Klaus D. Goepel, BPMSG, Contact, last update: Sep 4, 2017
389 thoughts on “Feedback”
Hi Klaus,
Your calculator is awesome, helped so much (BPMSG AHP online calculator). I am undertaking, what to some, would be a simple AHP model for ranking importance of several commodities and their value chains (agriculture) for decision making within an NGO, in order to focus concentration on the ones that have the most comparative advantage for beneficiaries.
It seems to be going well using your calculator (I am the farthest thing from a mathematician); I have been able to calculate the pairwise criteria and again the ranking alternatives (so, the criterion are things like market, input supply, government policy, logistics, etc… and the alternatives are the various value commodities such as chickens, pigs, horticulture, etc…), the problem occurs when I get to the section where I must calculate the ranking alternatives table against the criteria weights. I know that sounds stupid, but since I don’t have a simple calculator to do it for me I immediately get lost.
Dear Nathan,
thanks for your positive feedback. Let me try to explain the calculation of alternative ranking in short. In your hierarchy you should have for each sub-category or category a set of priorities, summing up to 100% (local priorities). They have to be converted into global priorities. Just multiply with the percentage of the weight (percentage) one category higher. When you do is for all criteria, the percentage of ALL criteria in the lowest hierarchy level should sum up to 100%.
Then take each alternative and add up the matching criteria percentage of each subcategory. The result gives the ranking of alternatives. Have you seen my AHP video on Youtube? I also explain it there. You can download the slides here. The procedure is shown from slide 18 to 25. Maybe it helps.
Regards, Klaus
Thanks for your template! I looking for some kind of Index.
I will leave a message after my works done with subject-title and contents.
From S. Korea, Grey Wolf!
I really like your concept of consensus indicator and added it to my list of AHP software must-have features as a part of collaborative voting. It would be great if you could add your comments to this list.
Regards, Dawid
Hi David, thanks for putting the link about my AHP consensus indicator on your blog. You can also download my published paper about practical application of AHP, where I gave some requirements for a software. I agree with Stuart: costs are most critical. That’s the reason I used Excel. Now I am working on an online version of the template, so that everyone can make a quick evaluation about priorities using AHP. Group decisions and measuring consensus in the RIGHT way is crucial. You can always do a mathematical averaging, but this is not the way to do it. Partitioning and clustering, as I have done, gives a good insight in the decision process.
A very nicely laid out spreadsheet! Am I correct in hypothesizing that the weights that the spreadsheet computes are appropriate for the same level in the hierarchy and separate spreadsheets should be used for lower level decompositions… such that, for example… if one were comparing location, fun factor, and price… and location decomposed into three factors such as familiarity, difficulty in getting there, and humidity…. the Level One spreadsheet would give the weights for Location, Fun Factor, and Price. The Level Two spreadsheet for Location would give the weights for Familiarity, Difficulty in Getting There, and Humidity. The score for each alternative (just following down the Location leg) would require the participants to score the alternative on Familiarity, Difficulty in Getting There, and Humidity… the Level 2 Spreadsheet could be used to apply the weights for these factors, result in the Location score for each alternative… and the weight for Location would be applied to the Location Score for each alternative when rolling up to the total score for each alternative once Fun Factor and Price have been similarly evaluated?
I am also curious as to your thoughts on how the the consensus element could be used particularly when the participants are from different levels within an organization or across different groups. For example, I would think that it would be very useful to demonstrate the disparity between how the “Business” and “IT” views the world when prioritizing projects.
Dear Eric,
thanks for your feedback.
Ad 1: You are correct. Each category or sub-category of the hierarchy requires a separate spreadsheet. Calculation of the global priorities has to be done manually (or in your own spreadsheet). New: online.
Ad 2: As the consensus indicator is based on Shannon entropy it can be decomposed into two components: within-group and between group entropy. This allows to identify clusters of high consensus (sub-goups of decision makers DM). In my published paper section 5.5 I give an example, where I found 3 sub-groups of DM with high consensus. Interestingly in this concrete project I can assign the 3 subgroups to their functions in the company: sales, senior management and finance. I used another spreadsheet, which I developed for partitioning Shannon entropy. You can find it here: BPMSG Diversity Calculator
Amazing AHP sheets.
I’m a brazilian engineering student and I thinking in recriate your file in portuguese, just for academic purposes. Wish me luck!
Keep going with the good work.
I don’t understand your diversity index calculator please can you help me for applying this index on my data.
My research topic is Diversity of Orthoptera in Gatwala forest.Please reply me as soon as possible
Dear Klaus
I am a Ph. D. student at Institute of Engineering, Nepal. I am using AHP in Evaluation of rural roads projects. I was calculating weight with simple Excel sheet matrix. I found your program is very useful for me and less time consuming and systematic way. But the problem is I have 89 participants and your program is designed for max. 20 participants. I tried to extend for 89 participants and I could not change the formula of matrix M19:V18……………..and others. Could you send me the excel program extended for 89 participants in my email sbbhandari@ioe.edu.np?
With kind regards
Sahadev
Dear Sahadev,
did you see my reply here?
Its troublesome to extend my template to more than 20 participants, but when you use one template for 20 participants each – in your case 4*20 + 1*9 = 5 templates – then you can use a “2-step consolidation”. The second step is to copy the consolidated matrices of the five templates into the 6th, treating each as one participant. Then you will get the consolidated result of all 89 participants. I could help you with the 2nd step, if necessary.
Dear Klaus
Thanks your comments regarding my problems. I tried to solve as your comments but I could not pass the second step as we can put the value from 1 to 9 integer number at ln1 scale and my consolidated matrix is in fractional number. I humbly request you to solve the problem.
With kind regards
Sahadev
Hi Klaus,
Your calculator is awesome, helped so much (BPMSG AHP online calculator). I am undertaking, what to some, would be a simple AHP model for ranking importance of several commodities and their value chains (agriculture) for decision making within an NGO, in order to focus concentration on the ones that have the most comparative advantage for beneficiaries.
It seems to be going well using your calculator (I am the farthest thing from a mathematician); I have been able to calculate the pairwise criteria and again the ranking alternatives (so, the criterion are things like market, input supply, government policy, logistics, etc… and the alternatives are the various value commodities such as chickens, pigs, horticulture, etc…), the problem occurs when I get to the section where I must calculate the ranking alternatives table against the criteria weights. I know that sounds stupid, but since I don’t have a simple calculator to do it for me I immediately get lost.
Any suggestions would be more than helpful.
Best
Nathan
Dear Nathan,
thanks for your positive feedback. Let me try to explain the calculation of alternative ranking in short. In your hierarchy you should have for each sub-category or category a set of priorities, summing up to 100% (local priorities). They have to be converted into global priorities. Just multiply with the percentage of the weight (percentage) one category higher. When you do is for all criteria, the percentage of ALL criteria in the lowest hierarchy level should sum up to 100%.
Then take each alternative and add up the matching criteria percentage of each subcategory. The result gives the ranking of alternatives. Have you seen my AHP video on Youtube? I also explain it there. You can download the slides here. The procedure is shown from slide 18 to 25. Maybe it helps.
Regards, Klaus
Now you can do it online! I have just released a new tool for it. here.
Thanks for your template! I looking for some kind of Index.
I will leave a message after my works done with subject-title and contents.
From S. Korea, Grey Wolf!
I really like your concept of consensus indicator and added it to my list of AHP software must-have features as a part of collaborative voting. It would be great if you could add your comments to this list.
Regards, Dawid
Hi David, thanks for putting the link about my AHP consensus indicator on your blog. You can also download my published paper about practical application of AHP, where I gave some requirements for a software. I agree with Stuart: costs are most critical. That’s the reason I used Excel. Now I am working on an online version of the template, so that everyone can make a quick evaluation about priorities using AHP. Group decisions and measuring consensus in the RIGHT way is crucial. You can always do a mathematical averaging, but this is not the way to do it. Partitioning and clustering, as I have done, gives a good insight in the decision process.
Excellent work. Simple and easy to understand. presentation is very interesting.
A very nicely laid out spreadsheet! Am I correct in hypothesizing that the weights that the spreadsheet computes are appropriate for the same level in the hierarchy and separate spreadsheets should be used for lower level decompositions… such that, for example… if one were comparing location, fun factor, and price… and location decomposed into three factors such as familiarity, difficulty in getting there, and humidity…. the Level One spreadsheet would give the weights for Location, Fun Factor, and Price. The Level Two spreadsheet for Location would give the weights for Familiarity, Difficulty in Getting There, and Humidity. The score for each alternative (just following down the Location leg) would require the participants to score the alternative on Familiarity, Difficulty in Getting There, and Humidity… the Level 2 Spreadsheet could be used to apply the weights for these factors, result in the Location score for each alternative… and the weight for Location would be applied to the Location Score for each alternative when rolling up to the total score for each alternative once Fun Factor and Price have been similarly evaluated?
I am also curious as to your thoughts on how the the consensus element could be used particularly when the participants are from different levels within an organization or across different groups. For example, I would think that it would be very useful to demonstrate the disparity between how the “Business” and “IT” views the world when prioritizing projects.
Dear Eric,
thanks for your feedback.
Ad 1: You are correct. Each category or sub-category of the hierarchy requires a separate spreadsheet. Calculation of the global priorities has to be done manually (or in your own spreadsheet). New: online.
Ad 2: As the consensus indicator is based on Shannon entropy it can be decomposed into two components: within-group and between group entropy. This allows to identify clusters of high consensus (sub-goups of decision makers DM). In my published paper section 5.5 I give an example, where I found 3 sub-groups of DM with high consensus. Interestingly in this concrete project I can assign the 3 subgroups to their functions in the company: sales, senior management and finance. I used another spreadsheet, which I developed for partitioning Shannon entropy. You can find it here: BPMSG Diversity Calculator
Your paper makes an interesting reading…Your Excel sheet helped me a lot in my research…Could you suggest some matterial for Fuzzy AHP?
Thanks for your feedback. Unfortunately I am not so familiar with Fuzzy AHP.
Amazing AHP sheets.
I’m a brazilian engineering student and I thinking in recriate your file in portuguese, just for academic purposes. Wish me luck!
Keep going with the good work.
I wish you luck! Once, translated, I could add it in for download …
Regards, Klaus
We are going to attempt to use this AHP template for weighting Bayesean priors of hardware failure data. Wish us luck.
I don’t understand your diversity index calculator please can you help me for applying this index on my data.
My research topic is Diversity of Orthoptera in Gatwala forest.Please reply me as soon as possible
Dear Klaus
I am a Ph. D. student at Institute of Engineering, Nepal. I am using AHP in Evaluation of rural roads projects. I was calculating weight with simple Excel sheet matrix. I found your program is very useful for me and less time consuming and systematic way. But the problem is I have 89 participants and your program is designed for max. 20 participants. I tried to extend for 89 participants and I could not change the formula of matrix M19:V18……………..and others. Could you send me the excel program extended for 89 participants in my email sbbhandari@ioe.edu.np?
With kind regards
Sahadev
Dear Sahadev,
did you see my reply here?
Its troublesome to extend my template to more than 20 participants, but when you use one template for 20 participants each – in your case 4*20 + 1*9 = 5 templates – then you can use a “2-step consolidation”. The second step is to copy the consolidated matrices of the five templates into the 6th, treating each as one participant. Then you will get the consolidated result of all 89 participants. I could help you with the 2nd step, if necessary.
Dear Klaus
Thanks your comments regarding my problems. I tried to solve as your comments but I could not pass the second step as we can put the value from 1 to 9 integer number at ln1 scale and my consolidated matrix is in fractional number. I humbly request you to solve the problem.
With kind regards
Sahadev