Tipping the scales … and seeing what falls out
By nature I tend toward impulsiveness. In an effort to compensate for this character flaw trait I try to take decisions of consequence somewhat seriously. I no longer remember where the techniques described in this article came from, but my process involves four activities:
- I list the alternatives for my decision
- I list the considerations that factor into my decision
- I assign weights to the considerations according to their importance to me (high weights signify greater importance)
- I score each alternative on each of the weighted considerations (high scores are favorable)
In my experience the process is iterative—that is to say, I keep adding and rewording considerations while I assign and change weights and scores. Sometimes additional alternatives come to mind as I mull over my priorities.
Once I’ve weighted and scored everything I sum up the total weighted score of each alternative. The alternative with the highest total weighted score “wins.” Sometimes the outcome surprises me, but most of the time it helps me to understand why I find a decision difficult: the total weighted scores of several alternatives are very close.
Sometimes my reaction is disappointment; I subconsciously want one alternative to prevail but the numbers don’t justify it. I take this to mean one of two things:
- Either there are unstated considerations that I should endeavor to dredge up and quantify, or
- I’m kidding myself about what really matters and I need to face facts
When this happens, I usually begin reweighting the considerations or adding ones that are missing. Frequently I find that I don’t want to admit what is really driving my impulses. This discipline of having either to write down and evaluate my motives or to abandon them outright aids me in my continuous internal war against idolatry.
Marriage also helps. Decision-making is more difficult for couples than for singles, of course, but given complementary strengths the effort pays off.
And speaking of marriage, the decision-making technique I’ve described here benefits not only me but also the better (but less bloggerific) half of Casa Jonsson. Whereas I tend to fly by the seat of my pants and not look back, Araceli’s tendency is to second-guess decisions. A disciplined process helps me to think things through conscientiously, and it helps her not to worry so much and to be more decisive.
When I wrote the first draft of this entry two months ago I included example decision matrices from recent family decisions: which neighborhood to move into in Spring 2000; whether I should go into business with friends in Winter 2002; whether to move from our current neighborhood in 2003. I didn’t need a decision matrix to help me realize that I don’t want The Internet Archive to cache and keep serving up that sensitive stuff until kingdom come. I wrote:
While it may not make for scintillating browsing, in some ways I consider these matrices to be among the frankest disclosures we’ve made on Casa Jonsson to date. Random thoughts and carefully worded opinions are one thing, but geological surveys of sacred ground are quite another.
But this entry would be incomplete without at least one example. I think I can live with letting you know why we just joined Netflix in favor of continuing to rent movies from our corner video store.
| Consideration | Weight | Alternatives | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Netflix | Walmart.com | Blockbuster Online | corner Hollywood Video | ||||||||||
| Raw | Normalized | Weighted | Raw | Normalized | Weighted | Raw | Normalized | Weighted | Raw | Normalized | Weighted | ||
| Summary | 2.2 | 2.5 | 1.6 | 1.8 | 1.7 | 1.9 | 1.9 | 1.9 | |||||
| Average monthly cost | 3 | ($11.99) | 3.0 | 3.0 | ($12.97) | 2.8 | 2.8 | ($14.99) | 2.4 | 2.4 | ($16.67) | 2.2 | 2.2 |
| Size of movie library | 3 | 40,000 | 3.0 | 3.0 | 20,000 | 1.5 | 1.5 | 30,000 | 2.3 | 2.3 | 10,000 | 0.8 | 0.8 |
| High-quality listings data | 2 | 3 | 3.0 | 2.0 | 2 | 2.0 | 1.3 | 2.5 | 2.5 | 1.7 | 2.5 | 2.5 | 1.7 |
| Rapid delivery | 2 | 2 | 2.0 | 1.3 | 1.75 | 1.8 | 1.2 | 1.75 | 1.8 | 1.2 | 3 | 3.0 | 2.0 |
| At-a-time movie limit | 1 | 2 | 1.5 | 0.5 | 2 | 1.5 | 0.5 | 3 | 2.3 | 0.8 | 4 | 3.0 | 1.0 |
| Community features | 1 | 3 | 3.0 | 1.0 | 1.5 | 1.5 | 0.5 | 1.5 | 1.5 | 0.5 | 0 | 0.0 | 0.0 |
| Monthly total movie limit | 1 | 4 | 1.5 | 0.5 | 8 | 3.0 | 1.0 | 8 | 3.0 | 1.0 | 8 | 3.0 | 1.0 |
| RSS support | 1 | 3 | 3.0 | 1.0 | 0 | 0.0 | 0.0 | 0 | 0.0 | 0.0 | 0 | 0.0 | 0.0 |
| VHS and DVD formats | 1 | 0 | 0.0 | 0.0 | 0 | 0.0 | 0.0 | 0 | 0.0 | 0.0 | 3 | 3.0 | 1.0 |
Google tells me that the above may be a form of Pugh matrix. But what do I know. People in need of industrial strength decision support advice ought to enlist the services of a professional, especially one for whom tossing a ball in the backyard becomes an exercise in projectile physics.
But if you’re interested in trying out my simple decision-making technique for yourself, here’s the above matrix in Excel workbook form. I’ve built in some helpful stuff like a summary page that automatically gives you a little happy face next to the best alternative(s).
Use positive numbers (greater than zero) for weights, and non-negative numbers (greater than or equal to zero) for scores. The Scored Considerations tab (see below) is for registering your opinion about aspects of the alternatives; these are subjective measures reduced to numeric scores.
The Quantified Considerations tab (see below) is for recording quantitative aspects of the alternatives; these are objective measures such as price, estimated return on investment, cubic feet of capacity, etc. When bigger is better—such as with return on investment—use positive numbers (greater than zero). When bigger is worse—such as with price—use negative numbers (less than zero). For example, the cost consideration below is expressed in negative numbers.
Happy decision-making!
Update: I revised the example based on Howard’s feedback. I also changed the way normalization works for quantified scores. Previously, the worst price (for example) of $16.67 normalized to 0 out of 3. This produced a disproportionate difference between the highest normalized score and the lowest normalized score among alternatives, considering that the best price ($11.99) wasn’t far from the worst. Now, the worst price of $16.67 normalizes to 2.2 out of 3. 
This is incredibly fascinating. You’ve put more work into quantifying and deciding from whom to rent videos than I have for quantifying and deciding basically every single decision I have ever made.
I guess that tells me something about my decision-making process.
My method is to list pros and cons in my head, and gnaw on that for a long time, and then make a snap decision based mostly on impulse.
I’ve always wanted to use the sort of system that you’ve got, but I always get too impatient and rush it. I admire your decision-making stylee.
I may have misrepresented myself here (for better or for worse, I don’t know). I don’t employ this method very often. I do have six more of these spreadsheets going back to Fall 2000, but invariably the decisions I take this seriously involve large investments of money and/or time. You could say that this is also true of watching movies, but I was more thorough here largely for the purpose of having a robust illustration of the technique.
My time may have been better spent analyzing how much time I should spend watching rented movies instead of doing something of more lasting value. :-( One consideration that doesn’t appear on my matrix is the side benefit of Netflix’s monthly rentals limit. It can be a governing device for a movie-lover like me who too often gives the rest of his life short shrift.
Despair not, Keef—all this rigor notwithstanding, you probably end up making better decisions most of the time than I do.
Nils, very impressive. Please submit your resume to our HR dept.
A few observations.
1. The difference between #1 & #2 seem tiny (weighted scores of 1.81 or 1.82). Accuracy does not equal precision. Perhaps 3 significant digits is way too much for something like this.
2. There is a concept of “Options” you may want to take into account. My guess is that the online selections require a credit card and automated billing. What are the chances that you’ll not take full advantage of the # of rentals per month? The Hollywood store may be the worst choice with a given set of assumptions. But what if you take a annual view? i.e., How many months will you rent 8 movies? How many months will you actually rent 4 movies? If you guys watch LOTS of movies, every month without fail, the analysis is good; but if there are months where you are leaving $ on the table, you might want to take that into account. Which selection gives you the option of “not paying in full that month”, and how important is that (and likely to happen)?
3. This may also be a case of satisficing instead of maximizing utility. Unless, of course, you get satisfaction from knowing that you are optimizing (which I suspect you do). It could be that speaking on absolute terms, spending $12–17/month for meeting your movie rental needs are sufficient. However, I suspect that it’s important to optimize. As such, the previous comment will have less meaning. If you have a month where you are traveling and somehow will not need to rent 8 movies; you’ll do it anyway. Maximizing utility itself brings you utility; satisficing is not good enough.
4. Some grocery stores are now doing $1/day rentals thru unmanned kisoks. Fewer selections, but $1/day! Another option is to do a Hollywood + $1/day rentals. Unless selection is important.
5. From economists, job interviewers, pollsters to management consultants, we like to ask “what did they actually do in the past?”. Then project from that into the future. What kind of movies did you actually rent? How many movies did you rent? Then make a list of movies you’ll likely rent in the next year, A real list is not important, but you might want to say “I rent 3 new releases/month” or “I will rent 20 foreign flicks in the next year, including these 5 must-sees”. How do your options stack up?
6. So you have the scores in this iterative process; then what do you do? Well, you might wat to timebox your decision — I will stop analysis by the end of April, for example. “If the delta between #1 & #2 is less than ‘x’, I will roll a dice”. Then when you decide, decide also that you’ll be happy with the decision (until you revisit the spreadshhet in two months!).
Good luck!
Now you know why they pay Howard the big bucks!
Good points, Howard. In light of your remarks I’ll admit that “satificing” is preferable here. The process is probably overkill since one family’s opinion is all that matters in this decision. Business decisions that involve many alternatives, considerations and objective measures call for full-blown optimization software and consulting.
Thanks for the feedback.
Somewhere this has the potential to go horribly wrong. Imagine if we all quantified our utility for every possible situation. This is how it might appear on an actuarial exam:
I-Mac (a professional basketball player) is currently visiting a city and trying to decide which night club to attend. I-Mac has determined that he has the following interpersonal utility funtion U = .5*E + 1.3*S + 2*(W/H-2)^2 + .8*HC + .5*P
Given that S = (.2,.8) (male, female) and indifference everywhere else, what proportion of tall caucasian women will maximize I-Mac’s interpersonal utility?
Second-guessing
Now that the decision is made I’ve learned about a new option that may upset my carefully stacked apple cart: Peerflix. It’s Netflix meets eBay. I don’t have a DVD library to speak of, but trading movies with a Netflix-like…