Avoid the Beta Version

by Susan Webber

After more than 450 interviews with executives of 100 software companies, a McKinsey team fails to deliver useful conclusions

Those who have a low tolerance for frustration avoid 1.0 software releases. No matter how useful the new features are, the user experience is inevitably compromised by design flaws and bugs. Secrets of Software Success is the 1.0 version of what could have been a very good book.

Unfortunately, books generally don't have 2.0 versions.

This book certainly aims high. A McKinsey team, headed by lead author and partner Detlev J. Hoch, interviewed over 450 executives at 100 software companies around the world to identify what separated software industry winners from losers.

Despite the exhaustive data-gathering, the team failed to distill the information into tightly-reasoned conclusions. Their work bears the hallmarks of a first progress review, the consulting version of the 1.0 software release, in which consultants present their work to date because they have yet to arrive at an answer.

Another problem is that it is not clear who the client is supposed to be. Although the book is targeted to software executives and other managers of technology, it is best suited to the general reader and programmers who want to gain a managerial perspective.

Even then, it risks insulting the reader's intelligence. For example, the bulk of the first chapter argues the unremarkable fact that "software - nothing but pure knowledge - largely drives and enables today's economy." And the book frequently lapses into a gushing, star-struck tone that grows old quickly.

As a result, Secrets of Success is a compendium of anecdotes and often-too-familiar fact. Too much time is spent recounting key moments in industry history, like the launch of the initial killer app, VisiCalc. It also dwells too long on the care and feeding of developers.

Hearing more tales of how software firms tolerate odd hours, throw elaborate parties, or offer "unique cultures" is not useful. Describing how firms create attractive cultures would be.

The book lacks critical analysis and the authors too often regurgitate what they were told in interviews.

The writers observe that great software companies have great managers, but do not illuminate how they differ from mere mortals.

The qualities they highlight, like "thriving on uncertainty" or "visionary," cannot be assessed objectively. In the mid-1980s, when Microsoft went public, no one would have singled out Bill Gates as an exceptional leader, It had taken Microsoft 10 years to get to $50 million in revenues - hardly a stellar record in a booming industry.

The insights that Secrets of Software Success does provide are not often at a level of detail that can help managers make better decisions. One of the book's better sections is on the importance of marketing, and some of their rules of thumb - for example, that successful firms hire on average two salesmen for every developer - are genuinely useful.

By contrast, the discussion of what the authors call "industry webs" - alliances of software and, occasionally, hardware producers to provide a single customer solution - is long on theory and short on practical advice.

And the notion that tactics often win the day is not new (for starters, see Amar Bhide's 1986 Harvard Business Review article, "Hustle As Strategy," Michael Schrage's work on rapid prototyping, or Evans and Wurster's Blown to Bits), nor does the book tell how to build an organization that executes well.

The book is marred by numerous errors (calling Unix a new operating system in 1987 and NeXT Computer an "Apple spin-off"). The authors also tend to force facts to fit their argument. For example, the book asserts that Apple's fall in market share from 1993 to 1997 resulted from the fact that Apple took 80% of its industry web for itself, while Microsoft took only 10%. This reasoning is quite a stretch.

Arguably, the main reason for Apple's decline was its products had offered premium performance at a premium price. With the introduction of Windows and its Mac-like look and feel, Microsoft closed much of the performance gap. Consumers switched to Microsoft to get a better deal. If sharing revenues with web partners was such a driving factor, how can the authors explain Apple's resurrection?

Similarly, Hock and his team argue that "componentware" is about to revolutionize the software industry. Perhaps it will some day, but its adoption has been much slower than anticipated.

Object-oriented programming has been the wave of the future for the last ten years. It simplifies development once reusable objects have been created. Object-oriented programming requires a different mindset, and many developers find it hard to adapt. It also requires more planning before coding begins and that planning can slow the development timetable, which is often not an acceptable tradeoff.

Similarly, in the chapter on trends, the McKinsey team argues that the software industry will consolidate into mega-firrms and R&D oriented boutiques.

How do they reconcile this view with their advice in the immediately preceding chapter to play nicely with web partners?

Finally, they completely ignored Linux and the open source movement, which is a radically different development paradigm and threatens to erode the power and profits of the major development firms.

Secrets of Software Success offers neither secrets nor pointers on how to succeed. Perhaps the writers are saving them for paying clients.