Oct 27, 2015
Published By : Marketing Magazine
Though global marketers haven’t yet perfected the art of turning data into insights, many of them are well on their way, according to the CEO of leading global data firm Fractal Analytics.
Founded in 2000 in India, Fractal has grown to become one of the world leaders in customer analytics and business intelligence, serving clients like Procter & Gamble, Kimberley Clark and CapitalOne across global markets where they operate.
In the past year, Fractal has expanded its presence in the Canadian market through a major global partnership with Aimia, which now offers Fractal’s customer behaviour analysis and big data marketing services to its loyalty clients, as well as using Fractal’s tech to provide insight on its Aeroplan program.
On a recent visit to Canada, Fractal’s CEO and co-founder Srikanth Velamakanni sat down with Marketing and talked about how close the multinational brands it works with are to achieving the dream of fully data-driven marketing.
He said when it comes to transaction data — the most abundant and powerful form of customer data that most companies have access to — the world’s top marketers have pretty much “cracked that problem.” The P&Gs of the world have reached a stage of maturity where customer data is properly captured at a large scale, converted into reliable, actionable insights in near-real time, and applied to most major strategic decisions.
Some of Fractal’s most progressive clients won’t commit to any major business move until their chief analytics officer gives the go-ahead. “Traditionally, strategic decision-making has been relegated to gut feeling. Today you’re seeing analytics playing a bigger and bigger role in strategy,” he said. “Should I do this today or not? That never used to be an analytics question.”
Where the frontrunners are focusing now is on learning to apply analytics not just at the executive level, but as a cultural practice across the organization. “They are looking to democratize the data, give people access to actionable insights in real-time — on their phones, on their ipads, whatever,” he said. All the top analytics and automation platforms — Fractal, SAP, SAS, Salesforce — have developed mobile dashboards for marketers and executives to check in on real-time revenue or spend projections whenever they need to make a decision, whether they’re in the boardroom or on the road.
“How do you make everyone more analytically driven rather than relying on intuition or gut feeling? I think that’s the biggest change that companies are trying to make,” Velamakanni said. “In the next three to five years businesses are going to get a lot better at that.”
Compared to the U.S. market, Canada has less reliable data about consumers. There are no retail data wholesalers like BlueKai or Datalogix, and organizations like telcos and credit card providers are much further behind in packaging and monetizing their data. The data available can often be outdated.
But, Velamakanni said no matter where you look, businesses never feel like they have enough data.
“With data it’s always a question of relative poverty. You talk to the credit card companies, they’ll say look, we don’t know what the consumer is really buying because we only see how much they paid,” he said. “If you go to the insurance company they say look, the credit card companies have so much more data, but our data is so sparse because we see much fewer transactions. Everybody is poor at some level.”
But, the reality that Fractal sees is everyone already has access to a mountain of data they aren’t leveraging. For some companies that means they haven’t fully exploited their own transaction data; for others it means they haven’t explored publicly accessible data sources like social media, or even census data, which when combined with internal data sources can provide powerful metrics including your share of the customer’s wallet.
“Most companies have not looked at their own data sources very well. They have not combined their email data or their unstructured data sources, they’ve not looked at those data sources at all,” Velamakanni said. “If you bring those types of data sources in I think you will get a lot of value — and then you can start looking at external sources of data.”
For companies that are just beginning to take advantage of their internal data, it can be tempting to set out an ambitious plan to comprehensively capture and analyze every contact point they have with the consumer. That’s a mistake, Velamakanni said, and it’s part of the reason many CMOs feel they have more data than they know what to do with.
“We see two kinds of approaches to adopting analytics. One is companies who say, ‘Let’s get the data together, and then once we’re done that let’s start solving some problems,’” he explained. “They are generally much less successful. You can continue to boil the ocean and you’ll never be done with that.
“The other set of people are much smarter. They’re essentially saying, ‘Let me solve a problem — let me get the data together just for that problem, and then let me build the infrastructure over time to enable that and more.’” The incremental approach, he said, leads to more demonstrable, short-term wins, which help build up support at the senior level, and at the same time slowly build towards the comprehensive infrastructure and expertise that businesses are ultimately looking for. It also helps instill the idea that building an analytics program is an ongoing progress — one that will never be finished, as there will always be more data to exploit.
As far as how reliable the data is, Velamakanni said it’s important not to get too caught up in whether every database entry is pristine and infallible. Analytics won’t ever be able to predict whether any one person will buy an iPad or an Android. What it’s intended to do is give you a better than 50/50 chance of finding someone who will.
“The idea is, can you be better than random? Can you be better than before?” he said. “That’s what you’re always looking for — you’re not looking for perfection. This is not an audit, it’s not a forensic investigation. … All information is not 100% accurate, but there is enough signal in that noise to make us better at what we’re doing today.”