Like my colleague, Bill Schmarzo, in his recent Point-Counterpoint: Eight (No, Nine!) Problems with Big Data blog where he takes on the New York Times, I felt the need to comment on an article I read in the Washington Post entitled “Five Myths about Big Data” by Samuel Arbesman [August 18, 2013, Outlook Section, Page B2, The Washington Post].
I feel some of the observations were off base—which I will address shortly. Some I actually agree with—but with a twist. Yet the article points out a key issue that we big data specialists have to deal with: these perceptions are out there. In addition, as conventional wisdom states, “perception is reality.” Coming from the author, who is an applied mathematician and network scientist, this commentary predictably takes a jaded view of the need for, and benefits of big data analytics. Let’s address each one of his myths:
| Alleged Myth about Big Data (per Samuel Arbesman) | Key Argument that Supports Myth |
My View |
| “Big data has a clear definition” | “There are a lot of definitions, a lot of confusion and industry experts often end up talking past one another.” | My knee-jerk reaction is “so what?!” Just take a business-focused approach. What is the business problem that a communications service provider is experiencing? How can you improve service and network quality for your best customers? Then, dig into the datasets with detailed analytics. Borrowing a phrase from that great movie, Apollo 13: “Work the problem!” |
| “Big data is new” | “Big data isn’t much more than a sexier version of statistics, with a few new tools that allow us to think more broadly about what data can be and how we generate it.” | I just don’t agree with the author on this. In the telecom world, typical network data provides fundamental information regarding alarms and faults—the most typical being hard outages. Big data analytic data techniques allow service providers to monitor network quality in two ways that are radically new:
1) Analyzing massive streams of network performance data enables network operators to detect subtle trends that, if ignored, lead to significant outages. 2) Big data analytics applied to data from a customer’s viewpoint provides deep insight into how your customers actually perceive network quality. Service provider monitoring systems are often unable to obtain this insight. |
| “Big data is revolutionary” | “When a phenomenon is large, we usually don’t need huge amounts of data to recognize it. Revolutionary for an individual? Probably not.” | First, I agree that big data by itself isn’t revolutionary. EMC’s approach is focused on coupling big data analytics with speed! Traditional business intelligence platforms can tease out trends from big datasets. These take time—and lots of it. The revolutionary step is performing these analytics in near-realtime. The result is immediate business impact—not weeks or months later. One telecom industry example: a service provider is able to recognize a service problem for high-value customers and proactively contact them to apologize and provide a credit—all within seconds of an event. |
| “Bigger data is better” | “Many interesting questions can be explored with little datasets. Too often, massive data sets lack a temporal dimension. We need long data, not just big data.” | I fully agree with the author’s rationale: massive datasets need a time element. What is unique in EMC’s approach (as described above) is the ability to 1) rapidly ingest massive data, 2) act on it based on analytic parameters, and 3) trigger business-focused actions in seconds. I’d say it this way: EMC’s approach is both big data and fast data. |
| “Big data means the end of scientific theories” | “Having more data won’t substitute for thinking hard, recognizing anomalies and exploring deep truths.” | Again, I agree (sort of) with the premise, but for different reasons. In short, it’s not just “big data.” You need the industry-focused analytics and knowledgeable data scientists and experts to configure big data analytics for a service provider. This is exactly the approach that EMC takes. We invest in the hard thinking and bring starter kits with pre-built analytic use cases, and then do local configuration for each specific customer. |
Other objections that I have come across in my discussions with communications provider senior executives – and my actual responses:
| Objection | The Perception behind the Objection |
How I Respond to Objections |
| “Why should I buy a big data platform vs. a point solution to do a specific implementation of analytics/actions for a specific function?” | Commercial off-the-shelf (COTS) solutions are lower cost, already have desired functionality and a roadmap for enhancements. | Your competitive differentiation comes from creating a unique value proposition for your customers. Yes, big data analytic solutions require some tailoring, but EMC solutions bring analytic use cases to every client. These require configuration – not software development.We welcome the opportunity to compete for value. And my final perspective is that you get what you pay for. |
| “Selecting a big data platform – for multiple functions – is hard to sell internally. More functions imply more people to be consulted and approve this large initiative.” | Obtaining corporate-wide executive agreement on a big data platform is difficult. Identifying stakeholders, determining requirements for multiple needs and developing an implementation roadmap are all daunting tasks. | I agree with McKinsey’s perspective: “Select a few high-potential areas in which to experiment with big data, and then rapidly scale successes.” |
In conclusion – don’t give up! Start small, celebrate success and evangelize!
Next Up: Transforming a Business area with IT as a Catalyst – Why is this so hard?
I hope you will join me and will pass on the link to your friends and networks. Please … subscribe, send me feedback, and check back for the next installment. If nothing else, I promise the Travel Tips will be extremely useful!
Today’s Travel Tip: “Must see” sights and favorite gifts to bring back from international travel
Typically, I am asked, “You’ve traveled to X. I am going there…what should I see and what should I bring back?” The following is by no means exhaustive but what I consider the top sights and gifts representative of the countries I have visited. Generally these gifts are readily available throughout the country as well as in the airport for last-minute shopping. We’ll start with the Americas:
|
Country |
Must See Sights |
Gift |
| Argentina | City tour of Buenos Aires, plus a dinner/tango show at Carlos Gardel’s.A concert at Teatro Colon, a spectacular opera/concert hall in Buenos Aires. | Havanna brand Dulce de Leche (a wonderful caramel spread)Leather goods made locally |
| Aruba | Any beach | Anything beach-related: you’ll treasure remembering the time on the beach |
| Brazil | Rio: Sugar Loaf, Corcovado (Christ the Redeemer statue), Ipanema and Copacabana beachesSao Paulo: Fabulous Sala Sao Paulo for an orchestra concert | Garoto-brand chocolate- and nut-covered bon-bonsHavaianas flip-flopsCoffee beans. |
| Canada | Toronto: CN Tower, Hockey game | Maple syrup assortments |
| Chile | Tour of Santiago including the old city. Winery tours. Skiing. | Chilean wine—my favorite is Concha y Toro.Lapis Lazuli, which is a blue stone found in only a few places in the world—including Chile. |
| Curacao | Scuba diving (with a 15-minute lesson, no prior experience is needed) in the National Aquarium amid manta rays, sharks, and giant sea turtles. (The sharks and sea turtles are in adjacent enclosures, but you can feed them through the nets) | Blue Curacao liqueurAnything beach-related |
| Mexico | Plaza de la Constitucion, or Zocalo, includes the National Palace and Metropolitan CathedralTeotihuacan, a series of pyramids about 25 miles outside the city | Tequila, coffee beansSilver jewelry |
| Panama | Panama Canal (Miraflores locks and visitor center)Rain forest excursion | Local artisan crafts. Note that the famous Panama hat, while popular, is typically made in Ecuador |
| Peru | Machu Picchu (via train from Cusco) is a “must see” side trip (allow 3 days)If you are only in Lima, then Huaca Pucllana restaurant at the ruins. | Silver jewelry |