Signal & The Noise

By Knowledge@Wharton | January 6, 2013

The negative aspect of the information age is that the world is too much with us. Data was generated in earlier periods, too. But very little of it was captured for posterity. Today, big data is getting bigger all the time.

But this unwieldy mass hides many a gem: Enter analytics, which many businesses are using as a way to make sense of the noise. From an organizational standpoint, analytics is defined as the application of advanced quantitative techniques on enterprise and third-party data to help managers make the best possible decision in a given situation.

LatentView Analytics, which has offices in Princeton, N.J. and San Jose, Calif., and a global delivery center in Chennai, India, is one of the cutting-edge companies in the business. CEO Venkat Viswanathan says that the consumer insights that come from analytics give a competitive advantage to firms that use them strategically.

After the recent presidential election, many commentators noted that in addition to Barack Obama, the real winner was Nate Silver, the New York Times blogger who predicted the results with stunning accuracy. Why did Silver’s model work so well and what does the success say about analytics — to take data and crunch those numbers to predict outcomes?

If you look at what Nate Silver is talking about, it’s this ability to tease out the signals from all the noise. I think analytics as a discipline is really coming [into its own] now as the technology, the ability to store data, the ability to meld together radio signals and come up with the insights are improving each year with changes in technology. We are able to apply this to various disciplines; the application for the elections is just a very recent one.

The Obama campaign itself also analyzed massive volumes of data and they did it to enormous effect. In fact, Time magazine recently wrote about how the clever use of analytics enabled Obama’s number crunchers, headed by Ravid Ghani, to do everything from raising one billion dollars in campaign funds to developing persuasive messages to get people to vote. What are some of the important lessons that business people can learn from such experiences?

In analytics, the number one ingredient that you need is vision from the top and sponsorship from the top. I think the Obama campaign probably had the belief that they could really compete using data and they had sponsorship right from the top. In fact, I remember having conversations with some of the members of the Obama campaign more than two-and-a-half years back. They were starting the preparations for the campaign and were collecting all the data sources. They were trying to understand where they should focus energies. What they have done is something that was doable even earlier. But no one had taken an organized approach to doing this. And they did it in a very systematic manner in terms of collecting all the data, ensuring they prioritized where they put limited funds to use, and getting the sponsorship that they needed. So, they had the right data, the right organization, the right process and the right vision. That made the difference.

What would you say this means for businesses that want to use analytics for equally effective outcomes?

I think the biggest lesson that we keep discussing with our clients is that we need the self-belief that has to come from the top. The organization needs sponsorship from the management that [focusing on analytics] is a direction that [the firm] wants to go. Look at how Amazon is using data as compared to others. Look at innovations like the Elastic Cloud. The biggest difference is that Amazon is proactive and aggressive in leveraging all the data assets. This comes from the vision they are laying out — this is the direction we want to take the business. So, I think from a business perspective the number one thing leaders can take away is the need to have a significant vision. Then build the infrastructure that is needed, which will allow you to collect all the data. And have a plan to find the talent.

What’s an example of a way in which a small company uses this Elastic Cloud?

There are many examples where businesses are built on top of the Elastic Cloud. Take Foursquare, which is a consumer service that allows you to make your presence noticed in retail locations. Their entire product is built on top of the Amazon Cloud. Even our own product innovation initiatives tend to rely on Amazon. It gives us the opportunity to build very scalable solutions. As we gather more and more data, we don’t need to scramble to find the hardware or software support. We already have it. We just need to turn on the tap. We are able to do calculations that used to take us weeks in a matter of minutes. That’s a significant productivity saver as well.

I hear a lot of buzz these days — and perhaps a little bit of hype — about big data and this is often in conjunction with trends like social networking, cloud computing and mobility. These are all described as forces that are, in many ways, reshaping the business world. Do you think these forces are converging? What are the challenges that companies face as a result of these forces?

I agree with you, there is a certain amount of hype to all the buzz that we hear around big data, but it is a very real trend. When you [look at] the sources that have created this data, you can see the dramatic change that is happening in the business landscape. Mobility, for instance — the number of smartphones that are available and being used today. Each of these smartphones is a computer creating real-time data with location information as well, which is a potential goldmine for multiple businesses that want to understand consumers. In five years, the social networks have gone from 100 million users to over a billion users. And each of them creates significant content, the user-generated content that [serve as] signals that businesses need to glean insights from. So, I think there is certainly a trend. The biggest difference this time is the speed at which things are changing. It used to take 10 years to reshape a particular sector; we are now seeing it being reshaped in three to four years. And with that, there is also the falling cost of data storage, the ability to compute that we just discussed, the access to talent … all of this leads to an opportunity where the pioneering companies can actually create business models. This gives them an advantage over other companies that are slow to adopt it.

More and more, companies are trying to figure out how to become social enterprises in the sense that they use their social networks to get their employees to share knowledge with one another or to engage with the customers and so on. And, predictably, this generates large volumes of data. How can companies mine this data and put it to good use?

I think there is certainly a lot of opportunity, but I would say the opportunity is different for different types of businesses. There are consumer businesses, which have a lot to gain because they have direct access to consumer insights in terms of what consumers are talking about. The part that you’re talking about is more relevant for large corporations where they can tap into the collective wisdom of their own employees and they can make the connections internally within their organizations and create this ability to trade knowledge very quickly within their organizations. So, systems like Yammer, which Microsoft has recently acquired, have created a platform where employees can share content in a format. I think more than the content itself, the beauty of these applications is that the relative ease and familiarity that people have with a platform like Facebook is replicated in enterprise space. If an organization creates certain incentive mechanisms where you are incentivized to share more of your knowledge and become an expert in a particular field, it is certainly possible for companies to tap into this.

When companies use social media, they often focus on metrics like Facebook friends or likes, Twitter followers and so on to track their success. What value do such metrics have? And what metrics should companies use to determine the effectiveness of their efforts?

The metrics that you mentioned are what I would call first generation metrics…. The absolute numbers themselves may have limited value. It depends on the sectors you operate in, the kind of consumers you have access to and the type of brand power that you have. So, I think it’s very important for companies to have benchmarks. One of our partners is a company called Unmetric and they’ve developed a social media benchmarking platform, where every brand can benchmark itself against four other selected brands.

But isn’t it possible to easily gain these metrics in the sense that aren’t there companies that will even sell services like increasing the number of likes? It skews the picture, doesn’t it?

Sure, which is why the absolute numbers don’t necessarily convey what is happening out there. What is important is to try and see how these metrics change over time and what is happening in the real world when this is happening? The other aspect is you need to go beyond basic metrics into what I call “engagement metrics,” because that’s when you understand whether the followers and “likers” of your brand are actually engaging with your brand. Do they have a point of view? Are they communicating that point of view in an articulate manner, which will then become an opportunity for us to derive insights? So, a lot of the work that we are doing today for some of our clients focuses on what I call the influencer analysis, where we are trying to identify whether it is a consumer world or a business-to-business world. Who are those influencers who shape opinion about our products and about our competitors’ products? We can engage with them and provide them all the content that is needed so that they make a more informed judgment.

Once you have identified the influencers, how do you work with them?

I think it’s no different from how companies have been engaging with influencers in the real world or the offline world over the past 50 years of marketing. So, you still need to give them all the inputs that are needed about your brand — maybe be proactive in addressing all the questions that they are raising about it. And then try and give them enough input so that they can potentially see your point of view in terms of what are the benefits and values that your product brings. And if they do buy into it, you create a virtuous cycle, which is going to lead to more opportunity for you and a better metric for your brand.

What are some of the most common mistakes that companies make with regard to analytics and how do you help them correct those?

I think an overemphasis on math and techniques is sometimes a bias that certain people have because of the backgrounds they come from and because of all the hype that is being created around predictive analysis. And this is something we need to hold back people on. We are saying it’s not about the math, that we need to go beyond the math. We need to keep the analysis grounded on what can be actionable and how we can take it to market. So, that’s one aspect we have looked at. Maybe ignoring external sources of data is another bias that we have seen. Companies seem to go after the low-hanging fruit, which is data that they have already collected within their business. But in reality, in such a complex world that we operate in today, it is very important to pick signals from other sources as well. They may have a lot to tell you in terms of what is happening in your business. The third aspect is analytics being run as a kind of technology or data initiative. It then tends to sit in a silo and is not integrated within an organization. It has to become mainstream for it to have an impact. So, it almost should not be called analytics if it has to have impact.

In any entrepreneurial journey, you have both successes and failures. Which has been your most instructive failure?

Very early on, when we were still relatively young as a company, we made a choice, which in hindsight we could have made differently: we spent a lot of time focusing on servicing a shallow market. We spent maybe the first two years — an enormous amount of time — focusing on India as a marketplace. I believe India has a lot of opportunity as a marketplace in the future. But maybe we got our timing wrong.