'Analytics is helping less experienced people get the same insights faster.'

Rajeev Jorapur, vice-president, MIS, Bajaj Auto, reveals how business analytics is enabling across-the-board gains for the motorcycle manufacturer from product development to sales, and the influence of social media on business planning. An interview by Nilesh Wadhwa.

How is Bajaj Auto leveraging business analytics?
Firstly, analytics is making things visible and secondly, establishing relations between different factors. With experience, one is able to anticipate a lot of causal effects. Analytics is helping us do away with the need for experience and even for a raw engineer or salesman. It is giving us a way of figuring out how things are co-related; for instance, is the buyer a Platina customer or a Dominar customer? An experienced salesperson would have figured this out based on his experience but with the help of analytics the need for a salesperson to have that experience can be done away with. Based on certain attributes, a salesperson can know if the buyer is a Dominar or a Pulsar customer.

This means analytics is making it easier for company  personnel to understand different needs, right from pre-development to sales. Is it more of an anticipated or calculative approach?
Analytics is enabling less experienced people to get the same insight that earlier required a lot of first-hand experience. For instance, if you take a machine, an experienced technician, only by hearing the sound of the machine, would understand that something is not right. Now with IoT, new technologies and machine learning, a raw engineer would be aware that as some parameters are not within defined limits, he needs to investigate. Therefore, he may arrive at the same results which an experienced person would have figured out from the sound. That is the second area where analytics has helped us.

The third area is throwing up results, which compels us to investigate or expect unexpected outcomes. It, therefore, leads us to investigate to better understand the second phenomenon.

Can you elaborate?
Yes, there was a classic case where we were discussing the performance of Bajaj dealers in a specific Indian state. We had a list of dealers who were underperforming and those who were performing well. This was known to the people working in the field. There was a brand manager who wanted to know how his brand was doing. So when that brand was selected, suddenly four or five dealers got graded out. We were using ‘Qlik Analytic’ that helped us learn that these four or five dealers were not selling a single unit of that brand (model). How is it that these four did not sell any and others are selling so many of the same brand? This may not have been evident in normal circumstances, the big data was there but it was not easily visible.

Many such instances happen whether you talk of top failure, some parts would appear in a list of dealers and they may not appear in the remaining dealers even though they are from the same geography. Therefore, that leads to further investigation.

In which other areas do analytics help automakers?
It is things that provoke us into investigating deeper and give us a better understanding of the context. In this example, it was the customer. Other examples are in the paint shops, production ready (models) where apparently not correlated parameters tend to show certain common trends, which leads to deeper understanding. This is the third area.

The fourth area is in the manufacturing. The ideal state is a single-piece floor where every vehicle in the assembly line is capable of being manufactured differently from the following vehicle. Earlier, it was batch manufacturing all vehicles of the same configuration. Therefore, manufacturing has to have the capability to say that if the first one is a blue Platina, can the next be a red one? This was at the colour and cosmetic changes level.

The next level was, can I have the Pulsar 220NS and a 180 on the same manufacturing line? The same brand but different models, functionally different but still the same.

A even further level of detail is: can I manufacture a Platina and a Dominar on the same line? This is ideal from the manufacturing perspective. From a customer perspective, segmentation was done at a very gross rate. For instance, the South is a brand-A market, North is a brand-B market, East is a C market. However, the reality is different. In south India, I may have a section of population that has come from the north and the preference is towards brand B and not brand C. Analytics is helping us to have sharper segments, more targeted and personalised and therefore help us move closer to mass-personal. I can give each customer a unique experience and do it at scale.

Broadly, these are the four areas where we have gained a lot from analytics.

Customer preferences have changed with people's exposure to a humungous amount of data. The vehicle market has also evolved. What are the key differences you have seen before and after using analytics?
There are two dimensions to this. On the one hand, analytics has made the task more difficult and on another dimension, it has made things easy. Where it has made things easy is data where the input is in controlled circumstances. If we analyse the conversion of footfalls in certain dealerships, we are able to get a fair estimate of the type of experience a person is going through in that dealership. We don’t need to ask a dealer or customer specifically about the experience. This input is controlled; in that sense, analytics has helped us.

Where it has made things difficult is where there is a lot of data from unknown sources from uncontrolled inputs like social media. You have platforms where people freely express their opinions; how credible is this information out there? Unless we are able to have a sense of credibility, the organisation doesn’t know whether to act on it or ignore it. In either case, it can be make or break. To that extent, I would say the capabilities analytics brings on the table can make things difficult for the decision maker. Earlier when this was not known, there was no decision to be taken but now it is in front of you. Now the business head has to decide whether to act on this or ignore it. Therefore, the task becomes more challenging.

Social media has given a voice to the consumer to freely express his/her opinion. How much input do you use from the social media or website for developing new products, customer experience and marketing activities? How much does this influence your decision?
It is one input to the product development, depending on the context. We tend to believe what is stated; we treat that as symptoms and it is up to the product managers and R&D to figure out what these symptoms are telling us. Are they pointing to the need for action and desire, which may be legitimate but from a brand perspective we may not like to fulfil the desire? For instance, if someone wants a Platina to behave like a Dominar, we cannot fulfill that. It has helped give additional input to the product development, R&D and manufacturing and, therefore, factor in the process of continuously improving either the experience or physical product, or the way we are doing business.

Which segments in the organisation use big data and how does it help you in your global marketing plan?
As our managing director Rajiv Bajaj said, we are an Indian company going global and therefore India is just one of the markets. Pretty much all the departments are using it, the context is different, the information they need is different, and the actions that they take are different.

If I start from the customer, we have various social pages and we actively listen to what people are saying on these pages;  that’s an input to how we position the product and the brand.

Then there is Bajaj-owned digital assets (a call centre, website). Let’s say there are people who are saying that they like these attributes of the brand very much, that input is very important today. There are people who appreciate some quality and there are some who say they had a problem with some component, dealership or servicing. If it is related to product quality, then the central quality assurance is very interested to figure out whether this is a one- off case, generic or local issue. They then get very interested and analytics helps them to go right up to the affected vehicles. If it is related to the customer's  experience with the dealer, service centre or channel partner, then the channel management team does the same investigative analysis: is this a one-off case, generic issue or a local issue?

Therefore, what analytics has helped is that first this information flows very fast. Secondly, you get the full context in terms of the extent of the problem, whether it is one customer or a section of customers, whether for one brand or multiple brands.

A lot of associated information is available easily, and it is credible because you are able to validate it with the person who has reported it. This is even more important in the international markets as geographically they are located far apart.

In global markets, we go to the distributors, who in turn go to the dealer, who then sells to the end customer. Thus, big data is helping us in finding out how the product is behaving in the international markets.

What are Bajaj Auto’s plans for electrification?
At this time, we cannot rule out the possibility. I cannot say when it (a Bajaj EV) will be introduced but I am sure it will come.

Do you think two-wheelers will drive early adoption of electro-mobility in India?
I am not sure about two-wheelers. I believe it will always be a blend, it will never be only electric vehicles. There will be petrol, CNG and diesel. I believe it will be one of the modes.

I don’t foresee no petrol vehicles or no diesel vehicles, at least in the near future. It will be a blend in the two-wheeler and three-wheeler vehicle space.

What is Bajaj Auto's expenditure on IoT, big data and other new technologies?
In the analytics space, I would put it as a percent of IT expense, 15-20 percent.

Bajaj Auto was the first Indian company to use robots for product development in batch manufacturing. How much has productivity been enhanced and how much of production is currently done through cobots?
Collaborative robots or cobots, specifically, are not my responsibility. They are now a part of our manufacturing design. Increasingly, we are using these collaborative robots which are robots working together with humans. As per my estimate, I think 45 or 50 cobots are in operation.

Currently which are Bajaj Auto’s main export markets?
They are fluctuating at the moment but Africa, Nigeria, Egypt, Columbia and Latin America are our big markets. We have also recently penetrated ASEAN (Malaysia and Indonesia). Sri Lanka is a big market too. We have been fortunate because when one part of the international markets is not doing well, it does not affect us as our export business is spread out to more than 90 countries. 

Can you shed some light on the Bajaj quadricycle and the markets where it is sold?
We are exporting the quadricycle; we are still not allowed to sell it in India. It is not that we don’t want to sell it here, we are awaiting the approval. We were hoping it would come by this time. We believe it will have a major impact in the way we see mobility. We have sold it in Turkey, East Europe and Sri Lanka.

Our intent was never to export it; it was supposed to be ‘Make in India, Sell in India’. Because we were not allowed to sell in India, we said it has to go out and sell and be proven in the overseas markets. Therefore, India may realise that it’s a good product. It was never meant to be an export-driven product.