John Parkinson: Budget Blacks and Blues
In the final quarter of every calendar year I send out a survey to my private research panel asking them about their (or their company's) technology spending plans for the coming year. The respondentsa 30-year collection of just over 600 friends, acquaintances, ex-colleagues, clients and other contacts in the technology world who are willing to respond to up to four surveys a yeararen't selected statistically. But they are interestingly representative of the technology landscape in North America. Generally, I get around 400 responses, and this last survey was no exception: I got 422. I've been looking hard at their answers for about a month now (I'm writing this just after the 2003 holidays), and I saw some patterns that were intriguing enough that I decided to make this the subject of my first column for 2004.
The most obvious finding: Almost no one expects to be spending more on technology in 2004 than they did in 2003. By playing with the analyses a bit, I'd estimate that the most optimistic overall prediction for a budgetary increase is plus 2 percent, the median is flat and the worst prediction is down about 6 percent. Just as interesting, however, is the fact that almost everyone also expects to be starting up major new projects in 2004 after a couple of years of deep drought. Respondents stated two main reasons for this: They still have too much old infrastructure that's at or close to the end of its useful life; and they have been able to save enough operational costs after three years of belt-tightening that they now have some investment headroom, even after fairly significant budget give-backs to their corporate parents.
But what exactly are they going to be spending their uncommitted dollars on? Here the picture (as always) gets less clear.
Quite a lot of budget is going to go into a rolling platform refresh aimed at retiring older technologies and replacing them with new onesas long as the new technologies both save money on an absolute basis and are relatively more productive than what they will replace (note that's and not or). By continuing with moderate to aggressive simplification, standardization and consolidation strategies, including renegotiating service-level agreements with service providers and better quality-of-service measurement, my respondents think they can squeeze between 12 percent and 20 percent annually out of their operational costs in 2004 and 2005 while improving reliability and availability. Lots of interest in resource virtualization strategies and operational automation.
Quite a lot of budgetand quite a lot of those operational cost savingsare going into rationalizing the application suite, updating it where possible, encapsulating old functionality where updates are not viable and integrating it (especially the data, but access, UI and security are all in play) everywhere. There are lots of plans for corporate portals of various kinds: business analytics, data rationalization and lightweight enterprise- application integration via Web services. A fair number are planning to retire old custom-developed back-office systems and replace them with ERP packages.
Almost all of this, however interesting, is essentially internal to the world of IT. It's about better plumbing and better housekeeping, about renovation and rationalization of sourcing. What I really wanted to know was what totally new, business-driven projects were in their plans. Here the scatter-gram is all over the place. After a month of playing with the responses via a variety of statistical models, I've come to the following conclusions.
Population membership isn't strongly correlated with industry, company size or age of IT executive. It isn't even strongly correlated with corporate performance, although I'll admit that I don't have enough data to be sure on some of the marginal correlations.
Population One are the habitual early adopters (and habitual optimists, I believe). In general, this population thinks of technology as a way to build competitive differentiation, as long as you keep adapting. They are pushing ahead with RFID, agent technologies, reworked services architectures for their core business systems, interprocess and intercorporate links, extended mobilityall the hot, and some of the hyped, topics from 2003. Most of them have technology budget increases, some fairly large. They represent about 15 percent of my respondentsalthough we have other survey data that indicates this population corresponds to only about 5 percent of all large- and medium-size companies.
Population Two are the pain avoiders (and probably habitual pessimists). They see information technology as a cost to be managed (if you have to) or avoided (if you can). They're mostly looking at stretching the life of their current assets, which are already older than average, replacing proprietary technology with "free" or open-source products and outsourcing their infrastructure, application maintenance and even new development to lowest-cost providers. A lot of their IT work is probably going to go offshore. Population Two makes up about 25 percent of my samplesmaller than the 35 percent of the total population of large- to medium-size corporations that we believe are pain avoiders.
Population Three is somewhere between these two extremes. They see information technology as a source of operational rather than strategic advantage, and they are generally incrementalists in their approach to planning for 2004 and beyond. There might be a couple of subpopulations buried in hereI can't really tellwith "preferred rate of change" being the discriminator. Their strategies seem to be focused on doing what the business side says it wants while building better links to their internal customers so that they can understand what the business actually needs. They are 60 percent of my sample and about the same proportion of the total corporate population. Their plans are about as scattered as the total sample, though with fewer outliers.
All three populations have some things in common. They are all using shorter ROI horizons than they did a few years ago. For Population One, it's generally one to three years (down from five or more in 2000). For Population Two, it's generally less than a year, although the reality looks as if this is a way to stop most projects rather than a belief in the possibility of quick returns. Population Three seems to be in the range of 9 to 18 monthswith real expectations around 12 to 18 months.
A statistically significant number of companies60 percentreported that they have given, or are planning to give, the IT organization to a business executive to run. They are expecting this to result in IT running more like a "real business function" (whatever that is) and less like (and I quote) a "money pit."
I'm thinking that this all adds up to a very different picture of enterprise IT over the next couple of yearsand a less growth-oriented picture than that painted by a lot of industry watchers. We'll continue to track this throughout 2004 and report back on it occasionally. Until we know different then, here's to a year of healthy budgets!
John Parkinson is chief technologist for the Americas at Cap Gemini Ernst & Young.
Illustration by John Kascht