≡ Menu

Making Complex Decisions – a method from my MBA days

This is a description of a process for facilitating multi-variable decision-making for any decision, thumb infection but I use choosing a software vendor as an example dear to my heart.

Choosing a software vendor is a difficult task. There are countless options suited to very different types of projects, needs and relationships. The costs range from astronomically high to impossibly low. Between you and the right vendor lies a fog of doubt created by (mostly true) horror stories of projects run amok. Pressing you forward is the hope that you could be on the verge of creating one of those (very real) dream relationships with a vendor that will make you the hero of your company or clients. What I am going to elaborate on in this article is the decision-making technique that I was taught in my MBA program so many years ago at the University of Illinois. The course was titled “Decision Making Under Uncertainty” and its principles are very applicable to the task of choosing an offshore software development partner.

By creating a layer of abstraction between yourself and the options and populating that abstraction layer with clear criteria, you can then view your options through the clarifying filter of your priorities. Okay, that might sound a little complicated. Basically what I am suggesting is that when faced with a decision that involves too many criteria to easily process in your head, a methodical and, equally importantly, a defensible and documented approach is more reasonable.

Most people make decisions based upon an evaluation of the options using some set of criteria that is meaningful to them. This approach is based on the assumption that the decision criteria are well known and can be matched easily against the performance features of the options. However, when there are many decision criteria and a large range of variability between options, this approach becomes very confusing and difficult to justify.

I suggest you begin the decision process by looking carefully at your evaluation criteria. When seeking a software vendor, the criteria may be:

  • Price
  • Cultural fit
  • Technical competencies
  • Timely delivery
  • Reliable process
  • Ease of communication

Next, decide how important each of these is to you. Some people advise ranking them. I disagree. In a list of ten items, ranking implies that item 2 is ten percent less important than item 1. To allow yourself more flexibility to express your true feelings about the relative importance of each criterion, I recommend taking a number and dividing it among the criteria. It doesn’t matter which number you use. I’ll use 50.

Price 10
Cultural fit 5
Technical Competencies 15
Timely delivery 25
Reliable process 20
Ease of communication 20


Oops, I ended up with a total of 95 and yet I am happy with the distribution. That’s not a problem. All we want is relative weighting and we get that by dividing each weight by the sum of all weights.

Price ,105
Accessibility ,053
Technical Competencies ,158
Timely delivery ,272
Reliable process ,211
Ease of communication ,211


I rounded so they don’t add to exactly 1, but your spreadsheet won’t round. Check your formula by making sure that the sum of the weighting numbers is one.

Are you starting so see the power of this method? You now know what you’re looking for. The next step is to look at your options and see how they measure up.

For each option, rate them. Think about how each vendor rates on each criterion. Again, use any scale you want, just so long as higher is better.

Price Accessibility Competency
DynaSoft 1 3 4
ITindia 2 2 3
software.ru 5 4 1


Next, you need to normalize your rankings. We do this the same way that we normalized the decision criteria. Assign to each vendor a value for each criterion that is equal to that criterion’s ranking divided by the sum of all the rankings for that criterion across all of the options. In this case, Dynasoft’s weighted ranking for Price is 1/(1+2+5). Repeat for each cell.

We get something like this:

Price Accessibility Competency
DynaSoft 0.125 0.333 0.500
ITindia 0.250 0.222 0.375
software.ru 0.625 0.444 0.125


Now a simple formula will rank your choices.

Rank of Option(x) = PriceNormalized (PriceWeight) + AccessibilityNormalized (AccessibilityWeight)

If the resulting ranking goes against your instincts or intuition, then rather then beat yourself up trying to justify to yourself your first impression, you can just go back to your rankings and your weightings and re-consider them. If you really think that ITindia is the best choice, then perhaps you ranked them too low or perhaps Price is more important to you than you had originally admitted. If, after careful consideration of your ranking and the way that you weighted your criteria the outcome still doesn’t match your gut feel, then consider the possibility that your gut is wrong.

This same method is also very useful for evaluating options as a group. I introduced a variation of it once to a large holding company in South Africa that used it to evaluate potential acquisitions. Some criteria were clear (NPV), others were not (strategic alignment, cost of cultural integration, skill of management team). By allowing many stakeholders (the board of directors) to establish their own criteria, average them, and then discuss the criteria and ratings objectively one by one, we avoided a lot of needless discussion regarding the relative merits of companies that were not really in the running but only had a few excellent points to recommend them.

Here’s a simple tool (written for me by Szymon Kurcab) that you can use to make complex decisions following the process outlined in this article.


Comments on this entry are closed.