Analytics, automation and anxiety

Enterprise Management Associates has been researching automation and analytics over the past four years. We have also been actively researching organizational and at times even psychological issues regarding the adoption of new technologies. So while preparing for a Webinar on automation, it occurred to me to look at the linkages among all three: analytics, automation and – to keep the alliteration going – anxiety.

I think I am stating the obvious when I suggest that there is a very strong link between analytics and automation. Some automation (probably still the bulk of automated capabilities today) involves scripting, where a single technical “guru” codifies specific areas for automation based on his or her domain expertise. But increasingly, automation will depend on powerful analytic engines to pinpoint and focus just how actions of various kinds should occur.

It’s important to state that automation can occur in many contexts. Not every area of automation requires an action taken on the infrastructure. In many cases, such as automated troubleshooting, automation can support shortened decision cycles and superior decision-making by collapsing false cues and myriad events and deriving a single recommended cause for action. Automation can also shorten and ease deployment cycles and minimize administrative overhead. And finally, automation can enable actions ranging from simple realignment of thresholds and trouble-ticket generation, to policy-based software distribution, to rerouting traffic around a point of failure, to fully self-healing infrastructures.

Advanced analytics in multiple forms, through multiple technologies, can help automation succeed. Technologies ranging from rules-based state machines, to neural-networking engines, to fuzzy logic, to case-based reasoning software, to data mining, to object-based technology and even fractals – all can help to support a more automated managed universe. Analytic tools support topology and root-cause analysis, performance management, configuration and change management, accounting and billing, and business impact analysis. Without these analytic tools, automation will remain at best a deviation from the norm – with deployments more like duct tape than deliberate strategic investments.

Now let’s look at the final “A,” anxiety. This is the mirror image of automated tools. Even though it’s a good thing – and in many ways a necessary thing for progress, automation can be seen as threatening to individuals and organizations in multiple ways. For many professionals, automation should be liberating and empowering, but those who view their job identities as defined by repetitive tasks may feel otherwise.

Similarly, while organizations seeking superior levels of efficiency and business alignment will find automation essential, those organizations with cultures strongly resistant to change and to new ways of doing things may find automation hard to swallow. These reluctant organizations and the individuals within them should be encouraged to look ahead – to see how to integrate automation into process change.

For IT organizations seeking guidance in this, there are many sources. EMA is planning to host an assessment tool related to IT evolution on its Web site before the end of September. There are of course other sources, such as the IT Infrastructure Library, that can provide detailed roadmaps for process enhancement.

So to make a very complex situation simple: Do look to invest in automation; don’t run from it. Look for the right type of analytics and technologies to support your requirements. And take the time to think through process and organization to plan and communicate value both internally and externally.