Picture this: Julie, a website visitor, decides to send a customer service e-mail. Given her browsing history, tone in her e-mail, and mailing list subscriptions, you know she is 70% likely to make a big purchase before the holidays. Her e-mail is routed to your best support agent – to provide support, and ultimately make a sale.
It might sound too good to be true, but if your company is collecting data about your customers, it’s not far from reality. Every day, your customers are making decisions on which emails to read, products to buy, and pages to visit. If you’ve centralized this data, you can predict which of those actions matter, and just how likely they are to make other decisions.
Below, we present a simple version of such a model, based on an online fashion retailer – a store that has thousands of Julies, and which we are helping achieve its business growth objectives. Let’s see what FashionCo can do to achieve its goals.
To see how such an analysis works, review the chart below. Every row represents an activity that a customer can perform. The percentages on each row represent the percentage chance that a customer performing that activity will commit whichever action corresponds to the column.
Suppose a customer clicks a link in an email (third row). Notice that there is a value of 1.40% under “purchase below $50”. This means that there is a 1.40% chance that a customer clicking an email link will make a purchase amounting to under $50 as their next activity.
The power of this graph is in showing the relative metrics across all activities. You can now see which activities are likely to lead to which other activities, and the color coding helps as well. Notice the dark 67.42% on the “customer support request” – this shows that over 67% of customers with a support request are likely to have a support request as their next interaction.
Given such a chart, let’s see what FashionCo can do to help its customers.
Finding 1: inactive customers “reactivate” through email and spending
Our “inactive” action is a bit of a misnomer – it actually represents a customer who has not interacted with the business in 30 days. Notice that within this chart, over 88% of inactive customers reactivate through receiving an email, while 6.54% of inactive customers reactivate through making a small purchase.
FashionCo clearly depends on email marketing to reactivate customers, yet a small proportion actually return to FashionCo to make a small purchase. Targeting customers with small deals could be a great way to generate more activity.
Finding 2: small purchasers need more love
Let’s go further and take a look at the small and large purchasers – in this case, defined as those who spend below and above $50 in one purchase, respectively. Notice the major differences in email send: nearly 60% of small purchasers interact with FashionCo by email following a major purchase, while only about 36% of large purchasers do the same.
Large purchasers are significantly more likely to make a second purchase worth more than $50 in the next 30 days (compared to nearly 0% of small purchasers), but are also more likely to become inactive. FashionCo could benefit from focusing on retaining these customers through better content or even targeted customer service e-mails.
Finding 3: purchasers become inactive too quickly
Let’s look at the chart by reading an individual column, rather than a row. Below, we’ve separated the “inactive” column to see which activity leads to the highest proportion of inactive customers.
Activity maps like these are crucial in helping businesses understand their customers and preempt whatever decisions they are likely to make. Not only will they help you prepare for good times and bad, but you can also use them to benchmark across genders, cities, or stores – build multiple maps for each of your stores and see where they over- and under-perform.