Today I'd like to browse into the waters of economics. There are certain concepts that I believe are highly relevant to the way we build software products. You might be wondering, how do these two fields come together, but bear with me, they are more connected than we often realize. Understanding some basic ideas of behavioral economics can help us steer software projects as it provides us with a different perspective on the prioritization and value-maximization process, which I write about frequently.
Before we get to the marginal utility we need to understand some basic concepts. What is a utility anyway? It is an economic term that refers to how much satisfaction a person will get from consuming a particular good or service. The important part to understand is it's the total utility one gets from consumption of certain goods. This is given by the sum of satisfaction one gets from consuming all the individual parts of it.
Let's look at an example. Assume you are eating a pizza. The total utility you get from that pizza is how much satisfaction it provides for your particular needs when eating all the single slices. In this case the need you satisfy is typically just hunger (need to eat and/or cravings).
How can we measure utility? Unfortunately, most of the time it is not directly possible. You can't objectively measure how much satisfaction eating a pizza can bring. On top of that it is highly individual. The satisfaction (utility) you get from eating pizza will probably differ greatly from the satisfaction that the same pizza will bring to Bob, who is allergic to gluten and ate lunch a few minutes ago.
But what we can do is to compare the value that different goods or services can bring to you. In a given moment you can compare the utility of a pizza compared to that of a steak. The one you'd rather eat is the one with the higher utility to you. You can also compare it to the satisfaction a particular water would bring. Or a car. Or a house. You can assign some abstract dimensionless value to each of the potential goods (similar as we do it with story points) and those can be compared.
It is obvious that utility, even for a single individual, varies greatly depending on the context. You might say a house would be of a much higher value than a bottle of water. But would you still insist on this if you were parched and lost somewhere in a desert? The famous line
My kingdom for a horse is a good demonstration of this phenomenon.
Utility is context dependent and that is also why people are willing to pay different money for the same good when the situation changes. Buying used skis is usually going to cost you less in summer than in winter because people don't demand them as much. Trying to sell ice cream on a beach in summer is an easier job than doing the same in winter time. Context matters for utility.
Knowing that, what is a marginal utility? This concept measures how much utility will the next unit of a good or service bring. The total utility is then the sum of all marginal utilities you get from individual parts of that good or service.
Back to that pizza. Eating a first slice will give you much satisfaction because you went from starving to having something to eat. That was the marginal utility of the first slice.
The second slice will probably be almost equally satisfying because you are still hungry. But will the 8th slice be as satisfying? Probably not because your hunger is gone by the time you get to it. So gradually we usually tend to get lower and lower marginal satisfaction of each additional unit of a particular good we consume.
An interesting fact is that marginal utility can also be negative. How much satisfaction would eating the 24th slice of pizza bring you? It would likely not be something you would voluntarily eat (unless you're attending some who-eats-the-most-pizza contest). That one extra slice of pizza could make you sick and therefore cause a negative experience. The marginal utility of the 24th slice would be negative and the total utility of the whole meal would start getting lower.
How does all this connect with software development?
We're building a software product that serves people. Their motivation to use it stems from the fact that they get some satisfaction out of it. Their needs get fulfilled when something that was once difficult gets easier thanks to our product.
When building a software product, we have limited resources to fulfill those needs. So we all know we should maximize the value of the time (or other resources) we invest into building it. That is no surprise.
How can the economical concept help us? I have seen a natural tendency by the product owner (and the rest of the development team including the stakeholders) to attempt to define backlog items as the final form of a software feature; a form that immediately brings users the
fullest and most
perfect possible behavior.
I won't go into deeper detail here, but if you are interested please check one of my
previous posts on this topic .
Understanding the marginal utility concept can help us overcome that psychological barrier of not wanting to reduce individual backlog items to atoms. We want to bake the whole pizza with prosciutto and mozzarella on top of it, serve it with water and wine on a table with a nice view of the sunset because that is what we imagine the product is supposed to look like eventually.
But let's view it from a different perspective. Why do users want our software product? If we use the pizza analogy, what is the
first slice of pizza that we can give them? The slice that will not make users feel fulfilled yet and at the same time cause them to not die of hunger anymore. We can apply the same approach to the whole product backlog preparation and simplification of the individual items.
Frequently a user story as it is written solves multiple users' needs at once without us realizing it. Remember the email example from the post I mentioned above? A typical user story could say that when you view an email, it gets marked as read, it gets rich-text formatted, embedded images get displayed etc. That is how you would imagine such a feature to ultimately behave.
What happens, when we apply the marginal utility concept to that story? What is the
first slice of pizza that we can develop and decide about the rest later? What aspect of that story has the highest
marginal utility for the user? Most likely the ability to actually read the text in that email. The
marginal utility of being able to read the content is typically greater than the one of rich-text formatting, image display, or marking messages as read. Suddenly it becomes obvious that the story can be broken down into more atomic slices.
By asking this question over and over again usually helps to make up the mind as to how worthwhile it is to invest effort into something. It also makes negotiations with stakeholders easier because it somewhat materializes the cost/benefit concept.
Know the users
As stated before, utility is highly individual and context-dependent. That is why the user stories start with
As Bob the truck driver, I'd like to have.... A software feature for one person will likely have a totally different value (utility) compared to another. That is the reason it is vital to know well who we are building the product for. To know their real-life problems.
Only then can we evaluate the marginal utility of our product's features relevantly
for them. The capability to view images embedded into emails is going to give wildly different satisfaction to a clerk receiving simple instructions in the text compared to a gallery owner wanting to receive previews of paintings from artists.
When we know the audience we can reasonably decide how we can split backlog items into more atomic ones and identify what
the next slice of pizza should be. That's also why getting real-world feedback as early as possible is so crucial for relevant decision-making. Describing utility highly contextual to a particular person is the reason why
user stories are called
user stories in the first place.
Managing the backlog well is a difficult job. We need to keep it meaningfully prepared, juggle with the item's values, compare the costs and returns. That's why it's good to take a chance and look at the same thing from a perspective by applying some basic economic concepts to it. I hope this sparked some inspiration for your own development process.