First, you have to study the processes in question; the area where you plan to invest the money.
A great way to do this is to chart out a sequence of work steps using TQM or Workflow type notation. Process flows should be annotated with volume, frequency, and duration measurements showing the volume of data, how often the data is entered or used, and how much time passes between key events in the process. I found the best way to represent this is using an annotated data flow diagram on a series of charts or using a charting program.
Once you understand the process, you can then ask ‘what if’ specific software technology would be deployed? How much more could be processed, how much faster, etc. Why would this happen?
Once you have this data, you can then try to figure out what the performance of the process is vs what it should be: actual vs standard.
Working through these questions together with workers eliminates risk, ambiguity, and communication problems. There are all kinds of forums to do this, such as Quality Improvement Teams, etc. I usually just ask the supervisor for permission, then sit down with workers and find out what is going on.
Once this work has been done for each process, you can then compile a list of all the sub-processes into an aggregate picture.
For example, the following table shows 10 subprocesses within a claims processing area. All of these processes deal with defective postal addresses.
The Old Volume column is the observed current frequency (occurrences per month) of each process (t0). The New Volume is the frequency that this process should have after the installation of Group 1 Software technology (ts). The Duration column show how long each occurrence takes.
The $/Hr column is the cost constant c for each process. This is either a unit cost that is fixed per occurrence, or a temporal cost per hour that must then be multiplied by the duration.
As discussed in the first post of this series, you can then multiply the volumes by the cost and derive current and new standard cost projections. Add all these and you can determine what the savings might be. In the example above, the value of a relatively simple correction in defective postal addresses is approximately $200,000 per year.
If the savings justify the change, you then have to ensure that your savings plan is realized.
One important step is to ‘instrument’ the processes. Install measurements on a daily or weekly basis that show the changes in volume and/or frequency. As the software is installed, you should start experiencing a steady shift towards the expected numbers.
If the expected benefit isn’t detected or the numbers aren’t what they need to be, you’ll need to sit back down with the knowledge workers and ask what is or is not working. Is the software not installed correctly? Are there gaps in training? What are the workers’ objections, if any, to change?
There are dozens of reasons why something could go wrong, only a few why they will go right. Successful implementation requires steady focus and teamwork, and, most importantly, quick corrective action.
The final step is to make the benefits visible in your company’s operating budget or profit and loss statement. Every accounting department has their own style, but in the end it has to fit into the Generally Accepted Accounting Practices (GAAP).
The simplest method is to use historical process performance numbers in your budget, then reflect the actuals as the improvements go into place. The difference then should be reported against the investment, in a separate report, as the final project ROI disclosure. Eventually you’ll be expected to re-adjust your budget to take advantage of the better performing process.
Next, we’ll take a look at some contemporary quality improvement programs or management programs such as Six Sigma, and how value calculations are accomplished under different conditions.
