imageContinuing along with my occasional discussion of Lean Six Sigma topics which apply well to farming, here is the next: the Pugh Matrix (or decision matrix method). We are often faced with making a decision between multiple choices which have complex variables. In engineering and manufacturing firms, obvious examples are deciding between two major design path choices, or selecting a vendor who will supply components long-term or perform some sub-contracted duty.

These are decisions where there are many pros and cons between all the choices, and it can be overwhelming trying to choose which is the best solution. The worry is that if we just default to our “common sense,” we may end up being biased and unable to make a truly objective choice. We may unconsciously place more importance on a certain consideration than other critical factors; and in the end, not select the best solution. With vendor selection, it can be easy to be swayed by one you know well and like; or by a good salesman. With design choices, the most assertive person in the room can sometimes sway the group in one direction. With farming, especially animal selection, the potential for bias towards our favorite animal, or best-looking animal, is huge.

In order to objectify a major decision, it helps to start by putting your decision criteria down on paper. What would make the perfect solution? What things are important to have or to consider? If you (or a team) can quantify all of the critical decision points, this helps to see the big picture when weighing the potential options.

Once we write down a list of all the considerations we’d like to weigh when making a major decision, we can score them according to importance. This lends to an easy-to-understand matrix; where we evaluate each potential choice against all the important factors, and weight each one objectively.

Of course, this is why it’s called a decision matrix. It’s a grid: the decision factors are rows, the potential choices are columns. There are many variations on how these are done: sometimes you just put a plus, minus or zero in each box. Some may allow a ++ or a – – to make a  wider spread. Sometimes you weight some factors more heavily than others with a multiplier. But the overall idea is, you decide the decision criteria and scoring weight ahead of time; before you start evaluating the choices against them.

Once you have the decision factors on paper, then you independently score each potential choice or solution option. Finally, you add up the overall score to help you see objectively which choice is superior to the others. The method is simple, yet profound.

When engineering teams use these for “big” decisions, the matrices can be large, complex, and take a lot of time to create and populate with data. This is the whole point, as it makes a very complicated decision very manageable. Ideally, decision matrices should be created and populated by a team, to help maintain objectivity. But, it turns out, the concepts of a Pugh Matrix can help us in everyday situations, and even an individual can do one.

I used one last summer when I had to choose between four rams I was interested in purchasing at a friend’s farm. Selecting breeding stock is a classic Pugh kind of case: there is never a perfect animal, only different combinations of pros and cons. The breeder sent me some photos, and tidbits of information and musing on each ram. I couldn’t decide, finding my mind revisiting the pros and cons of each one, and no obvious choice was rising to the top. This is a clue that a Pugh Matrix is in order, to organize the information in front of you into something sensible.

So, I just jotted down my most important considerations:

  • Look: does he look healthy, in good condition, no glaring conformation faults which impact productivity, does he have good meat characteristics?
  • Does he have NSIP metrics available on himself, or at least, his parents?
  • Is he RR (scrapie resistant, and will pass on resistance to all his lambs)?
  • Does his dam have a strong twinning record?
  • For the ram itself, and his sire and dam, do they have “is a” twin or triplet status? (Arguably, I’m allowing twinning genetics to appear in the list twice, thus double-weighting it, but that’s ok, because twinning genetics are very important.)
  • Is his birth weight within my ideal range (not too small, not too big)?
  • Are his 60 and 120 day weights strong?
  • Do I like what’s in his pedigree?
  • Is he completely unrelated to my ewes?

I scored each ram choice against all of these factors, using a zero for an I-don’t-know or neutral case; a –1 as a detractor, a +1 as a better-than-average; and sometimes a half- or quarter point- fraction if it was somewhere in between. I weighted each factor equally, for simplicity’s sake. The resulting “answer” looked like this:

image

This helped me see right away that there were two top contenders, and two lesser rams. Before I did this exercise, I could not see this at all: they all seemed on-par with each other for different reasons. And one of the two lower scorers, I thought, was the most attractive-looking ram; so there was my bias. I reserved the final choice for when I could see the rams in person; but I brought this printout with me, to anchor my decision-making in objective criteria.

The seller unfortunately did not have growth data on the rams, so I couldn’t incorporate this into my initial analysis. Once I got to see them in person, I could see that the second one was much larger than the rest; and was just beefier and more well-muscled. So, lacking better data, I made the assumption that he had the best growth genetics.

I don’t like making this assumption, for two reasons. For one, I’ve found that just because one animal looks heavier than another, doesn’t mean that’s proved right once he’s on the scale. Two, the two rams I was comparing were born two weeks apart, one was a twin, the other a triplet, and their dams were different ages. So it’s unfair to compare them head-to-head without first doing some adjustment for their age and rearing type. But I was evaluating them at six months of age; so these differences are starting to diminish by this time. They are a much bigger factor when comparing 60-day weights.

In the end, I ended up adjusting my scores of these two rams slightly: to discount the first ram for lesser appearance/condition once I saw him in person; and give the second ram extra credit for appearing to have the best growth. This made the second ram the clear winner:

image

One time in the past, I was in a similar situation of picking a ram from a set a friend had. I didn’t get data on them ahead of time, and just arrived there with them all sitting in a pen waiting for me to choose. They were all white, all about the same size, similar in body style and conformation, and there was very little to differentiate them visually. My friend couldn’t remember any stats about their parents off the top of her head. I waffled and waffled and finally ended up basically randomly picking one; as I felt helpless in the face of zero data.

I like having a tool much better: it allows you to quickly arrive at a conclusion, and feel confident that you made the best choice given the information available. The Pugh Matrix is a very simple and elegant way to make a data-driven decision.

Advertisements