A great Friday diversion. If you company is pondering how or why to use social data to enhance analytics understanding its impact is critical. This video does a nice job driving home the key points of how pervasive social is in our lives.
Analytic Response - Blog
Power users and self service line of business (LOB) users are always asking for more data access and easier to use tools that enable them to bypass traditional IT in favor of analytic freedom. Most of these users will push back hard on an environment that restricts thier ability to wander about in enterprise data.
Providing power and freedom to these users has incridible upside.
- Faster time to insight
- Less work for IT
- Aligning business knowledge with the analytic process
- pluggin more team members into data driven insights
These are just a few of the positives that come from self service or discovery based business intelligence. So what are the pitfalls of all of this freedom? I see two major disconnects that need to be addressed as this type of BI beomed more pervasive.
- Governance - Adding desktop data is a common choice for self service users. They often mash up information from a wide variety of sources many of which are not avialble to the rest of the enterprise. This can enhance or detract from the overall analysys and leave many in the dark as to how certain decisions are being made.
- Enterprise Value - The insights created by LOB and power users can be extremelly valuable to the rest of a team or the company in general. Self service solutions that don't supply a common and easy path to share and leverage these insights is missing a key feature. I am partial to solutions that balance freedom and control.
So the question is where does the control belong and how tight should it be. Does some governance and control add to the value? I think it does but the challenge is finding the sweet spot between freedom and a walled garden. If you want adoption you can't fence the power users in.
For your entertainment - Bing Crosby and the Andrew Sisters sing Don't Fence Me In.
Mashable has published coverage of the world wide patent race underway at leading technology companies. IBM maintains its first place position.
These Big Data Laws are written for the entertainment and perhaps education of vendors who are in this market and are briefing the analyst community about their solution. I Hope it helps and makes you laugh just a little bit.
Law #1 - The heavier the marketing message the lighter the technology.
Companies that lead with buzz words and marketing blather generally are struggling to deliver on the technology front. Decide early if you want to sell a marketing message or an innovative technology that will solve enterprise challenges. Sell the value not the message and stay away from marketing slogans like - Hadoop is Free!! (My POV - Its free like a puppy.)
Law #2 - The proper answer to the question "How many customers do you have?" is a numeric value.
The improper answer is anything that doesn't start with a number. The worst answer is a long convoluted narritive on how you are serving the needs of many industry segments while focusing on premier client penetration thru value added partner channels within high opportunity niche markets...blah blah blah. If you can't or won't provide a number I already know its less than 10 and I'm nervous it might be zero. I can't recommend you to my clients if I think they might become an experiment.
Law #3 - You are not the first, you are not the only and yes...you do have competitors.
Statements like these make analysts crazy and we come away thinking that you don't really understand the competitive terrain or the market in general. Steer clear of these types of declarations and focus on how you provide value and solve real business problems. (See Law #1 for clarification)
Law #4 - Analysts already know Big Data is really, really, really big and so do your prospects.
The size of today's data is old news. I already know what a Petabyte, Zettabyte and a Yottabyte are. I know about machine data, dark data, The Internet of Things, social data and sensor data. Big Data is about opportunities, being able to do workloads we could only dream of doing years ago at a speed and economic level that now makes it practicle. Educate us on what your company does and how you do it, lets skip the part where you explain how the world produces more data daily than the contents of the Library of Congress.
Law #5 - A connector to Hive is not a comprehensive Big Data strategy.
Hive is an interesting access point to Hadoop data, the ability to pass SQL into Hive opens the door for some interesting functionality but its not a comprehensive Big Data feature set. Hive is the low hanging fruit of Hadoop interaction and was the starting point for many vendors who needed/wanted to add a Big Data marketing message to their go-to-market strategy. (See Law #1 for clarification)
Stay tuned for more Laws and updates to this post -
Adoption of the Hybrid Data Ecosystem continues to grow. Vendors are working to deliver highly integrated ecosystems with platforms that provide the user an agile and flexible array of solutions to address today's complex and demanding workloads. The interesting part of the story is the division in how they approach this opportunity. Some vendors are building fully featured "Walled Garden" style solutions with all parts dependent upon one another. Its a nice strategy that may lead to better inetegration but you can't really get away from it once you are engaged. This type of lock-in to technology and infrastruture can be dangerous and in the long run very expensive for the consumer.
Others (only a few) are focusing on highly integrated environments that match the competition feature for feature but at the same time allow for portability. Pivotal is a company working to bring its clients portability with cloud infrastructure. A common fear of cloud adaptors is being locked into one infrastructure provider for the duration of their projects or forever. If you utilize Pivotal on its Cloud Foundry platform you have the ability to move from AWS to Rackspace if need be providing a level of flexibility that most companies would prefer over time. This strategy is also smart for companies that suspect a Cloud based program may migrate back behind the firewall at some point. Being locked-in to a Cloud provider will make this an impossible change or at the least terribly expensive.
So this begs the question, what's better? Walled gardens or portable infrastructure. The answer seems obvious to me but I'm interested to hear from you on this topic.