Focus on some basic elements of physics -- position, direction and movement -- to engineer the best performance from your IS department Try applying the six dimensions of measurement for each organisational unit in your IS department.
The statistics are sobering: only 18 cents of every dollar spent on IT in 1998 went toward value-producing spending. To help overcome this IT inflation, companies must manage their IT investments far better than they have in the past. IT investments ought to maximise value and benefits and be adaptable to continuous change. Today's IS organisation must design its process and cost structure to allow it to pick a performance target, hit it reliably and optimise the cost of doing so. In addition, the buzzword for managing IT today is agility. Agility, in my view, means the ability to turn on a dime, change rapidly and change the rate of change itself. With agility a high priority, where you are headed and how fast you are going is more important than where you are. It is also critical to know how fast your competitors are moving and where they are heading. Unfortunately, the measures and measurement reporting structures in place at most companies today make it difficult to manage with agility. To illustrate my point, forget about IT and IT metrics for a moment.
Imagine the focus is weight loss. Let's say you are faced with losing five kilos in 90 days. (Notice the target and time frame.) Let's look at two different scenarios.
Scenario 1: On the Way to Success
You weigh yourself every day (the current state measure). After one month, you've lost a total of two and a half kilos and are maintaining a program of losing about one to one and a half kilos every two weeks. It looks like you'll reach your goal. How do you know? By watching your current state, keeping your eye on the target and looking at the change rate, everything is proceeding according to plan. In fact, at the current rate, you will lose the target five kilos four to five weeks after the first month. The big assumption, of course, is that you'll be able to maintain the current rate of change. If that rate is higher, you'll close on your target more quickly. If it's lower, you'll close on the target more slowly. The key is always looking at the current state, the change rate, the target and the projected time to close on the target.
Scenario 2: On the Way to Failure
You weigh yourself every day. You look at your current state and the target but are not concerned with the rate of weight loss. After 60 days your weight hasn't changed. However, you assume that you can lose the five kilos in 30 days, regardless of past history. Now 88 days have passed, and your weight is still the same. Can you lose five kilos in two days? Probably not, unless you take radical steps. Another day passes, and you've lost half a kilo. You are faced with losing four and a half kilos in 24 hours. The situation is hopeless. The point of the two scenarios should be clear. Current-state measures tell only a small part of the story. Think of measurement as a tool for learning, navigation and managing your organisation's rate of change.
You might wonder how measurement can help an organisation attain its goals. It all depends on the ways it is applied. In my view, measurement is most useful in the following three applications: A. To provide a moving snapshot of key performance indicators - Current status - Position relative to targets - Movement toward targets B. To provide alerts to identify critical areas that affect your ability to reach your goals. The alerts will be derived from - Focusing on leading indicators (predictors of future performance, such as expected ROI), not lagging indicators (historical records of past performance, such as last year's market share) - Identifying expected ranges of acceptable performance for key indicators C. To communicate quantitative aspects of organisational goals across the enterprise in terms of targets and performance improvements Think back to the weight loss examples. The key in those examples is knowing the rate at which you are closing in on the target. However, even this is not enough if used alone. You need to know whether the current rate is sufficient to hit the target and whether the rate is being sustained. Then you'll be able to decide whether the rate needs to be altered or not. For example, should it be accelerated? If so, is that acceleration possible and can it be sustained? Key Dimensions To help answer these questions, measurement and measurement reporting should include five dimensions to be useful. 1. Current state. Shows current value of measure related to indicator 2. Target state. Shows desired value of indicator 3. Movement. Change in your performance since last period 4. Target achievement. Predicted amount of time to achieve target 5. External benchmark.
Shows industry standard value (if one is available) Notice that I've added the concept of an external benchmark to the list. Organisations need to know if the targets they are setting for themselves are within the frontiers of existing performance or in fact are stretching the limits of known performance. Using an external benchmark will help answer those questions. My list of five dimensions of measurement is an oversimplification of the situation from several perspectives. First, measurement is part of a process. The dimensions that I have suggested must be tied into a reporting mechanism. An organisation must identify its key performance areas and develop measures to report on them.
The measures must be reported frequently enough so that corrective action can be taken if necessary. One of the dimensions cannot be obtained through passive data collection. It is perhaps the most pivotal piece of all-the predicted time to achieve the target.
This value is the one that the person or team responsible for achieving the target must provide and watch with the most attention. By closely tracking predicted time, you will know whether you need to "ramp up" your rate. But the most important question is, What should you do if you are slipping a little, or worse, a lot? The answer can be found in a sixth and final dimension-the escalation indicator. 6. RAG (red, amber, green) escalation indicator. Shows whether the target achievement time is in predicted/ desired operating zone, the level of escalation and whether intervention is required. This is a leading indicator, and management should watch it closely. It lights up red, amber or green depending on the level of escalation (intervention) needed to deal with the slippage. By watching the RAG indicator, you will know if you are heading toward trouble -- perhaps the rate is too slow or you're heading in the wrong direction. Action can then be taken to change the rate, target or focus of the project.
Using the Six Dimensions
To better understand how the measurement reporting structure described might work, let's look at a hypothetical year 2000 project, the primary focus of which is getting core systems converted, tested and operational in time.
Suppose a company has 10,000 programs to convert and become functional in a 20-week period. Reporting should focus on: 1. Current state. 10,000 programs to convert, test and install 2. Target state. 0 at the end of 20 weeks (an average decrease of 500 per week) 3. Movement. Number of programs converted by the end of each week (the conversion rate) 4. Target achievement. Predicted amount of time to achieve target given the actual average conversion rate 5.
External benchmark. Maximum attainable rate per week based on industry experience 6. RAG indicator. Tied to whether the movement indicator showed at least an average of 500 conversions a week You should try this reporting structure for each organisational unit in your IS department. It will change the focus of your management discussions and, more important, change the focus of the people on the front lines. Remember, position, direction and velocity are the key elements of measurement and measurement reporting. By focusing on these dimensions, executives will become performance engineers, striving for agility in a fluid world.
Today's IT dollar supports fewer value-producing activities than in the pastIn a world of inflationary pressures, the money spent on information technology is no less immune than money spent on college tuition. The IT dollar has become diluted. The numbers from the Rubin Systems/Meta Group's annual Worldwide Benchmark Study tell it all. Between 1991 and 1995, for each dollar the average US company spent on IT, 70 cents went toward value-producing activities while 30 cents went down the tubes for the cost of failed projects, failed systems and the reworking of existing infrastructure. Consider the new wave of forces that have been acting on IT since 1995: year 2000 problems, the IT labour shortage and the increased complexity and change rate of technology. How have these forces led to IT-dollar inflation? In 1998, the year 2000 problem takes away about 25 cents of the value-producing spending of the average company, leaving 45 cents; the complexity and change rate of new IT cuts the value of the 45 cents in half, to 23 cents; and the IT labour shortage causes labour costs to go up by 20 per cent, leaving 18 cents. In short, each dollar spent on IT today has less value-producing potential than in the past.
The impact of these forces is starting to show up on the top line: company revenues. From 1991 to 1995, the revenues supported per IT dollar spent slowly crept up until, by 1995, each IT dollar spent by the average US company supported $US48 in revenues. As the dilution factors kicked in, this value has tumbled in the past three years to $US38.
Although IT spending in US companies has increased from an average of 2.3 per cent of gross revenue in 1996 to 2.6 per cent in 1997, this 13 per cent increase has not been enough to offset the declining value of each IT dollar.
In fact, demand by businesses for new systems continues to grow at 30 per cent per year. This means that demand is growing 2.5 times faster than IT budgets (without considering dilution of the IT dollar) and presents most companies with an impossible goal: how to meet their demands for new systems with insufficient IT budgets.
Howard A. Rubin chairs the computer science department at Hunter College in New York City and is a research fellow at Meta Group in the US
Join the CIO Australia group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.