This is a topic that I find highly valuable – as a tool for interacting with leadership, scoping solutions for small businesses as a consultant, and simply to understand and be able to communicate the value of your IT organization. I propose that while CIOs know the total cost of their IT organization, few of them can represent this as a function of revenue. I would like to direct you to a set of rules that says “you should spend 3-6% of revenue on IT” but such a rule set does not exist and with good reason.
What you should spend on your IT organization will vary depending upon your industry, maturity level, stage of business development, and simply how you generate revenue. For example, an insurance company likely requires a smaller cost allocation (by percentage) toward IT than an online media company. Yet even within an industry it depends on your model – Blue Nile (the largest online jewelry retailer) will invest much more in IT than your local jeweler just as Amazon.com will invest much more into technology than Tattered Cover.
Measuring Overall IT Cost as a Percentage of Revenue
This approach is best for working with leadership (see “Manage IT as a percent of revenue to relate to your CEO” by Michael Sisco on TechRepublic), defining an overall IT strategy, or in making comparisons to other similar companies in order to evaluate your overall efficiency and level of IT commitment. I would start with a top down approach. You must first evaluate IT as a whole and then delve into the division of costs within IT in order to expose specific inefficiencies. For example, a media company might spend as much as 6% of revenue on IT but if 4% of that is on internal infrastructure, 1% on development, and 1% on sustainability then it might mean you have an awfully inefficient infrastructure. On the other hand it might indicate that development at only 1% cost is extremely efficient and it might benefit you to temporarily invest more into development in order to increase your bottom line. If you are too focused on reducing development costs based off of the 6% figure then you will find yourself in constant frustration trying to correct inefficiencies that do not actually exist.
Evaluating Development via Cost-benefit Analysis (a Rationale Based on ROI)
This approach is essential for understanding the sustainability of an application and can be leveraged when proposing a project to the CFO via a simple cost-benefit analysis (CBA). For example: I have a single application with a development estimate of one resource (with an annual salary of $85k) for two months. Ongoing maintenance and support costs are estimated at one resource for three days per month, and $30 per month for the hardware resource (let’s say an Amazon EC2 m1.small instance for this example). This system is replacing an existing system which directly yields $2 million per year in revenue (and the existing system has a cost of $12k per month). As such I would propose the following to the CFO:
- Development Cost = $85k / 12 * 2 = $14,167
- Ongoing Support and Maintenance (Resource) = $85k / 365 * 3 = $699 per month
- Ongoing Hardware Cost = $30 per month
In this case the decision is a no-brainer as the up-front development cost is only roughly greater than the cost of a single month of sustaining the existing system and the ongoing costs ($729 per month) are less than 6.1% of the existing system. These two points will then converge in the first month of deployment yielding significant cost savings annually. This is more clear when you view this in a graph:
In this case, as long as I’m going to continue using this tool for the next three months, it is worth the development effort to build the new tool. It is important to understand this convergence point as well and apply it to the business road map. While this is in reference to a very small project other development efforts might address systems that cost hundreds of thousands of dollars to maintain and millions in up-front (capital) cost. If this convergence point is six years away and the company intends to divest those assets within three years then it might not be a warranted effort – unless of course it will increase the valuation of the asset during a divestiture.
Avoiding the Individual Resource Cost
I chose to weigh in on this just because it seems to be a common question. Leadership requests data and statistics at an individual resource level in an effort to identify individual inefficiencies. This to me is micro-management at its worst and an ineffective effort at best. This will often be disguised as a means of understanding the cost allocation on a project but I would retort that I can simply provide the overall resource cost on a project with a simple calculation (Resource Cost per Hour * Time Spent per Hour). If you must provide this then at least attempt to remove the names and titles as they provide no significant benefit to upper management and only increase the opportunity to further the micro-management.
More importantly large companies (outside of outsourcing agencies and web design firms) do not allocate resources to projects in this way. My team consists of software engineers, database administrators, architects, etc. and they’re allocated across multiple projects depending on the current need, time constraint, and primarily their abilities to: (1) effectively interact with the client on a particular project, and (2) leverage their specific experience as applicable to a specific project task/dependency.
As such I can certainly expose this in grave detail but it might not make a lot of sense to upper management and I cannot identify any rational need to provide this information when I could alternatively provide any of the calculated statistics they are more likely seeking. In short, please just avoid representing cost in this manner at all – it is a waste of time.
What Is Appropriate For You
How you evaluate your IT costs depends on your specific scenario and what it is you are trying to portray but understanding IT cost as a percentage of revenue is most valuable for presenting to clear picture to leadership or negotiating new contracts and defining a scope of work for new clients. The challenge is how to benchmark that against like companies within your same industry or sub-sector. Computer Economics provides compiled reports for a fee but they also have a free executive summary that will provide useful enough statistics for many industries. Other reports such as this “IT Spending 2010” report from Gartner provide an outlook of the breakdown and allocation of the IT budget itself. You’ll have to do your own digging but these will provide a good starting point.
Mark P. McDonald contributed a contrasting perspective in an article for Gartner titled “IT spend as a percent of revenue – a dubious metric at best”. In it he conjectures that “we need to recognize that the metric has no meaning because the numerator does not influence the denominator.” While I agree that it should be not used as a hardline for grading your IT spending it does provide a good baseline for comparison with like companies and a starting point for negotiations regarding realistic budgetary allocations.
- How Does Everybody Else Spend Their IT Dollars? (gartner.com)
- 2012 IT Investments: Leading CIOs Share Their Plans (cioinsight.com)
- CIOs more cautious on IT budgets, plan more outsourcing: survey (zdnet.com)
- AWS Cost Allocation for Customer Bills (aws.typepad.com)
- How to Project Revenue for Startups? (techpluto.com)