Insights for market share improvement of SMEs through Analytics
SME sector contributes to 52% of the GDP and 45% of the employment of Sri Lanka. The origins of these SMEs are diverse. Friends get together and start a business, or the business is passed down from generation to generation. Some SMEs are formed on geographical importance. e.g.: Tea factories.
For the purpose of this write-up, we have selected a retail clothing shop as an example to discuss a few insights gained through data. These analytical techniques can be applied to many other businesses.
When you talk of analytics in the Sri Lankan SME sector, the main challenge is the lack of data. But many of these SMEs are equipped with ERPs or at least a POS. So, business owners can think about starting analytics with the minimum data available with them with these systems and create results. This case study was done with limited internal data available with the ERP running at this retail clothing shop.
Descriptive analysis
We started with some descriptive statistics such as the items which are most profitable for the business.
Management decided to focus more on the sales of items with higher profit per item and came up with an action plan that included marketing of these items in the premises like stand boards, and billboards and also marketing these items using SMS for the registered customers.
Online ordering platform
The shop also had a small online portal where customers could order online. We also looked at the online ordering platform and identified that 86% of the customers had only ordered once using the platform. Some of these customers were contacted, and their pain points were identified and a smooth process was implemented aiming more customers using an online portal.
Customer classification
Shop has introduced a loyalty card for the customers but the data collected from customers at the issuance was not recorded in the system. So, the classification was done based on only two simple variables: spending and the number of visits to the shop. Based on that, customers were classified into 4 segments. If these SMEs could store this information in their systems, advanced classification methods like K-means clustering could have been possible.
Management decided to launch a customized SMS for customers in segments 1 and 2, with certain discounts to attract them on a regular basis. Since they are loyal customers to the organization, the more they can retain the customers, and continue to sell to, the more likely they are to achieve their business goal.
Furthermore, for customers in segments 3 and 4, management decided to introduce a tier of rewards that they can unlock when they spend more with the shop. Eg: Spend 15,000/- with us and get an umbrella free.
Market Basket analysis
This retail shop had the data granular to the customer receipt level. We did a market basket analysis to find out what customers tend to buy together.
Let’s look at the first record to understand what’s in the table.
Percentage of both items bought together: When we consider all sales contained in the item “Girls skirt 361”, 42% of these sales included the item “Girls T-shirt 2840” also.
Likelihood of a customer buying both items together: this is a statistical scale of measuring the likelihood of a customer buying both item 1 and item 2 together. The likelihood of a customer buying “Girls skirt 361” and “Girls T-shirt 2840” together is 14.1 times more than the chance of purchasing one item.
We identified many associations in the data set, However, for the demonstration purpose, we have only included 10 associations in the article.
Looking at these patterns, the management implemented a couple of strategies such as bundle offers with right pricing, placement of the products next to each other in both physical store and online store. These measures helped the management to reap the optimum benefits.
Social media data
The shop had good ratings on the social media platform and we also looked at what kind of goods attracted the customer most by looking at their reviews. By looking at this, the management could identify what features of their business best attract customers. They need to maintain the consistent quality of what customers like the most.
The management also decided to go through the negative reviews and improve some of their services to make customers more comfortable with their shopping experience.
This article discusses how an impact can be created with the minimum data available. Based on the availability of data, analytics methods can be much broader. However, Sri Lankan business owners in the SME sector do not see the value that can be created by analytics. Also, the right expertise might not be in these businesses to crunch the data.
Hope this article will help you to understand the value that can be created through analytics. These businesses can partner with consultancy firms like StrEdge, so they can unleash their true potential.
Article by: Gemunu Premarathna
Lead Consultant — Data Analytics at StrEdge
