Thursday, October 20, 2016

No data to big data

No data to big data

As data grows in size, it may also be wise to remember that not everyone in the team can look at a large database, do examine the use of data visualization tools

There was a time, not too long ago, when we had to go street to street looking for customer data. Honest truth. When working on a leading automobile brand in the early 2000s, we discovered that what the company had as customer data was in fact the data (often the registered office address) of the car leasing company. That was of no use to us in marketing, or at least of very limited use. We needed the names, addresses and phone numbers of the people who drove the cars. Did the dealers have the data? Well, we discovered that the dealers did have the data but it was on their service records (obviously when the car came for service it did not come from the leasing company).

This data looked rich, so for a few years, we employed field investigators to physically visit the 100+ dealer points and collect authentic customer data. This was revalidated by back-checks. We managed to assemble almost 75% of authentic names, addresses, phone numbers—the database of the direct marketing division of the agency which was entrusted with running the customer relationship programme for the said client. Was this good enough? A regional direct marketing expert was called to audit the programme. He was amazed at the way the data was collected and, to our delight, told us that 75% was indeed very high. Apparently, even in developed markets, manufacturers struggle to get real customer data—the challenge in those markets is that the dealer feels that the customer belongs as much to the dealership as to the brand.

From a situation of No Data we moved towards Some Data, but with a lot of physical effort.

The situation improved dramatically with the automation of all dealer points and state-of-the-art point-of-sale systems which collected not just the name of the owner (the leasing company) but also the name and address of the primary user of the car.

But just a name, an address, a phone number, a mobile number, an email ID, does not amount to Big Data, as we all know.

In the book Big Data in Practice, Bernard Marr chronicles the way Big Data is being used in 45 different companies (US-based) including the usual suspects like Google, Netflix, IBM Watson and Amazon. The book also speaks of interesting areas where Big Data is put to use. For example: Walmart, after analysing a humongous amount of buyer behaviour data, discovered that before a typhoon hit a town, the two top items bought and stocked by householders are batteries/torchlights and, wonder of wonders, strawberry shortcakes. We could have guessed batteries and torchlights, but not in our wildest dreams would we have cottoned on to strawberry shortcakes, supposed to be the No. 1 comfort food consumed in big quantities during an impending storm (wonder if Big Bazaar sells out of potato wafers before cyclonic storms?). Or take the case of Caesar casinos in Las Vegas; they figured out that by tracking data of regular visitors to their casino, the real potential to sell lies not with the high rollers but the average Joe who spends around $100-1,000 per visit. By focusing on this large cohort, Caesar became one of the best performing casino chains in the US. 

The author predicts, quite rightly I would imagine, that the term Big Data will soon become obsolete. What we know as Data will be Big Data. As we move from MB to GB to TB to pentabyte to exabyte (in the year 2000, just 3 exabytes of information was created) to zettabyte, our own definition of Big Data will keep morphing.

Where are we in India with the Big Data revolution? Given the huge talent we have in coding and analysis, are we way ahead of the curve?

Just last month, I got an exposure to two very different scenarios. In one, a family-owned, family-managed small enterprise, I discovered that the company had been collecting customer data on a regular basis and was also able to put it to good use. In yet another company, I was asked—‘Will customers willingly give their email id and mobile number?’

So here are just a few key pointers to the companies in the second bucket:

• Start collecting customer names, emails, mobile numbers from tomorrow
• Match the customer data with the purchase data
• Start building customer segments that can be monitored
• Test offers against these segments; maybe 10 a month?
• By using other data collection tools, see if you can add to the customer profile
• Target messages to them through social media
• Keep building a learning curve

As data grows in size, it may also be wise to remember that not everyone in the team can look at a large database, do examine the use of data visualization tools. They help the unlettered appreciate the finer aspects of data segmentation and analysis.

While you probably were aware of the fact that 70% or 80% of your sales came from your regular customers, you will now know who they are and will be able to focus a large part of your marketing budget at them.

So, instead of just mouthing empty words about Big Data, start building your business by using data more productively. It is really a lot simpler than you thought. Zettabytes can wait!

Ambi M.G. Parameswaran is brand strategist and founder Brand-Building.com. He will take stock of consumers, brands and advertising every month. The views expressed are personal.


Regards

Pralhad Jadhav
Senior Manager @ Library
Khaitan & Co


No comments:

Post a Comment