Most decision making should be based on logical approach, using data and methodologies, that is why consultants often use evaluation forms as a tool to simplify decision making. The evaluation forms are built on two main things: evaluation criteria and the weighting of each of those criteria. This tool can be used for Benchmarking comparison, vendor selection, or any other analytic decision making (business or personal such as buying a car/house, choosing a candidate -etc). I will describe below the steps to create an evaluation form and and I will share with you a basic example of how to compare a set of options.
1.1. Select the key criteria which are important for the evaluation
1.2. Give a weighting to each criteria
1.3. Select which of those criteria have a positive or negative impact on the scoring
The steps to create an evaluation form are described as follow:
1. Creating the model1.1. Select the key criteria which are important for the evaluation
1.2. Give a weighting to each criteria
1.3. Select which of those criteria have a positive or negative impact on the scoring
2. Data gathering and preparing
2.1. Gather and clean the data
2.2. Normalize the data
3. Total up the weighted average for each score, then total
1. Creating the model:
1.1. Select the key criteria which are important for the evaluation
We start by selecting the criteria on which the comparison will be based on. It is possible as well to group them into categories (or parent criteria), in blue, and sub-criteria, in grey.
1.3. Select which of those criteria have a positive or negative impact on the scoring
Not all criteria can have incremental effect of the final score. For instance, the higher the price, the lower the score should be.
2. Data gathering and preparing
The data gathering and cleaning in this example took me less than 15 minutes since it's a very simple model and I only used one source of data. It usually takes me more than 50% of the total time before reaching the actual result. Try to make your model as robust as possible to updates and new entries, it will save you a tremendous amount of time while re-rendering the results.
2.2. Normalize the data
Normalization takes numerical values and changes them into a common scale. For example, the Safety ranges from 1 to 5, and the Fuel take is between 58 and 70. The big difference in scale might cause problems while analyzing or modeling them. The normalization will give a common scale, between 0 and 1 in our case, across all criteria and make it possible to compare them.
The values between -1 and 0 in our example are those with a negative impact on the scoring. They will reduce the score once totaled up.
Normalization Formula = (value - min ) / (max - min)
Normalization with Impact Formula = Normalization Formula x Impact
Note:
You can follow me on Twitter and add me on LinkedIn.
Related posts:
Free Open Data Sources
2.1. Gather and clean the data
2.2. Normalize the data
3. Total up the weighted average for each score, then total
Example: Evaluation form for car selection
We will in this example compare few relatively similar brands and models to select which among those cars will be considered the best pick.
1. Creating the model:
1.1. Select the key criteria which are important for the evaluation
We start by selecting the criteria on which the comparison will be based on. It is possible as well to group them into categories (or parent criteria), in blue, and sub-criteria, in grey.
1.2. Give a weighting to each criteria
For each of those criteria, give a weighting (total of 100%). Sub-criteria have their separate weighting totaling 100% themselves.
Note: Weighting becomes optional when you have too many criteria.
Note: Weighting becomes optional when you have too many criteria.
Not all criteria can have incremental effect of the final score. For instance, the higher the price, the lower the score should be.
2.1. Gather and clean the data
2.2. Normalize the data
The values between -1 and 0 in our example are those with a negative impact on the scoring. They will reduce the score once totaled up.
Normalization Formula = (value - min ) / (max - min)
Normalization with Impact Formula = Normalization Formula x Impact
Note:
- Impact equals to 1 or -1 as displayed in the model
- Min and max values can be directly related to the sample of data you have, for example, the Fuel Tank min and max values are 58 and 70 respectively. They can also be forced for a wider range, for example, Safety min and max values are 1 and 5 respectively.
3. Total up the weighted average for each score, then total
Follow the next steps:
1- Multiply the normalized values of the sub-criteria (in grey) with their respective weighting.
2- For the criteria in blue, or parent criteria, if they don't have sub-criteria multiply them with their respective weighting exactly as you did for the sub-criteria.
3- For the other parent criteria having sub-criteria, sum their respective sub-criteria first, then multiply that value with its weighting.
4-Total all the parent criteria.
5- Multiply the total by 100 to make it easier to read.
I hope you found this helpful. You can download the Excel file in the following link: Evaluation Form. Please leave your comment below if you have any question or any recommendation to share.
Related posts:
Free Open Data Sources
Comments
Post a Comment