How to stop IT from going under: McKinsey's suggestion versus mine

Business and IT executives continue to worry. Based on McKinsey's 2015 survey, their article "Why CIOs should be business-strategy partners" reports that executives’ current perceptions of IT performance are decidedly negative. A few reasons:

  • “widespread concerns over IT effectiveness" beyond providing basic services and managing infrastructure
  • “fading confidence in IT’s ability to support key business activities, such as driving growth.”
  • “little awareness of or agreement on how IT can meaningfully shape a business’s future.”

Note that these are the same "findings" that organizations have complained about for years and years and years.

The McKinsey suggestions


Here are two (of three) suggestions provided in the article:

1. Reimagine CIO role so they help shape business strategy. Some ways to enable the change include "getting the CIO to report to the CEO, establishing clear partnerships between IT and corporate-strategy functions, and holding both business and IT leaders accountable for big business bets."

2. Develop IT’s business savvy. "Before they can contribute to overall business strategy, many CIOs and IT leaders may need to improve their understanding of the business.”

Again, it's important to note that these suggestions have been provided for years and years and years. Just google and you'll find a lot of evidence.

My suggestion


We're grateful for McKinsey's recent survey, but the findings and even the suggestions are not new. For years, other organizations such as Forrester, Gartner, and CIO.com too have reported similar findings and provided similar suggestions. Fixes have been attempted, but the problems keep recurring.

To me this suggests that instead of trying to fix IT governance, we should look for faults in IT practice. Software practice is strong in engineering and project management, but does it have a strong business phase? Does it have a method to translate business strategy into a realizable architecture? Without a translation method, how do we even know whether we're working on the right software projects in the first place?