Marico taps analytics for growth
- IndusInd Bank launches m-wallet with MobiKwik
- Bharat-22 ETF likely to be launched in November
- Donald Trump will decide on Federal Reserve chair ‘very shortly’
- Rahul Gandhi: The voice of Gujarat’s youth can’t be suppressed, purchased
- India Post Payments Bank appoints Vodafone M-Pesa’s Suresh Sethi as MD, CEO
Mumbai: Till about two months ago, Salim Gulam Mukhi, a key distributor of fast-moving consumer goods for Marico Ltd in Mumbai, used to place orders and wait “almost a day” before getting the goods delivered at his warehouse.
“Adding, or changing, orders was cumbersome,” recalled Mukhi, who has been a distributor for the maker of Parachute and Saffola oils for the past 24 years, and serves areas between Lower Parel right up to Virar—the last western suburb of Mumbai.
Mukhi today is a happy man because it takes just “10-15 minutes” for an order to be delivered. As a result, he can stock fewer goods and the lower inventory is helping him cut down on warehouse space. “It has also reduced the waiting time for trucks, which has resulted in lower investment and costs,” said Mukhi, whose annual turnover from the business is about Rs.4.5 crore.
He is just one of the 5,000-odd distributors and stockists of Marico who have benefited from an analytics-driven ‘Order Management Execution System’ that the company rolled out in December.
According to Mukesh Kriplani, Marico’s chief of business process transformation and IT, “This system is live for our distributors in the entire West region. And we are now extending the facility to the East.”
Orders were earlier placed at allotted intervals, and changes had to be communicated over the phone to all the stakeholders, including the distributor, depot manager and the officials concerned at Marico. “We have revamped the distributor system by pulling stock and sales information every two hours. This is visible to the distributor and he can effect changes within broad limits,” explained Kriplani.
In this first phase of Marico’s business process transformation, which the company hopes to complete by January 2016, “there was a lot of non-SAP data, for example, data from the product-wise schemes that are run by modern trade outlets”, according to Kriplani. “We had to link that data to sync with our product codes, since the actual sales are happening in their outlets.”
It’s not that Marico did not have access to such data earlier, but the analysis was not done in real time. Besides, storage of all that data would have proved very expensive.
Marico already had an enterprise resource planning (ERP) system from SAP AG, which it upgraded to SAP HANA to consolidate data, partly on the cloud. But the company also had to deal with unstructured data. Hence, the data from portable document formats (PDFs) and raw tables were copied from the existing system’s database into SAP HANA. “Now it is possible to pull all that data and store it on the cloud, resulting in lower storage costs,” said Kriplani.
Marico also uses Tableau, a product from business analytics firm Tableau Software, which helps it analyse a large amount of data on retail behaviour, sales and marketing, inventory movement and procurement of key inputs, and get detailed insights into the company’s performance in a visual format. The company also uses “sharepoint-based collaboration portals, ETL (extract-transform-load) tools, and R software”, said Kriplani.
According to Saugata Gupta, managing director and chief executive officer of Marico, “Business analytics, digital and automation is helping to transform our core operations, improve our consumer insights and innovation processes as well as help us take better decisions.”
Transformation of the company’s business processes, according to Kriplani, became necessary since operations were becoming “complex and challenging” with acquisitions and addition of new product categories.
For instance, “we acquired Paras Pharmaceuticals around three years ago and that was in a category that was different—more chemist and cosmetic. Moreover, our growth in foods such as Saffola has also been high. Again this is a slightly different category with different kinds of channels and different kinds of customers. So it was necessary to handle this wider category and wider range of SKUs (stock-keeping units) along with the traditional products that we make”, explained Kriplani.
On the sales side, Marico is implementing a ‘Sales Assortment Mix Analytics’ model in south India “as a pilot”, said Kriplani, to help the company predict selling and buying patterns across similar territories.
“We know, for instance, that consumers in the Pali Hill area buy certain kinds of products. Consumers in Jubilee Hills in Hyderabad may have similar buying patterns. The analytics model will be able to suggest if SKUs that sell in Pali Hill area can also be sold in Jubilee Hills. If it’s a ‘yes’, then you have cracked a new SKU and increased your reach. This particular model is in the pilot phase; we are doing it in one of the cities in south India and the early results are encouraging,” said Kriplani.
All such data, including the outlet format and pin code, is fed into the software. The model then runs an algorithm and presents a hypothesis that needs to be tested on the ground. One such test would be to match the sales forecast against actual sales, according to Kriplani.
Marico is also going beyond traditional trend series analysis of data by taking into account “all your sales marketing spends, maybe other events which are being held like promotions or maybe some other activities like certain specific festivals, or even certain price increases”. “All these can be inputted to get an output and a forecasting accuracy that will result in better service and faster movement of goods,” said Kriplani.
Alok Shende, founder director and principal analyst of Ascentius Consulting, believes that “R as a software environment for statistical computing and graphics brings down the ticket price for gaining insight from the data”.
“The R language is widely used among statisticians and data miners for data analysis and now has a robust community available in India. Conflation of R and SAP HANA brings together the power of open source R and in-memory SAP HANA, leading to a significant improvement in the kind of questions that can be asked of the data and the shorter time to derive insights,” said Shende.
In the second phase, Marico plans to sharpen its focus on social media analytics, besides enhancing the use of analytics in its factories.
“Manufacturing is all SAP-based. It is not that there are no sensors in the factory but if we are able to capture real-time updates, we will have to increase our storage space. Today, we collect data every hour. That is when we will have to do a pilot. But these are early days,” said Kriplani.
He insists that Marico’s distributors are benefiting from “the reduction in inventory days, which means your in-built turnover ratios lead to a better throughput”. He also points out that since the company’s categories range from foods to cosmetics—products that have an expiry ranging between six months and 36 months, “certain categories have to be fresher than others”.
“We now can make specific SKUs move faster and handle production accordingly. This also implies that our distributors will be assured of fresh stock, regardless of whether that falls in the food or cosmetics categories,” said Kriplani.
But when will Marico see the fruits of this business transformation? The company posted a full-year turnover of about Rs.5,700 crore as on 31 March. However, according to a 30 April Mint report, Marico’s January-March quarter growth was crimped by the roll-out of the new distribution model. Analysts believe it will be a while before the transformation helps Marico’s volume growth return to normal levels.