Product Evolution by lean selection

Back to products now 🙂

In spite of the odd sounding title of this post, I will start by admitting the contents aren’t novel. I am providing a different vehicle to communicate the key tenets of Lean Product Management from the first principles. I love the vehicle used here (evolution) and my intention in writing the blog is more to enjoy the analogy. I intend to map my learnings in both the fields of Biology and Product Management in retrospect.  A few key tenets on Lean Product Management are given below.

1. Love the problem not the solution
2. Always focus on de-risking the initiative against the available parameters
3. Make small incremental changes to your product and test them
4. Set yourself to quickly respond to feedback

One of the key aspects I try to understand when I am looking at a product is how did this product evolve to the current state? What could the next increment? In that vein, I find the current view of the evolution of a product is a complete antithesis of the word evolution. Therefore, I want to synthesise my understanding of the biological evolution by natural selection and share my perspective on this topic. To me the understanding product evolution is critical to create an environment where the product can sustain and thrive. The understanding also gives us the framework to identify the causalities that led to the success and failure of an initiative.

The proposition

A product is an organism

I consider a product as an organism. A single-celled product is one which solves one problem and cannot be decomposed to multiple components or sub-products. A multi-celled, complex product, however, can be decomposed into a lot of sub-products each solving independent problems. I have detailed my views on products and their composition in a previous blog.

Why this analogy comparing products and organisms?

The broad goal for organisms and products is survival. An organism evolves to maintain a balance in the ecosystem. When the ecosystem isn’t balanced, then the entire system perishes. An organism has to protect itself from its competitors or predators. However, those predators and competitors are needed for the species to maintain the balance. An organism adapts to small and incremental changes to the ecosystem. However, when there is a drastic change fundamentally destroying the primal needs of the species, then the species go extinct. An organism constantly alters itself to test how it performs in the ecosystem. If that change works then it continues and if it doesn’t then it fails. A product operates in an analogous way. It forms out of a need and lives in tune with the environment.

In summary, both these entities are comparable for the below reasons.

1. Live within the ecosystem
2. Save themselves from predators/competitors but also need them for maintaining balance
3. Go extinct as the ecosystem dramatically changes their causal factors
4. Go through continuous incremental changes to test their sustainability

Why the evolution analogy?

Evolution by natural selection is defined as

Organisms best adapted to their environment tend to survive and transmit their genetic characters in increasing numbers to succeeding generations while those less adapted tend to be eliminated.

If I replace organisms with products that perfectly explains the situation on the product space. Also, evolution has some basic traits which make it an amazing vehicle to ride on for narrating the concepts around product management. Evolution creates an organism through random small changes which have survived the test of time.

1. Evolution is non-linear
2. Evolution has no foresight
3. Macroevolution is caused by microevolution over time
4. Evolution is not keen on creating a solution but solving for maintaining equilibrium
5. Evolution happens in small increments and short feedback cycles.

Product Evolution – the prevailing opinion

When the term Product Evolution is used in organisations these days, it communicates one of the below two meaning.

1. A plan of how the product will change over time
2. A roadmap of features and a plan which also communicates the strategy, evaluation and approach

Both these meanings have a fundamental issue. They are both predictive exercises. While evolution is an approach towards sustenance, these two are exercises to figure an approach to get to an outcome which is assumed to give success. This not only assume the problem but also the solution and when it will be needed. Product Teams have to instead set outcomes as boundaries and let the process of evolution guide them.

Since we are in the topic of evolution, I am amused by the irony caused by our prefrontal cortex. The prefrontal cortex is responsible for planning, complex cognitive behaviour and decision making amongst others. I feel humans tend to overestimate the capability of the prefrontal cortex by assuming it can be used to predict the future. No individual has ever been able to do it. We can all fit our recollections with the findings to create a pattern. This is also the reason why planning is different from prediction.

Product Evolution by Lean Selection

My Assumptions: The base intuitions of this approach

I am making two basic assumptions here in order to proceed with this approach.

• People have to reach an outcome. The product is a tool to achieve that outcome.
• The tools adapt to the changes in the environment over time. If they don’t they perish.

First, I would like to illustrate these intuitions of mins with an example. People want to have the streets lit during the nights so that they can see clearly and commute. Initially, everyone carried their source of light. Then, authority got centralised. The government or community had oil lamps for lighting the streets. As technology evolved, people had fluorescent lamps powered by electricity. Now, we have solar powered independent units which power LED bulbs. The outcome is always lit street and it has always been met. The tool to meet the outcome has changed with new inventions entering the environment. If a company made oil lamps and never changed the tool with times to help customers get the same outcome, they would have been out of business when the fluorescent lamps started coming up.

Defining Product Evolution

According to me, Product Evolution is a result of incremental changes to the product. If these changes are guided by the principles of careful selection to sustain in the ecosystem then the product continues to live. The primary attribute to achieve this sustenance is the admission that we cannot have the foresight to predict the tool in the long run.

The best equation to evolution by natural selection is the Price equation which was defined by George Robert Price. The Price equation to get the change in traits across generations is given below.

Z =1/n ∑_i  iZ_iΔ𝑧 = Z’ – Z

Δ𝑧=1/𝑤(𝑐𝑜𝑣(𝑤𝑖,𝑧𝑖)+𝐸(𝑤𝑖Δ𝑧𝑖))

where z stands for a trait, and w stands for the fitness level needed for that trait and i is the index of the subpopulation. Z’ presents the collective traits of the next generation. Cov is the covariance function and E is the expectations function. The expectation function presents factors other than the direct lineage which affects the traits.

The same equation holds good for a product too. If I have to simplify this for a product, then I will end up with

P’ = P + Δp

Δp = ∑_i  O_iWhere O is a product trait to reach an outcome. The product team will ensure that Δp continues to improve the covariance function so that the traits improve the fitness level of the product.

Though there are plenty of valid criticism of the Price equation, I feel it is the best approximation we have got to mathematically represent the Evolution by Natural Selection. A change in a trait is influence by how fit the trait is with the population and the expectations of the trait in the population.

In the product context, a product increment is a collection of traits each intending to cause an outcome. A product increment which doesn’t increase the product’s fitness quotient in the population will have to be removed. The smaller the quantum of change the shorter the feedback cycle and smaller the impact both positive and negative.

Lean Selection – A way to achieve this outcome

As mentioned before a product can be single-celled or multi-celled. Consider each cell as an individual component or sub-product. Each cell can therefore individually evolve. Let us take the example of the eye. The evolution of the eye happened in parallel with the evolution of the brain or any other organ. Each organ has a specific goal and goes through incremental changes. The collective communication system between the organs called the neural network also evolves with time. The product evolution has a similar approach. The biological evolution is also fundamentally lean. By lean I mean it is risk-averse, economical and controlled. For example, the entire population doesn’t go through the same change, only a tiny fraction does. If that change helps the population then it propagates. I call it the ‘Lean Selection Experiment’.

The core principles of lean Selection experiment are

1. Identify a need
2. Make minor changes to one component
3. Test it with a small population
4. Get feedback
5. Decide whether to drop those changes or continue

A look at Lean Selection Experiment: Eye as a product

I want to first explain the evolution of the human eye. Charles Darwin once said, till this day the eye makes me shudder. Each step in evolution happens as an experiment. The only difference between Lean Selection and Natural Selection is that Natural Selection is a random process while LeanExperiment is a more controlled process. Natural Selection doesn’t actually have an understanding of the problem space. In Lean Selection, we can artificially do that by performing generative research and running tests.

Eye: Experiment #1

1. Need: Need to know when the is a movement in my environment to protect myself
2. Solution: Develop a  sheet of light-sensitive cells to detect the difference between light and dark.
3. Test: A small number of species develop a light-sensitive sheet
4. Feedback: Can detect a change in environment
5. Decision: Continue with the change

Eye: Experiment #2

1. Need: Want to know the direction from a predator is coming
2. Solution: Bend the sheet into an indentation
3. Test: A small number of species get a light-sensitive indentation
4.  Feedback: Can detect the direction of the predator based on the shadow
5. Decision: Continue with the change

Eye: Experiment #3

1. Need: Need to know the nature of the predator
2. Solution: Bend the sheet into a deeper cup
3. Test: A small number of species get a light-sensitive cup
4.  Feedback: Can detect the rough shape of the predator
5. Decision: Continue with the change

Eye: Experiment #4a

1. Need: Need to know a more precise nature of the predator
2. Solution: Bend the sheet into a pinhole camera kind of eye (Mollusc called Nautilus)
3. Test: A small number of species get a pin-holed camera
4.  Feedback: Can detect the shape of the predator
5. Decision: Continue with the change

Eye: Experiment #4b

1. Need: Need to protect the eye
2. Solution: Have a transparent sheet to protect
3. Test: A small number of species with a transparent sheet protecting the eye
4.  Feedback: Eye is protected
5. Decision: Continue with the change

Eye: Experiment #5

1. Need: Need to a clearer image
2. Solution: Expand the sheet to make it into a lens
3. Test: A small number of species with an eye lens
4.  Feedback: The image is clearer
5. Decision: Continue with the change

Eye: Experiment #6

1. Need: Need to a clearer image
2. Solution: Form a coating to form a reflecting tissue
3. Test: A small number of species with a reflector eye – eg. a Scallop
4.  Feedback: The image is clearer
5. Decision: Continue with the change

These experiments are not linear which means Experiment 5 is not after Experiment 6. Also, one could stop after a specific outcome is met. Further, this evolution of the eye happens in parallel to the other components of the organism. When we run multiple such tests we end up with multiple solutions for the same problem.

I suggest reading this text to get a complete understanding of the evolution of the eye.

MVP in different contexts

Since its conception, the word minimum viable products has metamorphosised into multiple acronyms.

1. Minimum Viable Experience (MVE)
2. Minimum Compliant Product(MCP)
3. Minimum Reliable Product(MRP)
4. Minimum Delightable Product (MDP)
5. Minimum Testable Product (MTP)

I am not sure what was the intention of each of these but for me, an MVP is a minimum set of outcomes which gives me enough feedback to take a call on whether to proceed with the next experiment or otherwise. In some instance, it could just be a single outcome and in others a set of outcomes.

I sometimes get into a discussion on how to compete with other players using MVP. If one intends to compete with Google search engine or Amazon eCommerce or eBay auctions, then their starting position is incorrect. One can never achieve get MVP this way. The starting position could be to check if there is a market for secure searching or a market for searching academic journals.

Here is a design for an experiment

Search Academic Journal: Experiment #1

1. Need: Need to search peer-reviewed academic journals on a specific subject
2. Solution: Manually create a collection of academic journals. Create a mechanism to search these journals
3. Test: Share it with the University students who study the subject and measure the usage
4.  Feedback: Talk to the students to learn more.
5. Decision: Continue if the students come back repeatedly.

As you might see in the above context, I am storing the documents. I am not crawling any server to find these documents. I have manually stored and indexed these documents. This is a way to come up with an MVP without intending to directly compete.

What is required for this setup to succeed

There are 3 critical components for this approach to succeed.

1. Infrastructure
2. Team
3. Culture

Infrastructure

Teams need the right infrastructure to run these experiments. The infrastructure should allow teams to test a small sample of the population, roll back, scale the changes without much effort. If infrastructure becomes a bottleneck then the experiments tend to get bloated thereby eroding the principle.

Team

The team needs to have dedicated capabilities to perform the following functions.

1. Research customers
2. Identify the experiment
3. Design the experiment
4. Develop the solution
5. Measure the success
6. Manage the finance
7. Coordinate with other teams

Culture

The team culture is another key component for the success of this. There are 7 principles

1. Love the problem not the solution.
2. Make small incremental changes to your product and test them
3. Plan the experiment and not the solution
4. Set yourself to respond to feedback both within the team and from the customers
5. Respond just in time to avoid waste
6. Do not fill your time with solutions (product managers with more stories, designers with solutions and developers with lines of code)
7. Collaborate with other members instead of competing with them

Conclusion

In conclusion, I feel enabling a product to evolve through the lean selection criteria is the best way to create a healthy sustainable product. The lean selection criteria drive product teams to explore the changes in the business environment, constantly engage with customers, run small experiments and adapt to the feedback. This pushes the team to constantly look at outcomes rather than solutions.