To calculate the ROI, you must convert approved business measures to money.
We are now in the data analysis phase. The next step is where we want to decide if we are going to convert this data to monetary values. We are going to assume, yes, we are going to convert the data to monetary values.
In this situation we decided, yes, we are going to convert our data into a monetary value if it is not already a monetary value. Now we are still going to use our design thinking principle number six. We are going to use the same process model, and we are going to use the same tool kit that we used before to decide about the credibility of our impact data. This is the data we will report. This time the question is “How do we convert data to monetary value?” Let us determine a unit or measure and convert that into money.
Using our Business Alignment V Model, we are still working with the impact data. We have impact data already isolated, so we can claim that this is what our program produced. Now we want to convert that isolated impact data to monetary values.
Four Types of Hard Data
There are four different types of hard data. These are output, time, cost, and quality. These are the types of impact data that we will be looking at if we have hard data. Now the thing about hard data is they are objective. If we are going to go pick apples, we know how many apples we picked. Hard data are easy to measure and quantify, they are easy to assign a monetary value, and they are common measures used by organizations. They have a lot of credibility with management. Now, on the other hand, we have soft measures.
Please reference the list of soft measures in the Application Guide. Soft measures are subjective, and they are a little bit more difficult to measure and quantify. But you can do it. You also can assign monetary values to them; it is just a little bit more of a challenge to do. Assigning monetary values to these soft measures is a little more difficult, and they have less credibility as a performance measure. They still have credibility and they are typically behaviorally oriented.
We are dealing with hard data and soft data and most of the time we can convert the soft data into hard data. We will note this on the ROI Analysis Plan worksheet.
Again, this is our ROI analysis plan that we started working on during the planning phase. In the last lesson, we examined how this is used to plan how to isolate the effects of the program, which is the first step in the process of analysis. Now what we are looking at is how are we going to convert this data into monetary values. This is an important distinction at this next step. Using this analysis plan upfront, we are working with others to determine how we are going to convert the data to monetary values. We typically work with finance or some other branch or division in your organization to help us to do this.
Types of Data Conversion Methods.
Standard values are those values that the organization has agreed on are true or they agree on the value. They understand that if we produce one widget, it costs us this much. If we have a waste of one widget, this is how much one unit of waste costs. Or, if we are talking about how long it takes to do something, we know how much we pay the employees, so we know how much it costs to do it. These are standard values, and we do not have to demonstrate that they are valid numbers. The other methods that we have require a little bit more work to accomplish. They add to the cost of data conversion, which goes into the cost of our evaluation. They relate to historical cost. This is where you go in and find what things have cost in the past. You can go to experts, either inside or outside your organization, who have the expertise to be able to give us a value. You can go to external databases, such as trade publications or professional organizations that may have it. You could also link it with other measures. We have estimates. Estimates are a great technique to use as well. They are further down the credibility scale, but they still have value. Whenever we are dealing with estimates, we must make sure that we adjust for bias as we have done in the past with isolation. We also must make sure that we take care of those extremes or unsubstantiated claims.
These are the different data conversion methods available. Now we will show you how to link data with other measures.
In this example of a predictive model by Sears years ago and published by the Harvard Business Review, they wanted to know if they could look at employee attitudes and be able to make predictions about customer perception, which would then make predictions about the revenue base at the Sears stores. What they discovered is that if people had a good attitude about their job and a good attitude about the company where they worked, it influenced customer behavior and caused them to shop at Sears. If there was a five-unit increase in employees’ attitudes about their job and the company, that is going to influence their behavior and cause them to stay. This five-unit increase would, in turn, cause a 1.3-unit increase in the customer's impression of the store by the employees being more helpful, talking up the merchandise, and showing increased value. This gave the customers a positive impression causing them to make a positive store recommendation to other people, and help them to come back to the store, causing a 0.5 increase in revenue growth
What they discovered is that increasing an employee's attitude, defined as a compelling place to work, would drive an increase in customer satisfaction impressions, making it a compelling place to shop, which in turn, would drive an increase in revenue growth, making it a compelling place to invest and also to shop. It worked well. They had five different task forces, 80 videotaped customer focus groups, 80,000 employees’ surveys, and a million ideas from employees. Sears looked at 20 years of data. They put a lot of work and effort into this. If you look at Sears today, you know they have fallen on hard times. If you look at why, and you go back and look at this model, you will find out what was in this model is also predictive of the reverse of what would happen with Sears. You will know their growth and people's perceptions of working there by going back to the impressions that people had in the workplace. Now, a lot of people say it was online shopping, but there is a lot more to it than that. Online shopping caused a decrease in customer flow to the stores and affected employee's attitudes. It all works together.
You have predictive indicators of taking data from one source and transferred over to something else. Whenever we are dealing with converting data to monetary values, we are always thinking about one unit of measure. Boil it down to one unit of measure. What are we trying to improve here? When we are converting data to monetary values, what does and where does that unit of measure come from?
Five Steps to The Money.
What we must do is figure out how much one unit of measure costs. The five steps show how we can convert a unit of measure into money. We get those units of measures from the level four data and we focus on one unit of measure. We convert the unit to value using the techniques we examined of how we can put a price tag on it, whether it is a standard value or something else. Then we want to figure out how much change in the performance has occurred. The last step is to annualize it. If we average it on a monthly basis, we multiply it by 12 and then we multiply that by the total value.
In this example, we have one grievance. We are having a problem, two organizations where people complain about things and when they file a grievance, HR must investigate. You go to HR and they said, how much does one grievance cost? They said each grievance, on average, costs about $6,500. That is good information. There is a program in place to reduce grievances. We want to reduce the causes for grievances. What we discovered after our program is on average, we were able to reduce the number of grievances by 10 a month. That is great, but we cannot claim all 10 because we are going to isolate. How many of these 10 are a result of our program? We are told it is seven. We can legitimately claim seven and defend it. Now the question is, what is the annualized change in performance? If it is seven per month, we multiply that seven times 12 it gives us 84. We can avoid the $6,500 costs for grievances by avoiding 84 grievances. What is the value of that? We said it was $6,500 per grievance, which means that we are going to be saving over half a million dollars through our program. If our program costs less than a half a million dollars, we get a positive ROI.
In another example, for those of you in sales, if you have one sale, you know it is a 30% profit margin. Everybody agrees that is what we use in the organization, and your program was able to increase the revenue per month by $20,000. How much are we dealing with annualizing for the monthly revenue increase? After we isolate it, we have figured out that it is $240,000, because the revenue per month was isolated. We are looking at $240,000 over 12 months. What is the value of that? Well, we cannot claim all of that because that's revenue. We can only claim the profit. So, you multiply the revenue by 0.3 and that means that we have been able to improve the profits of the organization by $72,000. Again, if our training program costs less than that it is a positive ROI.
Data Conversion Issues
When we are converting data, we always want to use the most credible sources. Remember, credibility is in the eye of the beholder. Define who is credible during the planning phase. Get everybody to agree. We also want to make sure that we use the most conservative option available if we have two credible sources. The whole idea is we want to drive that ROI down. Then, we will also want to adjust for the time value of money. The finance department will be happy to help you there. Last, we must know when to stop this process. We cannot just keep spending money doing this. It will nickel and dime us until our expenses exceed our costs. We must be careful when we are dealing with data conversion.