by Sneha Jha

CEAT’s Investment in In-memory ERP Brings it On the Fast Lane

Jun 19, 20147 mins
BusinessEmail ClientsEnergy Industry

As a sluggish ERP began to swallow CEAT’s revenue, its CIO took the wheel and steered the company toward in-memory ERP, changing the company’s fortunes. 

Niranjan Bhalivade was a worried man. The frown lines on his forehead got deeper as he hunched over his laptop to read an e-mail from Tamal Saha, SVP, sales and marketing, at CEAT.  The e-mail—marked to the MD and LOB heads of the company—was Bhalivade’s worst nightmare. Saha had written about the company’s snail-paced ERP and how it was beginning to gnaw at CEAT’s profits. As the CIO of the company, Bhalivade wasn’t thrilled.   During the company’s monthly closure, the slow ERP delayed dealer billings. Saha notified that even though they had orders, some of the dealers missed their slab discounts. Discounts are given based on the invoicing done by the dealers. In slab discounts, the off take of particular SKU’s or product categories are important. If the invoicing is not done on time, the dealers will miss the slab numbers and would get less discounts on the product which they could sell in the market. This would result in serious issues with these trade partners and loss of business. And the frontline sales team had to bear the cross.  As he reached the end of the e-mail, Bhalivade slumped heavily on his swivel chair and groaned in frustration. Once again the problem had reared its ugly head. A couple of months back, during the company’s quarterly closing in March, invoice processing, billing, and report execution took a very long time. On an average, over 3,700 invoices were generated in a day. At month end, the number swelled to 7,300. Due to system performance issues it took about 7-10 minutes for a single invoice to get through. And a large number of invoices at monthly and quarterly closures only aggravated the problem.   “The average daily billing is Rs 14 crore, making every hour critical for business. Hence, the risk of revenue loss increases at the quarter or year end. The sales staff would stay up after hours to complete the billing. Despite that, reports would get ready only the next day,” says Bhalivade.  And this inefficiency pushed the sales team over the edge.  The tyre manufacturer also ran 148 discount schemes which are executed for 4,000 customers in more than 900 SKUs. The discount execution process took nothing less than 72 hours.  This had a huge impact on business. Even a one hour delay during quarter or year end could lead to losses to the tune of Rs 6-8 crore. Apart from the potential revenue loss, CEAT was losing 775 man days a year. And this was when all the systems were working at 100 percent load. As CEAT’s dawdling ERP began to eat into the company’s profits, Bhalivade knew he had to do something—and fast. Pedal to the Metal A company of CEAT’s calibre hardly needs an introduction. A part of RPG Enterprises, CEAT owns three manufacturing facilities in India. Its Sri Lanka and Bangladesh operations have added 18 percent additional users.  Business was growing at a blistering pace. And IT had to enable business growth to handle the additional 60 percent data volume and 20 percent ERP users.  At the same time, the business required a high availability and 100 percent uptime with faster processing power. And that solely depended on CEAT’s ERP system.  The tales of the company’s sluggish ERP didn’t really come as a rude shock to Bhalivade. He knew that the system was dragging the company’s business down. Which is why he had already planned and budgeted for a system upgrade project. But Saha’s mail set the alarm bell ringing. “We thought if it has to be done then let’s do it now, the sooner the better,” says Bhalivade. But was upgrading server memory really the answer?  For one, it wouldn’t be a self-sustaining move and it wouldn’t shrink the response time. “If it took seven minutes for the invoice to be generated now, then it would take five minutes after the server upgrade. But the time taken for generating an invoice would not reduce drastically,” says Bhalivade. Worse, after six months, it would require another upgrade and that means the company would have to dole out more cash.  That wasn’t Bhalivade’s biggest problem. What he wanted to do was ambitious and bold: Cut down the time to execute reports from seven minutes to 10 seconds. It was clear that a mere server memory upgrade wasn’t enough to do the deed. This provoked Bhalivade to look beyond the tried-and-tested.   And that brought him to a fairly new but powerful technology: In-memory ERP. “We decided to take a big leap and run our ERP from the memory itself. This meant that it doesn’t have to go back to the hard drive, search the data, process it, and then give it back to the user. And this also ensures that all calculations happen on the fly,” he says. That’s exactly what CEAT wanted. An excited Bhalivade pulled up his sleeves. It was time to get down to business. Picking Up Pace Bhalivade wasn’t dispirited by the enormity of the project. In fact, the level of difficulty egged him on to prove his team’s mettle.  That’s why he was undaunted by the fact that he had to migrate 1,000 users to in-memory ERP in just 100 days. Or that the project had to go beyond borders to include users in Sri Lanka and Bangladesh.  Bhalivade constituted a 33-strong project team and organized a three-day internal training program. His objective was to shrink the time taken in report execution from seven minutes to 10 seconds. He gave clear instruction to his team not to transport any program that took even 11 seconds in the pre-production system without his approval. It had to be brought down to 10 seconds before it got transported on the new platform. Clearly, Bhalivade did not want to fall short of target.  His team revamped around 480 programs and optimized them on in-memory technology. Before the project went live in January this year, Bhalivade arranged for two major mock drills in December. Planned with military precision, the mock run covered 150 users in India for 90 minutes. The drill threw up a lot of glitches which were resolved within a week.  The migration was executed in 52 hours. Bhalivade wanted it to be a seamless transition for business so he undertook the migration on Friday evening after the system shut down. He expected to accomplish the task by 11.30 am on Monday. However, he completed it by 7 am. The project had a tight timeline of 100 days but it was completed a day before the deadline. The new technology platform is well equipped to process large data volumes. Its enhanced capacity helped CEAT address its ERP performance issues by resolving the problem of slow computing and data presentation. It provides a strong IT support and scale for the company’s rapid growth and business development.  The benefits of the project outweigh the cost. Earlier the company had six-seven racks of servers. With the new technology the entire ERP has come down to just one server. The company has realized considerable savings in power and datacenter costs. “One of the direct benefits of the project is that we have saved 60 percent power and datacenter cost.  We pre-empted that the ROI of the project would be achieved in 18 months but within six months the project justified the investment,” he claims.  Efficiency has been increased to the tune of 50 percent because of the faster ERP. Invoicing time reduced from seven minutes to 10 seconds with all possible and required controls, locks, and validations. There are no deviations in performance even in peak seasons or quarter end transactional rush.  “After going live, we introduced our first round of discounts. Earlier it took us 72 hours to execute the process.  But we were able to complete the discounts along with credit notes within 10 minutes,” shares Bhalivade. The CFAs and territory leaders no longer have to put in extra hours. So their work-life balance has also improved. And as CEAT drives on the fast lane, Bhalivade is a happy man.        Sneha Jha is special correspondent. Send feedback on this feature to

Earlier, it took us 72 hours to execute the discount process. With in-memory, we are able to complete it within 10 minutes.