Measurement for SAM – Practical Techniques to Optimize Software Efficiency

By Phil Hames

People talk about lean processes and efficiency with regard to IT but in reality there is confusion about what this actually means. There tends to be a focus on effectiveness or “get the job done” for most IT professionals – their prime concern is keeping everything running. If systems go down everyone shouts, so, of course they want to ensure that everything is working rather than saving every last penny. However, efficiency need not be achieved at the cost of effectiveness and often the two aspects are complementary.

Software licensing is a bit like the tip of the iceberg: you see the cost of the licenses but it is harder to quantify the hidden costs such as the costs of support staff, security, hosting, hardware and software updates, staff costs, procurement, etc. It is even worse for the organization when you purchase software assets and then not actually use them. If you can optimize your software usage then you will also give cost efficiencies throughout your IT.

Make IT Simple

Many financial departments have the view that IT is expensive and money may be being wasted. To other people IT is invisible, making it hard to determine cost efficiency. In general, managing software assets is not simple when they are treated as invisible. Senior managers can be confused and frustrated trying to figure out whether they are investing wisely in their IT assets. IT managers will improve their managerial credibility if they can show that they are maximizing the efficient use of their software assets.

Many IT managers will say “My area is already efficient; we are meeting all our KPIs and keeping to budget.” This really means that the IT area is effective, but not necessarily efficient. Two people can be effective in getting a job done but they can have different efficiency levels. For example, the two people could travel from A to B and both get there and are effective in reaching the goal. However, one takes roads which are less direct, using more fuel and taking longer than the other. Therefore, one is less efficient than the other. In summary, effectiveness is a measurement of whether you are doing a task whereas efficiency measures how well you are doing the task.

Our goal as managers should be to achieve effectiveness in the most efficient way, in other words: know where you are and where you want to get to.

With any task, you have to set goals. These goals should be easy to understand and measureable. You also need to measure where you are starting from. The difference between these two measures is your progress towards improving efficiency over time.

For example, you are trying to become more efficient in the usage of a software asset and begin with a measurement of 64% usage in November. By the following May, usage is at 85% for the same software product and you have achieved an increase in efficiency over a six month period.

1. Start with the Money

If you are going to prioritize your drive for efficiency, then step one is to go where the money is, i.e., get the biggest savings first if you can. This is a basic prioritization strategy that can be done using a common sense management technique called Pareto analysis.

To use Pareto analysis on software licensing efficiency, you begin by identifying the variance between what you use on a regular basis and what you have licensed (see Figure 1). You may find that in some cases, you are using more software than you have licensed. In that case, you have a negative variance and you have bought too few licenses. Whether you find a negative or positive variance, both represent a measure of inefficiency.

2. Biggest Opportunity

Your next step is to apply a cost to each variance. To do the Pareto analysis, rank of all your costed variances. If you do this in a graph, you will see that your biggest cost saving opportunities is grouped at one end of the graph and you can easily see the largest (see Figure 2).

One final way to show the information is to group your software titles and potential savings by supplier and rank the suppliers for total potential savings (see Figure 3). Showing this as a pie chart will show you which suppliers you want to look at first as often just a few discussions will bring you large results.

3. Create an Action Plan

A problem solving process will help you create an action plan and implement it. The simplex problem solving process was developed by Min Basadur including four key stages:

  • Assess – if you have already done your data gathering and Pareto analysis you have done this stage. You know what you are using and what you are paying for and how much potentially can be saved
  • Analyze – You have also done your rankings to see where the best opportunities lie. You will need to do some further work in identifying if any of these saving are possible in the time frame that you have. For example, if you have just signed a three year agreement for software licenses with no option to change the terms and conditions, then you can put this supplier to one side and move to the next opportunity. Any opportunities in your action plan timeframe can be included in your planning process (see Figure 4)
  • Planning – This means generating the ideas for efficiency improvement and filtering them to create a list of actions. The actions need timeframes, the people responsible and efficiency targets (ideally measureable in cost savings). The actions that deliver the best cost savings in the timeframe you are working on should be your priority
  • Implement – The next task is executing the actions: you implement and measure the results. This could be a supplier re-negotiation. You will know what you want to achieve before going into the negotiation and then you can measure your success against this. The most important aspect is to close the loop and analyze your efficiency after you have implemented your action plan. Then, go through the process again for your next round of actions to improve your efficiency

Actions for efficiency improvements are not a one-off exercise and you should be seeking new actions. Inefficiencies just happen if you don’t have a process of continuous improvement.

4. Kaizen – Continuous Improvement

Although Kaizen started in production management, it can be applied to many other disciplines including software efficiency management.

Your problem solving process is providing innovative actions and making sure that they get implemented and assessed. Sometimes you will make many small improvements and at other times you may have big wins which provide major step changes. Overall, the cumulative effect will show how much progress you are making over time (see Figure 5).

It is quite easy to show your monthly efficiency improvements and your cumulative improvements against the plan and what has actually been delivered. Such a chart is a great way to show the overall effect of your efficiency improvement over time and get buy-in from everyone concerned.

5. Histograms

There is another simple way to show how effective you are being in driving efficiency across software utilization across an organization; this is by using histograms (see Figure 6).

In software efficiency we are mainly concerned about variations against a zero difference between what has been licensed and what is actually being used (not what has been deployed). There are two key things to look for in these histograms:

  1. The breadth of the base – If you have a short fat distribution curve, then you have a wide variation. This is a sign of inefficiency as you have lots of variations from your optimum point (zero variations). If your distribution curve is tall and thin with most software products near to zero variance then you are highly efficient
  2. The skew of the graph – If your histogram is skewed towards having too many licenses then this means you are consistently over-purchasing. If you are skewed to the negative values then you are not purchasing enough licenses.
6. Software Efficiency Rating Chart

If you are going to create one chart to show your software efficiency then it should be a software efficiency rating chart (see Figure 7). It shows a widely understood way of showing efficiency against some standard variables rated A to G. A and B levels are benchmark whereas D to G mean you should have an active improvement process in place.


According to Peter Drucker, “What gets measured gets managed.”

Peter Drucker was right when he said what gets measured gets managed, but once you have measured things, you need the right tools to analyze the data so that you can make the best decisions quickly. Then, you need to implement and measure the outcomes of your actions. Once you make some progress, don’t stop there; start the cycle again with your continuous improvement process.

About the Author

Phil Hames is the Director of The Business Software Centre Limited.