top of page
  • Writer's pictureAdam Timlett

Beyond Feedback: Leveraging Biology in Tech Org with Life History Theory

Updated: Jul 9, 2023

What do the following have in common?

  • The James Webb Space Telescope

  • Bananas

  • Sliced Bread

  • Ice cream in a cone

  • The Empire State Building, NY

  • King Charles III

JWST, Source: NASA


The answer is that they all go efficiently and reliably from one life stage to another.

Form and Purpose in Nature

People sometimes ask what the purpose of an organism is. One answer, is to look at what an organism’s ‘purpose in life’ is, and for higher organisms the answer is food and sex.

Organisms don't have a purpose they are just viable, which is how they get to exist. Often they are viable in a wide variety of different environmental conditions, just as often, they are only viable in quite a specialist niche in their environment.

Technology people want to separate what to build from how to build it

In lots of technology projects what to build, how to build it are considered two logically separate things. We often tell people not to solutionise: "Don’t tell us what to build, but tell us what you want, then we will figure out how to build it". Content & method of delivery: 2 separate things. What it is and how it gets there are separate.

Comparison with Natural strategy

What would this strategy look like in Nature? Perhaps like the thought experiment by machine learning researcher Kenneth Stanley of a ‘forgiving earth’. In 'forgiving earth', every organism that one can imagine exists, anything that someone asks to exists gets produced. There are no constraints on form to be produced. Most things would make no sense, however, Kenneth says the world would be filled with grey goo. But a small amount of things would be things we recognise that do exist and likely lots of things we've never seen. In reality, however, Nature imposes constraints on what is viable and things that are not fitting to these constraints, are not seen. There’s no grey goo, (unless it’s actually some sort of surprisingly sophisticated slime mould in a cave somewhere).

So, if we look at these constraints it turns out there are patterns, and the pattern is the precise opposite of the idea of separating cleanly, form and how that form is delivered. Instead, what it is and how it got there are deeply intertwined.

Everything that is viable in Nature, bears the hallmarks at maturity of the organism, of the way that it got to maturity.

The same is true of all the examples I gave in the pop quiz at the start:

  • A banana that is half peeled bears the hallmarks of how it was stored and transported.

  • The James Webb Telescope once unfurled, bears the hallmarks of how it got to Lagrange Point 2 in outerspace.

  • King Charles III bears the hallmarks of how he got to be King. (By waiting. A long time).

One major example or hallmark of an organic system is that it is 'designed' or takes the form, of an organism with the capacity to generate meaningful feedback, especially from generating options as part of the design/feature of the organism itself. And we see this most clearly with Sex. Unless we're actually from outer space, that is how all of us got here.

What is sex for? The role of Sex in generating feedback

Sex is the efficient generation of variability and portfolios of options which are part of the 'design' if you will of the organism that allow it to generate enough individual differences from one generation to the next for viable organisms to make up some fraction of the population. That's what sex is effectively 'for', otherwise we could just reproduce asexually like bacteria. Asexual reproduction works for bacteria because of the huge numbers of cells that can be generated that way, such that a meaningful number will have mutations just from errors. Sex works differently by deliberately shuffling the genes in each generation. Even though it is far less efficient than asexual reproduction, you guarantee a lot of differences in just one generation.

The key point here is that the 'design' of the organism makes significant concessions to the whole life history of the organism. To deal with the problem of generating enough variation, the content or form is changed.

In the case of Sex, Nature makes major concessions to the problem of generating meaningful feedback and this fundamentally affects the form or ‘design’, if you will, of organisms. Form of delivery and form at maturity are not logically separated, in contrast to typical business thought.

Form of delivery affects form at maturity

In the concept of tech, the separation of form at maturity from form of delivery means there are arbitrary designs with no concession of whether they will ever be able to measure and generate the appropriate feedback and so be viable. It’s like animals from Kenneth Stanley’s ‘forgiving earth’; We often see technology products delivered of a form that can’t generate meaningful feedback. Tech org thinkers are often complicit in producing these products because, they say to the business, “tell us what you want, whatever it is, and then we will figure out how to build this arbitrary design.”

Some agile thinkers recognise this issue, such as Mary Poppendeick, and she did a whole analysis on the Empire State Building and how the design of the building made concessions to the problem of how to build it quickly. The design at maturity was changed by the need to avoid creating unnecessary dependencies between different aspects of the building. The building bears the hallmarks at maturity of how it was built so quickly and efficiently, (such as the large external electric junction boxes which are not integrated inside the walls but sitting outside the internal walls).

Actions have consequences, sometimes

So what are the consequences for tech org if we take this further? For example, in a lot of agile type thinking, proponents will argue that you should seek to make lots of small adjustments and constantly respond to feedback. But in a large majority of actual business contexts any signal in your feedback from a small change delivered will be completely swamped by noise. Your moves, or the content being delivered, if small, will often just be far too small for meaningful feedback to be generated. Any stats you use, if the changes are too small will just be random information, not meaningful feedback. Hence, the form required for getting viable feedback, partly dictates the content to deliver to receive that feedback.

Just like sex, you should look at implementing forms of content that involves making big enough changes in one generation to be able to generate real impact that can actually be measured. However, note that this depends on analysing for that life history stage (the getting feedback life stage), what size of content is necessary to get reliable feedback.

In different business contexts what change is big enough, varies, just as different organisms adopt different life history strategies. For example a business with a large number of users of a product may generate meaningful feedback from a small change, (like asexual bacteria). But a business producing a product for internal use will not, it needs the change to be pretty big, in just one generation.

Life History Theory

This is where we can leverage theory in biology to help us to analyse the problem. Getting reliable feedback is part of the constraints to do with the life history of an organism. Hence, the organism has a series of hoops to jump through at each stage of its life and reliable feedback from the environment is one of those hoops, for the population as a whole. In biology a theory called ‘life history theory’ has been constructed to analyse different strategies for doing this. Each 'hoop' at each life stage imposes a constraint on the forms that are taken of organisms and often there are trade-offs between different parts of the life history. Different organisms adopt different strategies to get through all the necessary hoops.

Hence, each part of the strategy for survival over the whole life history is reflected in the 'design' if you will, of the organism. Because it has no purpose other than viability, there is no distinction between what the organism is ‘for’ (e.g. what form it takes at maturity) and concessions to that purpose of form to choosing viable forms that allow a method of delivering that form at maturity.

The right concessions to make on the form at maturity depend on the particular strategy chosen.

Life History Theory also gives us certain concepts that we can leverage to help us to understand how form should be analysed in terms of various stages of the product. The most important thing to understand is that there are many different viable strategies, just as there are many different kinds of organism. So rather than focus on one strategy, the idea is to characterise and understand what the different available strategies look like and what determines if that is a good strategy for that niche. Then you can choose the life history strategy that actually suits your business context.

r/K life history strategies

The r/K theory is the idea from life history theory, that Nature has two basic potentially successful strategies when producing offspring. One is suited to a highly uncertain environment, and means producing as many offspring as you can. The other works in a more stable environment and means making sure the largest number of offspring reach maturity.

r refers to the gradient of an exponential growth rate. K refers to the carrying capacity of the environment, the number of organisms it can theoretically support.

r strategy

An r strategy maximises reproductive rate and suits highly uncertain environments. An example is salmon.

K strategy

A K strategy maximises efficiency of the carrying capacity of the environment by bringing offspring to maturity with a high success rate. Examples are elephants and humans.

Applying this to business contexts

Salmon are r strategists, elephants are K strategists. In general we can think in terms of life history strategies of a whole population. Is the quality going to be high enough and uncertainty about the environment low enough for a K strategy? For an r strategy, is the number of options we explore large enough to get a fraction reliably returning value in an uncertain environment. i.e. is r high enough?

So, there is a clear trade off in business too, and the strategy choice is reminiscent of the r/K heuristic in biology. The key to leveraging biological research is to understand that every business situation is different, and we need to identify the right strategy for the context.

By the way, if you're not thinking in terms of portfolios of options to generate feedback, and just sequences of feedback and iteration like a PID controller or some cybernetic control system, then you are still thinking mechanically rather than organically. You're still more K than r because r implies parallel options not sequential options.

The K as information content of products in economics

Let’s look at what this means for the design of delivery teams. In the work of complexity economist Cesar Hidalgo he contrasts high knowledge or information content products versus low information content products, and connects this to long-term GDP. Again, this directly relates to carrying capacity but of an information channel, such as the cognitive capacity of a person or team doing work. We might say that if a business adopts a high K strategy it seeks to produce products that are the result of a very high information content and high expertise in the teams that produced them. A low K product has relatively less information content in it, such as the case of mining natural resources. A high K product requires a team member with specialist knowledge to mature and reach expert level with a high success rate.

We also see within tech work, how some products requires staff with a high knowledge and expertise level. If you produce products with a high K strategy, for that particular life history strategy to work, you need to ensure your staff can transfer that knowledge, just as animals must transfer skills if they are K strategy animals like from parent elephant to next generation elephant. But Cesar Hidalgo shows this is hard to do. It's surprisingly hard to transfer knowledge from one country to another to copy another country’s industry. Eg try building modern high performance semi-conductor factories outside Taiwan. It’s not easy to do. You need a network of talent and businesses around the main engineering firm and decades of expertise in a location to be able to generate the high information content required by that semi-conductor product.

Domain Driven Design comes in here

This is where documentation comes in and 'ubiquitous language' to raise that success rate of the new members of the team receiving a high amount of information content from the prior members of the team. These are the tell tale signs of the code repository at maturity, that show how that product might actually reliably get to maturity and reliably adds new features in a high K strategy.

On the other hand, you may utilise a relatively more r type strategy, e.g. If the skills required are generic, and the expertise required is lower, and the talent pool is far larger. Then you can adopt a different life history strategy, in which turnover is acceptable or even a positive thing, because it refreshes the energy levels of staff, etc. Examples are where deploying reconfigurations of a system for businesses. It is implementations and operational. Not designing very complex things from scratch. But above a certain level of product market complexity, this more r type strategy isn’t as viable.

Take home messages

· To understand which life history strategy suits your business construct analytical models of different life stage 'hoops' to get an intuitive grasp of what each life stage looks like. For example, how big should new features be to generate sufficient feedback to see the signal over the noise for your product?

· Then plan a strategy over the whole life history of the product, where each hoop should be successfully negotiated and designs are informed by the needs of the strategy. For example, a K or an r strategy for reaching product maturity.

· This might include 'Domain Driven Design' type work as a potentially viable component of a life history K strategy to increase the K of your product and reduce wasted effort and staff failing to reach maturity in their expertise in supporting a product.

· Then, once you intuitively understand what you’re trying to measure and what’s important to your specific strategy, you need metrics to check if your product’s life history strategy is on track.

Recent Posts

See All


bottom of page