Some Lean Biology

How animals select food and how business needs to prioritize features

Adjugo, Chris Verlinden

It is a very strange coincidence. My only contribution to science is the Digestive Rate Model for Optimal Foraging (DRM, as opposed to the Contingency Model of Optimal Foraging, CM). Did NOT make a major splash in the world of science, although somebody did apply it to a real life situation and found it to correctly predict behavior, so at least I was not wrong.

In my previous life as a biologist, I once was asked to look into an appropriate way to manage hunting Roe Deer. Turned out to be somewhat trickier than expected, one side of it being that Roe deer eat a very wide variety of plants, so it was kind of difficult to see if they cause damage or not. Found some papers on how their digestive tract works, what they can extract from it, and then realized that they can really benefit by selecting (parts of plants) that are easy to digest. Tried to get it published, got the feedback that I needed to link it to existing theory of optimal foraging, the Contingency Model.

The CM assumes that animals select food that can be eaten fastest, so they will leave food items aside if in the mean time they can eat something else faster. In other words, maximize the eating rate. This is the fastest possible rate of food ingestion, so the theory was not really challenged.

Except then by my theory, DRM, which states that animals would select food that yielded most energy during digestion, so would select food to optimize their digestive rate, not their food ingestion rate. And so they would select food items that would result in the highest energy yield during digestion. Meaning, for herbivores such as a roe deer, selecting plant parts with low fibre content, as these digest easily and are most nutritious. Food that passes fastest through the digestive tract would yield more benefit than food with the same energy content but digesting more slowly. In short, they should select items with a high energy yield divided by the time it would need to digest the item, the factor 'digestibility'.

Then, very many years later, I became a SAfe Program Consultant and learned about flow and prioritizing features and Epics using WSJF (Weighted Shortest Job First, see my previous blog post). WSJF is the Cost of Delay divided by the duration it would take to realize it. And it gradually dawned on me that WSJF and 'digestibility' are essentially the same. A digestive tract is a flow system, exactly what SAFe sets up in the Agile Release Train. Animals need to be selective and choose the most digestible food to get the most out of the capacity of the digestive system.  . And that is exactly what business needs to do to get the best result from a backlog of items: choose those items that can be realized fastest and generate the most value per capacity and time.

I remember how difficult I found to generate and explain the DRM flow model. Flow is not a concept we are aware of, we do not get what optimizing for flow means. I found no analogous type of model at the time and the DRM never really struck a chord. Blindness to flow is everywhere and I have to admit that this part of SAFe took embarassingly long for me to grasp. And even SAFe is sometimes a bit wobbly about it, defining WSJF sometimes as Cost of Delay over Duration and sometimes as CoD over cost.

In my paper I also concluded that you cannot optimize behavior by optimizing each individual type of behavior. You cannot optimize eating by eating as fast as possible. The optimization of behavior needs to take into account all alternatives and strike a balance. Now, as lean-agile consultant, I talk about local optimization versus optimizing the entire value chain. The same thing, really...