customers have the same probability to reactivate – in our example 100K customers with a 15% chance of reactivating).
Since we’re referring to the churn segment, the relevant campaigns are, indeed, reactivation campaigns, although the concept we tried to convey here does not relate to the content of the campaign, but to its cost versus benefit. Since we do not have information about the cost of each brand’s marketing campaigns, and since this cost changes across communication channels, we decided to demonstrate the concept based on the number of campaigns rather than the cost of each one.
Please note that as the amount of time elapsed since the customer’s last order increases, the overall number of campaigns we’d recommend sending to such a customer decreases (the aggregated number of lifetime campaigns increases, of course).
Thanks for your questions!