IT Spending Benchmarks Are Only An Imperfect First Step

I’ve been seeing a mushrooming of requests from Forrester’s CIO clients for IT spending benchmarks. These CIOs attempt to defend their budgets against repeated requests to cut, cut, cut. Their hope is that a spending benchmark will show that their budgets are reasonable given their size and industry – giving them ammo to fend off perceptions that “our IT spending must be too high.”

What I tell these CIOs is that spending benchmarks are only a first step in determining the appropriateness of IT budgets – and not a simple one at that. The reality is that a benchmark tells you only what you are spending in relation to the average of a group of ‘similar’ companies. (see Forrester’s “US IT Spending Benchmarks For 2008”). It really tells you nothing about whether it is the right or wrong amount, whether it’s being spent on the right things, or what benefits you are getting from this investment. And what constitutes a ‘similar’ company is not straight-forward: is it just same size/same industry? Same georgraphy? Same business operating model or strategy? Same prospects for growth or contraction? These are all business characteristics, but what about IT characteristics like the degree of automation, use of packaged versus custom software, outsourcing or history of past M&As?

Benchmark_3 While to a CFO, industry, geography, and company size may look like appropriate criteria for determining peers, when it comes to IT I think that the most important criteria is business strategy as it impacts IT intensity (how important is IT to the business) or IT archetype. For example, a company pursuing a strategy of competitive differentiation through product innovation versus one that focuses on having the lowest prices would have different requirements for IT, so while they may be in the same industry they might not necessarily be peers when it comes to IT spending. This was made powerfully clear to me in a recent consulting engagement where the client’s non-profit status did not correlate with its IT intensity. In other non-profits they compared themselves to, IT was a support function, but in this one it was at the core of their mission.

Forrester helps our clients by assembling benchmark peers that are as similar as we can find while still providing a reasonable number (more than 10) of firms to compare against. We wrap this analysis with an examination of how they are different - adding color to the numbers in a fashion that a CFO can understand. But then we tell them that to make sound decisions about their IT spending, they have to go to a step further and answer more difficult questions about whether they are spending on the right things and getting the expected benefits - questions that only the business, not IT, can answer. CIOs should work with their CFOs to help this analysis by:

  1. Understanding IT costs at the service, not the budget line level
  2. Implementing portfolio analysis to measure alignment
  3. Implementing a benefits realization process to measure actual vs. expected benefits from IT investments

CIOs that can satisfactorily complete step two, i.e. consistently demonstrate tight alignment with the business and deliver expected benefits, will ultimately make the benchmarking step irrelevant.

What pressures are you getting from your CFO to benchmark your spending? What are you telling them?

Comments

re: IT Spending Benchmarks Are Only An Imperfect First Step

I absolutely agree that IT spending benchmarks are nonsense. You can benchmark spending all you want and it will at best reduce IT's ability to service its users. There is only one way to reduce IT spending: "Change!" And the answer here is too: "Yes, we can!"So what is this change I am talking about? It is a technology paradigm shift away from complex solutions. Away from Java, XML, and most of all SOA integration of all the unsatisfactory systems of yesterday. SOA is like server virtualization that has the terrible side effect that all those software pieces that should be dumped can now be kept alive. SOA does the same, just at a substantially higher price.How can software become simpler? Get away from coding applications, customizing standard products and integrating all those old crappy solutions with SOA. I see many of you rolling their eyes. Why do we have those old systems? Tell me! Admit it! Simply because you can't maintain them and you are glad that they are still running ... sort of! You installed and customized CRM, BPM, ECM, BR and BI and now you want to integrate them? Good luck! All existing customized versions of Siebel are all UNMAINTAINABLE. SOA will hardwire all those systems together until the business is sold off and the new CIO is brave enough to dump them.The paradigm shift is to get rid of coding. But that is not an empty suggestion. I am talking about reality. I am talking about our Papyrus Platform that maps a business strategy to a business architecture and deploys it via life-cycle management into production. All content and process is audited and goals are monitored. Anything else? Yes, open that system to the business user. That will take a little longer but as we are not talking about code maintenance it is feasible.Yes, IT is too expensive. We all agree. So how about solving the problem and not just compare if your symptoms are not bigger than those of other businesses. That would be intelligent way to deal with the cost of IT.Max J. Pucher, Chief ArchitectISIS Papyrus Software