Meet the poster boy of data-driven retailing in India

Could you tell us about the new sources of data that are emerging? How are they coming together, and how do they link back to sales?

There are a few areas that are emerging as critical. Social media is one of them.

For most brands, including us, in earlier phases, it was all about building size and scale - put crudely, ‘getting likes’. We all thought once we have a large, engaged community, 80 per cent of the work is done.  Over the last one year, the realisation has dawned: ‘What is the end from social media we are working towards?’ This year (2013) has been about life beyond likes.

The second part of the story, is how we could link the data from social with actual buying. Sadly, the (social media) companies themselves don’t seem to have done enough on this front.

The day we turned 1 million fans, we gave out a coupon only for our Facebook fans. On one weekend, we had 10,000 people coming in, and shopping for an average of Rs 2,000 after showing the coupon – which is Rs 2 crore of incremental turnover. That made us believe that there is some possibility in the space. We took it a step further.

We have three million people in our First Citizen (loyalty programme) database, who contribute to around 71 per cent of our sales. On the other hand, we had five million Facebook fans. We always knew that there must be some commonality between the two. We tried hard, but we could not find anyone in the industry who could tell us how to take this further. We figured that there is a way, of de-duping across these two. We picked two variables – e-mail and cell phone number. We have those details, because we charge people Rs 300 to enter the programme; they don’t just fill up the form with arbitrary details.

We de-duped this information against the Facebook data. We saw a 50 per cent (approximately) overlap. We said this is good – because Facebook is about the mind, First Citizen is about physical shopping. If we could start drawing this link between the two, we would be 10 steps ahead of the game.

So, if we had a Miss Sharma in our database, and we know she is also on Facebook, now, not only would we know that she is heavily into, say, ethnic wear, we would also know that she recently delivered a baby. Or that her husband got a promotion. In shopping, having a baby is a huge life event. Till now, I only knew about it when you shopped for maternity wear, or later maybe for baby-related things. That is too late to influence the shopping.

If I can marry these two – loyalty data and Facebook data – not only do I have the action, I also have what’s on her mind. Using this, we created some segments out of this 50 per cent overlap. One of the segments is the ‘Inactives’. These are people who are not shopping with us, but did so sometime back. Today they are fans, but are not shopping with us. We could target them with offers and communication. We could use them as a controlled group and measure the lift. Those are seeds we have already sown. We have done some experiments and are getting decent traction.

If we want to connect social media up with business, this is the path. Frankly, we are just at the tip of the iceberg. There is a whole lot we will be doing in this space.

Analytics has been spoken about for a while now...

You may say analytics is an old story, but it is not. I would say it is the biggest pillar of our business today. It is an asset that we are sitting on. The data we have is for three million customers, and goes back 18 to 19 years for a lot of them. This is real user data – not market research data. There is no forecasting or estimation; what we have is universe data.

Tell us about its application today.

This data is multifarious in its use for us. To begin with, we started using it a lot for adjacency analysis. It threw up things like people who bought a suit, coming back within 10 days to buy a shirt or a belt, or a pair of shoes. So what we do now is, when people buy suits and jackets from us, we trigger a message on an accessory to go with it. We have been doing things on adjacency earlier too - it has only gotten bigger, more robust and smart over the last year.

Our process is that we convert data into an insight. If you jump into petabytes of data, you are certain to sink. We get into data with a hypothesis. Let’s take something very simple: a guy buying formal shirts from us, must also be buying formal trousers. Data is then used to accept the hypothesis, or modify it and sharpen the insight. Only ‘X’ number of people buying shirts were buying trousers, as we realised.

We don’t jump into the data and look for trends. We see trends in the marketplace, and we verify those trends through data.

We realised that working women as a segment, is different from housewives. You might intuitively agree - there are some expected differences. But there was one category, which we chanced upon. We realised that working women also buy more toys for their kids than housewives.

All of this is about intelligent targeting. In India, communities behave differently. A Malayali is different from a Tamilian; a Bengali is quite different from a Punjabi. Most retailers have a Durga Puja celebration in Kolkata. There are a lot of Bengalis who are not based in West Bengal, and there are pockets in other cities where there would a sizeable number of them – like Powai and Lokhandwala in Mumbai, Chittaranjan Park in Delhi. Bengalis are identifiable by their surnames. The top 50 Bengali surnames would account for 90 per cent of the community. We got the top 50 surnames, and de-duped those against our base. We found ‘X’ number of Bengalis who don’t stay in Kolkata. We realised the big pockets in other cities.

Irrespective of whether they are in Bengal, Bengalis still treat Durga Puja as a time for dressing up, going out, celebrating. That’s a great opportunity for us as a retail chain to come into the picture. We targeted Bengalis outside Bengal during Durga Puja, and got an increase of Rs 1.5 crore in sales.

The other community we tapped was Muslims. Eid for them is a big consumption opportunity. We targeted them specifically and the incremental revenue was about Rs 1 crore.

The same logic has now been taken forward with a variety of communities. The idea is to target communities with different stimuli, because one size does not fit all. Advertising often locks you into the ‘one size fits all’ approach. The game is not about that one fancy piece of creative, that one big idea, that one big insight. It’s actually those thousand smaller insights, thousand different stimuli that are measurable.

Which are the agencies you work with on analytics?

Much of it is in-house. We farm out bits and pieces, but 98 per cent of what we do is in-house. One is it is our own customer data, which is confidential. We also realise that it is a significant differentiator in our business.

How have returns from analytics gone up in the last 12 months?

Last financial year, we did Rs 40 to 50 crore worth of incremental turnover from analytics. This financial year, we’re looking at Rs 85 to 100 crore.

Eighty five crore sounds like a big figure. But it is actually broken down into 50 or 60 initiatives. And the best part is that it is completely measurable. You clearly know what you put in and what you got, which, in the area of mass advertising, can be much trickier to determine.

Are analytics investments bucketed under marketing? Has marketing expenditure gone up in the last year?

Marketing for us has two or three parts. One is the mass marketing efforts (including digital), which is typically 1.5 per cent of sales. We spend about 1.5 per cent more on loyalty, and all the initiatives I spoke about. We have historically been at around 3 per cent and that hasn’t changed. But the contribution of loyalty, analytics and digital has crept up to become bigger. This is not to say that we are off mass media – we are certainly there and that’s not going away.

We’re seeing more malls and stores. Has that impacted footfalls, conversion and yield?

We measure what is called ‘customer entry’. The number of people walking in is not growing at a fast pace – there are not too many new people coming in all the time. The reason for this is also the number of stores. We had one store in Mumbai; now we have 10 (including one each in Kalyan and Thane). Catchments are shrinking. There is a breaking up of audiences.

However, because of targeted marketing, the conversion is increasing. So are the ticket sizes and basket sizes.

Are there trends in upcountry / metros? Do you also see merchandise getting localised by region?

In bigger towns, we have always had a very large selection of formal wear. You can imagine why – there is a substantial amount of working population. When we initially went into smaller towns, we went in with that kind of a thought process. We quickly realised that in towns like Vijayawada, Coimbatore or Baroda, there aren’t that many corporate working individuals. There are people from business – imports/exports, small business and so on. They don’t have this need for formal-formal wear. There, we have switched gears and gone for a much more casual avatar. We see much more of denims, T-shirts and casual wear. We also see significant differences in women customers – smaller towns have a far lesser number of working women. Ethnic wear is big in big towns, but not growing that fast. In smaller towns, ethnic is very, very large. For women there, western formals are of no use.

Because we have this data, we have an understanding of who is consuming what. Brands used to ask us if slim fits are being bought by young customers. The young are buying it, but it is not just them. We discovered that slim fit is actually being bought by the middle aged guys. If you look only at transactions, you will never get to know that. This changes a lot of things for brands.

In places where you launch a new store, what is the adoption of the First Citizen programme?

Wherever we have opened a new store, we already have a sizeable First Citizen customer base. Take Baroda for instance – there were already tens of thousands of First Citizen members before we launched the first store. They would have shopped either at Ahmedabad or even our Mumbai stores. So even before opening, we have a handle in terms of what Baroda wants. That gives us a very good inkling into merchandise planning. You don’t have to go there and learn.

How has the role of the CMO changed?

As a CMO, you are a lot more integrated into the business vis a vis somebody who is doing just advertising. The opening up of loyalty and analytics, and to some extent social and digital, have changed the way a CMO thinks. From a personal perspective, what I was doing five years ago, and what I am doing today, are radically different.

How has this data helped in personalisation?

It makes a big difference. For our preview sales, we used to send out a mass mailer to, let’s say, 2.5 million people. If it’s personalised with your name on it (not on the name stamp, but inside), the probability of your opening it is that much more. We actually then proved it in the business. We saw a 30 per cent difference in response rates between the two.

Is there a difference in ticket size between First citizen members and non-members?

Non-First Citizen ticket size is 30 per cent lower than a First Citizen. So we know that the minute a person becomes a loyal customer, he’s going to give you 30 per cent more.

Has discretionary spending come down?

I wouldn’t say so. The categories where spending has come down, from what I read, is automobiles, travel and holidaying, real estate and the like. But in April-September 2013, we have had a 14 per cent like-to-like sales growth. That’s a number we haven’t done for many quarters. The theory to that could be that people are still willing to buy into smaller luxuries. You could delay buying a car, but nothing stops you from buying that lipstick. It’s reflecting in the merchandise that’s getting sold.