Five reasons that Big Data initiatives go wrong

Gartner’s 2012 hype cycle is not optimistic on Big Data*. The technology is shown as near the Gartner emerging technologies hypecycle 2012“peak of inflated expectations” just about to go over the cliff into the “trough of disillusionment”. The bubble seems ready to pop when the skepticism is so commonplace and the warnings so dire. Even so… this is as yet a young technology and it is difficult to predict its trajectory. I will however list the five major reasons that analytical initiatives fail. Take a look at the list anddo your own readiness assessment.

Top 5 reasons why analytical initiatives fail

Reasons #5: The data quality is suspect : The question to ask is if the data measurements reasonably represent the measured real-word phenomena. In one case study we knew the applet on the installed unit was malfunctioning in some of the installations. The timestamps were junk. That said, roughly 60% of the readings were still usable and a quantum leap in the business’ knowledge of their customer behavioral patterns. The early wins realized from the analysis in that phase helped sustain momentum of the initiative. The next generation of installations had much better data quality.  If the representation is reasonable, the next question is if these data are relevant and/or adequate to the targeted business (see also Reason #1 below).

Reason #4: You lack the technology : So you have a data warehouse but all it does is store and manage data. You lack the pipes to bring the data into your existing processes. Your hardware vendor of course will be glad to take another seven figures and another eighteen months to implement. Till then you twiddle the thumbs. My biased opinion is to try out cloud analytics with a partner like us. You get a proof of concept in a couple of months and if you see the benefits, you continue with a public or private cloud implementation. It’s the shortest path to benefits realization with zero! IT overhead.

Reason #3: You lack the resources to work with the data : The other big reason analytical initiatives fail. Your IT team has fantastic operational skills or your analytical specialist is heads-down driving those killer predictive models. Everyone has their hands full making sure deadlines are met. And you dont have the budget to hire a new team for a project with indeterminate value-add. Integrating this data into the processes is a full-time job and it cannot be done by part-time resources.

Reason #2: It is difficult to use the findings from the data in day-to-day work : This one is Tire swing graphic poor product developmentbig. At the end of the day,  analytical insights have to be usable by the end-user. This could be the call center operator, or the marketing manager, or the product manager, or the regional sales manager, etc. If the initiative is carried out without involvement from the end-user THE INITIATIVE WILL FAIL. You got to build with the end in sight and the end-user needs to be involved throughout the design and delivery. The famous tree swing graphic literally paints the picture [click on image to see attribution].

Reason #1: there are organizational barriers to making use of the data : It comes down to leadership, or the lack thereof to be more precise. This is without fail the single biggest reason that analytical initiatives fail.  If the initiative is considered a technical initiative led by the technology team, chances are this will be a cash sink. A line of business executive who runs a revenue center and who has line-of-sight visibility on the use of the analytical insights is the best champion. Bonus: If you have such a champion Reasons #2-#4 will evaporate.

Are you ready for Big Data?

Test your analytical maturity below. If you score at a Level 4 or higher, you are ready to use Big Data. If you are Level 3 the odds are 50-50 that you can get some benefit from Big Data. Level 2 or lower, forget Big Data. You need to get the fundamentals right first.

Analytical maturity calculator

 

Definition

I have heard different versions but most prefer the definition from IDC – Big Data technologies encompass hardware and software that integrate, organize, manage, analyze and present data characterized by the “Four Vs”

  • Volume
  • Variety
  • Velocity
  • Value

In our experience, the best “Big Data” implementations (especially in respect to Value delivered) have been in the area of consumer telematics – streaming usage data from consumers delivered to the company for use in product design, marketing and pricing decisions.

To see how industry leaders are leveraging consumer telematics Big Data download the case study below.

Case study shows how consumer telematics augments market research

 

Leave a Reply

Your email address will not be published. Required fields are marked *