Customised Shopping – Online returns


Real life case studies

Interesting examples where data science is being used in real life situations to provide insight, help with making important decisions and improve lives.

Customised shopping - online returns

This case study is based on the work completed by a Scottish business that focused on helping companies manage their online returns. The premise of the case study is that retail businesses often focus on sales at the expense of managing returns.

One of the retailers interviewed discovered that they were spending £25 million a year shipping stock to and from returners to achieve zero sales. For another company, they discovered that reducing returns by 11% would have the same bottom-line impact as increasing sales by 20%. And we all know how is much easier to manage an existing customer than it is to acquire a new customer.

Let’s bring this to life with a simple scenario:

From a sales perspective, Craig is a much better customer as you get 20 sales to Chloë ‘s 1. But once you factor in the cost of the return of the 19 items from Craig, it turns out Craig is not only getting you less sales, he is actually costing you money.

By looking at the whole picture, and tracking the right indicators, we get clearer insights into who is the best customer.

Now, let’s extend our thinking a bit further and consider what we can do to get a better business outcome. We could offer Chloë and Craig different experiences to try and influence their behaviour.

When Chloë  uses our shopping app, she will see everything we have for sale. And when Chloë  is about to buy something, we can offer her incentives to buy more e.g. get the 2nd one half price, or buy a third one and get free shipping. And we do this because we know that Chloë  is considered a shopper who tends to keep what she orders.

Craig could get a very different experience. When Craig shops he only sees a subset of the available stock, has a limit to the number of items he can buy, and may notice that the cost of shipping has increased to factor in the inevitable cost of returns, because we know that Craig is considered a shopper who tends to return most of the items he orders. 

This is just a very quick example to get us thinking about what indicators to track and about how data can help you decide what actions to take to influence behaviour.

If you think that this idea of customised shopping is still to come, I would suggest you think again. How many times have you noticed a hotel, or a flight, go up in price when you look at it for a second time?