Running a bit behind on getting blog posts out this week, but I thought I’d try to get one out early about the MIT CIO symposium I attended yesterday.
Jeff Cutler did a great job of summarizing what went on at the Big Data session and wrote it up in a blog post, with pictures and everything.
A couple of the key elements discussed were:
- Correlation is NOT causality.
- When dealing with Big Data, you need to Measure, Experiment, Analyze and Replicate. Having expectations is important.
I did ask the panel a question about their view on:
When you expand your knowledge about the organizational behavior and management process using big data techniques, management may be one of the most well understood and best targets for automation so what are the implications for business and business schools. I don’t think they really understood the question, since their answer was targeted at a whole other set of issues related to how management uses the output of Big Data efforts.
I did talk to Erik (the moderator) after the session and he agreed that this is an area where organizations have significant opportunity in the future. When you think about business processes and the data available, there is structured and yet-to-be-structured data as well as well understood, yet-to-be-understood processes and chaotic behavior (I almost said chaotic process, but if it is chaotic it can’t be a process). Most of that management work is ripe for automation, freeing up people to work on other creative (leadership) tasks.
Lately, I’ve been a number of conversations with people about the strategic use of technologies. I mentioned the criteria I use to evaluate trends and technologies. We then typically get into a discussion about the difference in impact between some of the technologies that are much discussed today and how the tactical use differs from the strategic use.
- Analytics – Although you may need to gather more data and keep it longer, there is not enough attention space to sustain the effort unless you simplify, automate and focus attention only on what needs human involvement. Time to action/decisions has to be the measure of impact.
- Cloud – Although it may reduce costs in certain circumstances, the strategic impact of cloud techniques (whether it is infrastructure, processes or people) is to increase flexibility. If through the use of cloud techniques you end up increasing the flexibility, it cannot be sustained.
- Mobility – The mobility strategy for a business has to focus on improving the access to corporate information and reducing the latency in the decision-making process. If the focus remains on the devices, it will also fail.
These current technology directions (and others) have a strategic side and a tactical manifestation – make sure you know what is important to your business over the long haul when creating your plan of attack. If you want to reach the top you still go up one step at a time, but it is easy to lose sight of the goal along the way. Identify the metrics to measure progress and then measure the impact along the way and make adjustments.
When I was writing this post I felt it was a bit risky, since these technologies are viewed as so important today. The real point of the post is to view them strategically and not just a buzzword or fad. This tactical approach may be the reason that for some organizations, innovation is not working out.
Yesterday, the New York Times had an article titled: Pills Tracked From Doctor to Patient to Aid Drug Marketing. The articles discussed how the new analytics capabilities are allowing drug marketers to locate influential doctors by their social behavior as well as patient behavior. This article was a good example of where additional insight can be used to define action.
I normally view much of the “Big Data” trend to be focused too much on insight and not enough on action, but this article did talk through some of the interesting issues at least as it relates to the healthcare provider market.
Recently, I’ve been working with our executive briefing folks and a number of others on megatrends – the industry independent trends that will shape our lives in the future and their effect on business decision making. These will naturally shape how technology is consumed as well.
In the process, a number of meta-drivers fell out that may shape the megatrends. Yes, this is turning into a convoluted network of interactions and that is why some models to assess these interactions are so important. These categories for these meta-drivers seemed to be:
- Engagement – this is what drives social, concepts like flow and maybe even the Internet of Things
- Simplicity – addressing the limitations of our ability to consume
- Efficiency – this embraces the concept of abundance and scarcity
- Flexibility – the need to adjust quickly (probably the sustained driving factor for cloud techniques)
- Security – we all know about this, if you don’t feel safe almost nothing else matters
- Visibility – the need for contextual understanding in order to act (one of the reasons for the current focus on Big Data)
Are these too simple? What have I left out?? It surprised me how old some of the links I identified were to link to this post.
It seems like many of our decisions could use an indicator showing how they increase or decrease these categories. We could use this as part of defining our expectations.
How many times do we make decisions that increase security but radically decrease flexibility or visibility, for example? You hear that discussion about our personal as well as our business lives today.
This week I had the opportunity to attend one of Leon Kappelman’s classes at the University of North Texas to participate in interactions with students about their senior project/presentation. The teams of students were covering a number of topics like BYOD, Cloud adoption, Biometric based security… All topics where I felt fairly comfortable.
One presentation was focused on Data Management in the Age of Big Data and they had one concept well understood that many analysts miss.
The opportunity for better decision making.
The team focused on 5 key issues. The lack of:
- Data Governance
- Data Quality standards and management
- Data Architecture and Security
- Operations support
- Business buy-in
We had quite a discussion about the business buy-in issue, since we needed them to explain why it would get this far without buy-in but they explained that the issue orbited around business culture and the implications advanced analytic techniques would have on the culture.
I was happy to see these students internalized these concepts, and hope the organizations they move into after graduation are ready for their perspective.