The next time you’re downtown, stop and look around you: people, stores, banks, transit, restaurants, stoplights -- all of them constantly generating and consuming data. Now think back to the people, all of them with their own destinations, purpose, concerns, needs and schedules -- more data.
Since 2000, when he co-founded Fractal Analytics in Mumbai, Srikanth Velamakanni has been looking closely at the data that defines our lives, our jobs, our towns, even ourselves. Fractal has helped scores of clients sort it all out to better serve each customer, analyzing client data along with its in-house data warehouse to provide near real-time solutions.
Fractal moved to New Jersey in 2005, then relocated again to San Mateo, Calif. in 2010 to be closer to Silicon Valley. It just opened an office in Rome, will soon expand to Switzerland and has already opened in Canada. Today it provides data analytics services to companies with revenues of $10 billion to $100 billion in sales, deriving 55 percent of its revenue from the retail/packaged goods sector, 40 percent in financial services/insurance and 5 percent from technology and telecom.
Fractal's New Market
The company is about to branch into the area of life sciences and healthcare, perhaps the hottest sector in the US economy with the implementation of national healthcare underway. Just as data can help a retailer understand a customer’s needs better, Fractal believes it can help doctors to better understand the needs of their patients.
CMSWire had he chance to sit down with Velamakanni and Fractal CMO Careen Foster to discuss what companies to do to better meet the needs of their customers while leaving consumers in control of the data in their lives.
Murphy: When I think of predictive analytics, I think of it as part of a rainbow. These days, I think of customer expectation management as being most of that rainbow. How do you think about customer expectations in relation to analytics?
Velamakanni: When we talk about customer expectations, there are two aspects to it. Obviously, there's a whole host of data that we have about customers, their transactions with the firm, their data transactions and other things. Using all that, we can have a 360-degree view of the customer -- what they like, what they don't like -- we can get really, really specific. Are you a Nikon guy or a Canon guy? Do you like iPhone or Android? Microsoft of Google? Every single thing. As a retailer, you can understand your customer in a similar way -- hundreds of thousands of variables. So the first part is to understand customers truly by understanding transaction data.
The second part is about solving their problems in real time, depending on what they're doing right now. So if I'm standing on the street and searching for a restaurant at 12:30 and I have a preference for Chinese cuisine and I tend to spend $100 per meal, now I know which of the 50 restaurants in this neighborhood works best for me. I can solve the customers problem in real time, using all the stuff I know about this customer. Once you understand customers and you can solve their problems, you can get the kind of loyalty that we all want from the customer.
There was a study published a year or so ago that said between Google, Facebook and Amazon, people are twice as likely to share their information with Amazon as with Google or Facebook because the relationship with Amazon is clear -- they're selling and we're buying. It's also clear they won't misuse your information. They will use your information to make your experience better. It's not entirely clear what Facebook and Google will do with your data. Sometimes you feel they will use the information for other purposes, or that you are the product, stuff like that. If you can crack these two pieces of the puzzle, I think you're really meeting their expectations. I know you wrote the book Web Rules. In many ways, what you wrote in that book is what is happening now. It was very futuristic back then, but it's all real now.
Murphy: I was going to ask you about customer expectations for privacy, but you got there first. There are some generational differences, and perhaps some cultural and geographic differences, between people who do or don't give their information online. I think the reason I don't like it when Facebook tries to push ads on my page is because their analytics aren't as good, and the ads don't match my interests.
Velamakanni: Customers expect to be in control. If you can be transparent with them, say "this is what I'm collecting" and "this is what I'll delete." Google is doing a little of that. You can see all the data, and you can say "please don't use this information or this information." Not everyone uses [those options], but they're definitely giving the control to you. You can control what we share and what we don't share. That's what customers expect, and they want to understand that if I let you use this information, this is what I will get in return. I think most of the world is there.
With Facebook, I don't have any trust in what is happening to my information. All the privacy controls are hidden in so many different places. They're trying to make it better, but a year ago it was not possible to control how your information would be used.
Murphy: We are somewhere down the path towards real-time analytics, but we haven't mastered the art yet. I suspect it will be years before it's perfected. Someday, when I'm looking for a quiet place for us to talk near Union Square, I'll be able to go to my phone and get that. But that takes a lot of information and to crunch all that data in real-time is something that is simply not possible today. How long do you think it will be before we get to that totally connected, IoT kind of world?
Velamakanni: The ultimate vision may be many years. My sense is the ultimate vision will be redefined. We don't even know what's possible now with billions of sensors and all these cameras. It is changing. Our expectations 10 years ago were very different from today, and I think we're already close to meeting many of the expectations from 10 or 15 years ago. It's a moving goal post and I really don't know if there's a date where we can say we will meet customer expectations for real-time analytics. But it's getting really, really sophisticated as we speak. Look at maps, for example. What was our expectation 15 or 20 years ago for maps, and look at what's happening today with real-time traffic reports.
This morning, I was supposed to be some place at 8 am, and I was driving with a friend who was using Google Maps with Waze included, giving real-time traffic information. But we ended up reaching the place later than if we had not used Google Maps for one reason: Google Maps doesn't include car-pool information. It gives you the best route if you are not using the car-pool lane. Maybe somebody will write that app and it will be the next Waze for $1 billion. But you can see that expectations have changed. My expectation this morning was that it should have realized we were in the car-pool lane and given us better information. But that's because I have already done so much that my expectation has changed.
Murphy: The thing about expectations is that if you don't have any, then everything is fine. If you have expectations, and they're not met, then the company is in trouble. You can tell so much about what your customer are thinking, but you don't know the rest of it -- and that hole is the hard part. If an analytics system knew we were looking for a restaurant to sit and talk, it could have given us a lot of suggestions, but most wouldn't have met our expectations.
Velamakanni: Today there is no way to even ask this question. To say, "I want a quiet place" -- systems don't understand human language and intentions. Yes, there's a long way to go.
Foster: We were just talking about this with one of the online websites for travel -- the need to break out even more granular information in preferences. For example, I don't just want a fitness center, I want very specific fitness center stuff. How do you tailor it so I can find the one hotel that is going to meet my needs?
Murphy: If I was searching for blue jean prices on my website, and I drove to the mall and got a text saying "20 percent off on all blue jeans" today, I might check it out.
Velamakanni: I think that one thing the customer expects, but doesn't always get, is that the company will have all the information in all the channels. So if I search for something and you go to the store, your expectation is that all that information is reflected in how the store knows me. If I see 20 percent off online, and I go to the store, and the 20 percent off is not there, it doesn't make sense. The expectation is that my experience is consistent and the organization has my information at hand at all points in time. It's such a basic expectation that all of us have, and it's not fulfilled right now. My web experience is different and if I call the call center, and they have no idea what happened, I have to start all over again.
Murphy: Fractal is in a crowded space. There are a lot of companies with services like what you're doing. How can you play in field? Is it getting too crowded? Will there be a shakeout? Are you looking for an acquisition?
Velamakanni: This is a hot space from the standpoint of clients trying to solve these problems. The expectations are huge for what this space can deliver. There are maybe 50 books on big data right now, and they all talk about what Google is doing, what Netflix is doing, or what somebody is doing. When you look at the real-world companies, there is a huge gap. Most of the companies we talk to are $10 billion to $100 billion in revenue. Most companies we talk to are in the middle of that range. Our goal is to move them from the lower left-hand quadrant to the upper right-hand quadrant. [Editor's note: By comparison, Google had 2013 sales of $59.8 billion.]
Murphy: It's probably hard to sell when you compete against big companies -- big blue companies, for example. They can say "Here's the solution, we can put it together." How do you compete against that?
Velamakanni: We sound exactly like IBM would sound, except there's another zero on their price tag. IBM can do everything that we can. Unfortunately for them, they have this software business, they have this consulting business, they have this hardware business. On paper, do they have the capability to do exactly what we're saying? They do. Operationally, can they actually achieve it? It's really hard for them unless you're a Fortune 100 company and you're willing to say, "I'll write a $50 million check." Most of our companies aren't ready to write a $50 million check. We can start with maybe $1 million and scale-up for $10 million, $50 million or whatever they want.
Murphy: You guys have the analytics part, but you don't have a lot of the other parts. If you go into a company, they already have a CRM system, they already have a content management system, they already have a sales management system, they have a contact database and they have collaboration tools. So you come in and you talk about analytics. They must say "Well, that sounds great, but how do we hook all this stuff up?" What's your response to that?
Velamakanni: That's actually a key advantage for us. We don't have anything else to sell, no other fish to fry. We say, we can work with all the data that you have. Now that they have invested in all the technology in the world, and haven't solved their problems, we can get this data, solve the problems as a result. And we deliver it all as a service to them -- there's no integration work. Whatever software is required, we do at our end. We will plug right back into their data warehouses, or campaign management systems or whatever else they have. We' completely agnostic about their data infrastructure. All we want is to populate their data infrastructure with the data and the analytics that they want. We fit in quite well with what they have.
Murphy: Is everything you do cloud-based?
Velamakanni: Most of it is cloud-based. The data and analytics are running on our on-premises cloud. Only when clients are comfortable are we willing to put it in the public cloud. It's very rare. So far, our clients haven't really embraced that.