IT Dilemmas [John Levy]

As an organization grows, it has to increase the scale of IT activities, including software development. As a result, it faces these dilemmas:

1. Software development is engineering, but IT is an expense.

If your only management metric is level of expense, rather than return on investment, then you won’t be able to differentiate between winners and losers in software development. Some organizations have the CIO reporting to the CFO, because they are completely oriented to control of expenses, and IT is a large expense. But this misses the fact that investment in software, whether for internal use or for sale as a product, should be managed the way that R&D (Engineering) is managed – as an investment in development that must ultimately produce a return.

Engineering is not purely execution. R&D/Engineering involves a degree of creativity along with a professional sense of discipline about implementation. That’s why recruiting and retaining people who can balance innovation with execution is as important in Engineering as in, say, Marketing.

2. Outsourcing and offshoring work when the activities of software development can be broken into component parts and split across multiple organizations without loss of innovation or relevance of the product.

In forward-looking organizations, IT-implemented tools provide a strategic advantage to the business. To achieve this, IT strategy and architecture must be very close to the top decision makers. In this context, the question of outsourcing becomes, “can we succeed by having people we have not hired directly execute some part of our implementation?” To the extent that the impact of software or IT on the organization’s strategic goals is not considered, the decision to outsource can lead to failure in many ways.

3. If no one in top management understands how IT works, then all the IT decisions will be made based on financial reports. As organizations get big, they become more and more finance oriented. When someone at the VP level or higher has enough technical background to understand IT implementation, then IT strategy has a much better chance of being realistic. Of course, the CIO should have a basic knowledge of technology as well, because vendors will attempt to sell things which may not be appropriate to the organization. The main problem, however, goes back to item 1 above: managing Engineering is different from simply controlling expenses.

4. Traditional Project Management doesn’t fit well with Agile software development.

When we do software development in an Agile framework, we have fixed time & budget and variable functionality. Almost no one who has been doing traditional Project Management intuitively understands how to manage a project this way. Short Agile iteration cycles confound traditional PM tools and measurements. We need the PM world to come up with a new division: Agile Project Management, with a focus on projects that are not characterized by large sets of requirements up front, and have rapid fixed cycles with variable functionality.

5. It is difficult to overcome an addiction to short-term metrics.

IT managers prefer short-term pain-reduction (ship it!) over longer-term goals (is it what the users want?) every time. This leads to the never-ending cycle of code debt (code that needs to be cleaned up and/or re-written). To overcome this phenomenon we need to understand the psychology of addiction and to get top management to support a change to proper metrics. For further discussion of this, I recommend reading Gerry Weinberg’s Quality Software Management, Volume 3: Congruent Action (New York: Dorset House, 1994).

All 5 of these items also apply to software development in Engineering departments. But the effects are most widespread in IT organizations. How is your IT organization doing?

__________________________________________
John Levy consults on managing Agile development and is a frequent expert witness in computer & software patent cases. He has 30 years’ experience as a consultant and manager at Quantum, Apple, Tandem and DEC. His book on managing development, Get Out of the Way, is due out in 2009. More info is at http://johnlevyconsulting.com/