Dividing, they gain

Customer segmentation effectively magnifies the impact of a brand.
Customer segmentation

Customer segmentation is all about managing and understanding large sets of data. Segmentation allows brands to run multiple campaigns targeted at different groups, which are most likely to make proper utilisation of the communication, instead of bombarding the entire consumer population with unnecessary messages.

“We use micro-segmentation for one of our F&B clients. We have divided their customer database into thousands of smaller groups based on their demographics as well as purchase behaviour, taste preference, etc, and send them the offers that are more suited to them. As a result, since launch, the client has witnessed a 3-5 times better hit-rates as against mass campaigns,” says Aneesh Reddy, CEO and Co-Founder, Capillary Technologies.

Domino’s Pizza India Limited case is a trendsetter when we talk about customer segmentation. Taking an overview, following were the business measures available to the managers to analyse data – sale value, number of pies or side items, number of orders, average ticket value and food cost, and parameters like time, customer category, geography, product, etc.

Pawan Gadia, CEO at Ferns N Petals, says segmentation enables one to take a holistic view. “It gives more drilled data, enabling one to achieve higher numbers. In our case, 20 per cent of the customers are the ones who buy more than five times in a year, and this chunk contributes 40 per cent of the top line sales,” he adds.

Shailen Amin, CEO, Bestylish, gives an interesting insight and says it is very important to realise as to when you should start segmenting your customer base. Having a decent repeat buying trend is a good sign, for instance.


Predicting clusters

It is understood that the whole exercise is not really about what customers have bought, but it is about what they are searching for, what are the items that they are viewing, items they are keeping in their carts and coming to them later, etc. “It is about trying to draw clusters of people depending on their behaviour. There could be a mixed shopper or predominantly a gadget freak. We try and predict such relevant clusters,” says Muralikrishnan B, Country Manager, eBay India.

The way to go is using predictive algorithms, which entail you to do this on a scalable manner. You cannot segment customers on a spreadsheet. At eBay, they indulge in doing the ‘churn analysis’. If a customer has not shopped at your site for 75 days, the probability that he will never come back to you is 60 per cent; it takes a lot of data to understand this and this is just information. The question is – what can you do to convert this information into insight driven and actionable information? You can pick up from options like calling the person/sending a mail/offer, etc.

Muralikrishnan adds, “Offline stores don’t have this kind of information, like how many customers have bought men’s leather jacket as well as women’s shoes. Being in the e-commerce space, we have this luxury of information.” The approach needs to be intelligent, evolved, and worked across large masses of data. Gadia says the way to deal with it varies from organisation to organisation. “We need to plug the data into the tool, given that data could be both online or offline. One can use customised tools or have a team as well. We have hired an outside agency for serving the purpose,” he said.

Mayur Patel, India Country Manager, PayPal, says, “We focus more around how we can enable PayPal services to different consumers who want to transact online. We take into consideration consumer preferences like using bank account/credit card/debit card, etc. We look at those preferences to make PayPal services more relevant to the customer.” PayPal is a global e-commerce business, which allows payments and money transfers to be made through the internet.

Gadia further says that while the research entails having thrice-a-day online research and once-a-day offline research, they take a time frame of approximately one year to create a relevant set of database (like from one Valentine’s Day to the next Valentine’s Day). They delve into the data on a monthly basis for customer segmentation. Muralikrishnan says it takes about 3-6 months for re-working on segmented data – re-doing and re-clustering people – because once you’ve identified the pattern, you need to give it some time for the hypothesis to play out.

Amin says at Bestylish, they have an in-house CRM background person, who takes care of the segmentation, while they are also considering having outside expertise to analyse the data. He points out that CRM requires a bit of infrastructure – taking data, segmenting it, slicing it and dicing it – and that is an investment.

Affordability makes it easy

The primary reason that has been stopping the brands from going for customer segmentation is the lack of technology and the cost involved. Until a few years ago, only the large retailers like Tesco and Target were able to afford customer segmentation. “Now, with the advent of newer technologies, more and more retailers have begun to adopt smarter business intelligence tools for customer segmentation. Also, many retailers aren’t aware of the right parameters to segregate their customers, which we can arrive at using statistical and analytics tools,” comments Reddy.

How does it work when you’re partnering with a player like Capillary, which has a team of analysts who undertake customer segmentation for clients? “We begin by defining the most important parameters and then create the structure required for the segmentation. The team then uses statistical tools to sort the existing database into defined clusters and study their strongest preferences. These findings are then applied to designing marketing campaigns, which have a strong appeal to the targeted customers,” says Reddy.


Important criteria

Number of transactions (ticket value)

The category the consumer likes (category of large purchase)

Consumer’s age

What/how much constitutes the purchase

Buying for themselves/others

Pulling factor

Time of last transaction/number of times shopped last month

Other criteria

Top cities

Item selling most

Time/month/payment mode


Products being searched


Dos and don’ts

Amin says they rework on segmentation every quarter or so and don’t indulge in it too often. “You should give it time while you understand the data and trends.” He emphasises on capturing as much as data as possible and says considering too little data should be a complete no-no. People behave the same, be it offline or online; the key is understanding and bridging the gap. Gadia seconds Amin and says having a correct database is of utmost importance and one should not jump on to it. He says that a decent amount of turnover is an important criterion as it enables one to analyse mature data, focusing on products and traffic. Gadia also focuses on employing the correct tools, not relying on incomplete information, and having an expert opinion.

It is not to be forgotten that an online customer has a chance of visiting an offline store and vice versa. Gadia says the secret of bridging the gap in both the cases is providing the right kind of voice support.

Focusing on capturing consumer information and blending it to fine tune your business model is the game changer and will act as a catalyst in the path to profitability, given that ‘profitable’ is not the expectation in the first two years.

A win-win situation

Experts say the biggest benefit to the brand is a higher response rate in every campaign. The brands can run hundreds of campaigns targeted at different clusters and micro-clusters at the same time and generate faster and higher sales. “This is definitely an advantage as compared to running one single campaign for the entire customer population. As more segments are created, you can expect an increase in conversion rate. In most cases, the clients have witnessed a reduction of 4 per cent in customer dropouts and a 3 per cent growth in transactions,” says Reddy.

He says for different categories like apparel, footwear, etc, his team designs a new structure based on the customer characteristics and client requirements. Since no two clients are same, neither are their customers, the segmentation model cannot be replicated. He explains, “In the apparel business itself, we have clients who deal in premium clothing as well as the clients who are into affordable wear. The brand appeal in each case is different; the customer expectation from each brand is different. Therefore, we need to design a new customer segmentation map for everyone.”

Muralikrishnan says the difference lies in the granularity. “The granularity of segmentation is finer in a niche site,” he says.

While in offline segmentation, the focus is mainly on demographics and customer purchase behaviour, in online segmentation, it is more about the customer psychographics (level of tech savviness) and their interests online and offline. Recently, Capillary has launched a new solution that allows retailers to integrate their loyalty programme online and offline. This allows them to use social platforms to recognise and reward their customers and get a 360-degree reach.

Amin says, “It is an exercise, wherein you are trying to pull your customer back, the benefit being that if you can segment your customer, it becomes a personalised pull.

Publish Date
Not Sponsored
Live: People Reading Now