Avoid Becoming A One-trick Pony: Embrace Your Business' Need For Real-time

 

The term “one-trick pony” allegedly originated back in the 19th-century days of the traveling circus, where low-end ones were sarcastically called “dog and pony shows.” The really bad ones got the reputation of having a pony that only knew one trick. Today many IT shops are in danger of becoming like those sad circuses, having one or at least a very limited set of technology tricks to help their firms seize opportunities quickly. For example, I routinely talk to business people who say they avoid IT at all costs when they have new analytic needs; at these firms, IT has only one response to all new requests for data – update the data warehouse or a data mart in a slow and expensive waterfall development process.

One term keeps occurring, as I talk to businesses about this issue — they want to be real-time.  We’ve been using the term for years to talk about a wide range of things, from embedded C to extreme, low-latency analytics. I think all of these miss what the business is really after — the ability to use more information more quickly to take rapid action in response to unanticipated changes in their environment. Five-year technology strategies are out; but many can’t get their head around this new world, which is why a recent Forrester study showed that IT is increasingly losing control of technology spend. How do we get back in the game?

Companies like Barclays Wealth Management, Sears, and USAA are redefining their architecture with new tricks to be responsive in real-time by:

  • Increasing the number, granularity, and scale of predictive models. Investment firms have been doing this for years to predict the behavior of securities, but now we see those same concepts being applied to customers. What if you had a model for each individual customer? How responsive could your firm be? How profitable?
  • Shifting their data architecture from layer cake to hub-and-spoke. Leading firms are replacing their 2005 ETL and data warehouse centered “layer cake” architecture with a hyper-flexible hub-and-spoke platform that provides multiple cost/performance options for making more data available to their business more quickly.
  • Closing the gap between detecting a change and acting on it. At its heart, this is the essence of tomorrow’s real-time business. The CTO of Sears put it well: “This is a whole new ball game.” In this new game we need to be more responsive our changes in customer behavior and markets.

Emerging technologies like cloud, advanced analytics, and mobile have captured our imagination because they deliver real-time capability with agility and new cost options, but there is more coming very soon. To cope with this change, we need to embrace emerging technology that does away with entrenched limitations and let go of tired mantras like the “single source of truth.” In a real-time world, there is way too much information and potential value.

Come see see my keynote speech on this topic at our EA forum event in Washington, D.C., on May 6 and 7, then dialogue with senior leaders from Barclays, Sears, and USAA in a panel afterward. In the meantime, let me know if you are seeing this trend, what you think about it, and what your firm is doing to cope.

 

Comments

Real-time Network Enterprise

Hi Brian,

I couldn't agree with you more. The Enterprise seems to have a growing gap in perception / requirements. Entrenched interests in status quo seem to overlook the fact all the cool innovation is happening outside of them.

I've long argued that the "horses for courses" argument results in a stable of expensive one-trick ponies - specialization, verticalization, increased network complexity, latency and expense.... (application middleware vendors dream, customers nightmare).

If we hope to realize the benefits of the Cloud, we need a Network Application Model - the ability to link, evaluate and respond to diverse, distributed application resources on demand.

Stateful middleware stacks sub-optimize Cloud infrastructure. Until we change the application model the enterprise won't get what it wants - smart, connected, adaptable applications.

Best,
Dave
http://www.enterpriseweb.com

Centralized vs. distributed

I gave a speech on this at our forum, and the point was made that real time needs distributed architectures. Had got me thinking. Nice comment, thanks.