Your customer behaviour at your fingertips
With rapid growing Pakistani fashion industry, it is inevitable that customer trends and preferences would evolve accordingly. Customers not only have access to more choices in terms of brands and styles, but also in the way they purchase these items. Various factors now play a role in the buyer's decision-making process, ranging from budget, outlet accessibility, fashion trends, weather, and external events in the country. Couple this with rising competition, and we are faced with the challenge of understanding our customers better and rewarding the most loyal ones.
Challenge
Our client, a leading retail shoe brand, was faced with the task of managing a huge database of customer data and segmenting it in a way that would aid in business decision-making. The client wanted us to carry out this exercise for the customers who had a loyalty card, as they had been promoting the loyalty card program for a few years. At the time of our analysis, there were 20,000 loyalty card holders, but the data had to be organized and structured for it to become more meaningful for the company.
Solution
We developed a segmentation model based on the agreed parameters with our client. Our key objective was to take a holistic view of loyalty transaction data, which was a result of several variables (loyalty transaction data, card issuance data, loyalty points data, external factors such as weather). We had to merge all these data sets into one organized form where they would correlate with each other to give us actionable insights.
We considered the minutest of details such as which days of the week/months/seasons had the most shoppers, which categories of shoes saw the most sales and when, how did pricing and promotions play a role etc. But most importantly, we looked at the loyalty tenure of customers. Loyalty tenure defines the length of relationship between the customers and the brand. We saw which customers had been most active for which months, and which of them were likely to fall out of the loyalty chain altogether.
Moreover, we used proprietary customer segmentation techniques to divide the customers into nine segments:
The first 8 categories made approximately 66% of the total loyalty revenue and comprised 31% of customers. Each of these categories was then further studied in detail, in terms of average transaction/customer, most preferred season/outlet/location for shopping, and many more.
Result
Some of the results provided deep and valuable insights into customer purchase behaviour. For instance, in our regional analysis we found out that Karachi has the lowest tenure of loyalty relationship with more than 50% of cardholders ending the relationship within 10 months. This could possibly be because of more competition, dispersed locations, and differing tastes of the varied demographics in the city. To dig deeper, we conducted a store-wise analysis which further detailed which stores were contributing to the low loyalty of customers in Karachi. Thus, we had detailed and accurate action points for the management – they knew where to allocate their resources to improve performance.
Moreover, we helped unveil some critical customer insights based on the data. For instance, we found out that regardless of the which segment the customer belonged to, more than 60% of the loyalty sales were coming from the top 4 segments, clearly indicating where the investment should go. Similarly, unique findings like peak sales from key customers during wedding season would be useful for targeted promotions and market basket analysis to up trade customers for more purchases.
Our data also helped identify the 'Almost Lost' or 'At Risk' customers who were likely to fall out of the loyalty program and decrease/stop spending on the brand. Once identified, these customers can then be sent tailored messages or discounts to win them back and result in incremental sales.
