Why the build-measure-learn process is like cocaine for designers
We are now very familiar with the idea that social media can cause the release of Dopamine from our brain when we receive comments, views and especially ‘likes’ from posts we put up. Dopamine is one of 20 neurotransmitters that carries urgent messages between neurons, nerves and other cells in the body. Essentially dopamine relates to the reward we receive for an action, and according to The Sun, dopamine hits the decision-making area of the brain called the orbital frontal cortex — the same section activated when cocaine addicts are shown a bag of the Class A drug. Now how do I get this hit from work you ask?
From 2017–2020 I worked for a large company that had many teams working within the marketing department. For most of that time I was a marketing designer in the design team, but within the wider marketing team there was also an analytics team and digital performance team that ran A/B tests on everything from changing the colour of a button to swapping out content on a page. Now we have all heard of teams being siloed and this is a perfect example! For the three years I was in this team not once did I have a discussion relating to testing the content I was designing, and only in the last year did I realise I could go over and have a chat to the analytics team and get them to pull the numbers on page views, button clicks and churn.
Fast forward to now where I have moved to a digital agency and have the title of Product Designer. One of the many things I have been introduced to here is the build-measure-learn cycle, and also apparent is the lack of analytics, testing and research teams. Here the analytics, testing and research is under my remit and it’s my job to make sure anything I want to get statistics from is event-tagged by developers and is placed into an analytics dashboard for monitoring.
The build-Measure-Learn cycle
The build-measure-learn cycle works like this…
You can start at any point, but let’s use the example of building a sign-up flow to an app. I create my designs. There are 4 pages that appear consecutively that allow the user to enter their details, get added to the database and create a login which lets them skip this process in the future and get instant access by clicking the app icon.
Once I have created these pages I need each page and each button tagged so I can measure the number (or %) of people clicking through the flow and get an accurate reading of how many people can complete this successfully vs how many drop out.
Once I have these numbers I can work out exactly where people may be having difficulty (they may press a back button, take an extortionate amount of time, have an error occur or just quit the app) and I can learn from these findings.
Perhaps people are turning away when I ask them for their bank details before asking for their name? Perhaps people are getting confused with the amount of information on the page? Perhaps the button to continue through the flow is just not that obvious. Whatever it may be we can learn from our stats, hypothesise, build something new, and measure to complete the cycle again. Does it work better this time? Are we having more people complete the flow successfully?
The Addiction
And this is where the dopamine kicks in. Remember when I said “Essentially dopamine relates to the reward we receive for an action”. Well, here’s a quick explanation of how it works.
When we get rewarded:
An example of this would be when I first looked at my analytics dashboard and saw that 1000 people were using my app.
When we expect a reward:
An example of this would be when I start checking my analytics dashboard after each feature I release for my app, expecting my user activity to increase each time.
When we get irregular rewards:
An example of this would be releasing new features in my app without validating them first. If they were successful 100% of the time, I would expect the reward (which is what we saw in the last graph). If there was uncertainty as to if they would work (lets says a 50% chance), I would get an extra high dose of dopamine during the work and as I check my analytics, from the anticipation and excitement of not knowing if I was successful.
Wait, so how does this work?
Once I have built my sign-up flow and can get (almost) live stats, I can measure how many people are successfully navigating my design at any moment in the day. Every morning I can open my computer, check out my dashboard and see that another 1000 people successfully completed the task they set out to do. This gives me that hit.
I can learn from the data, improve my design, roll it out, and see that 1200 people successfully completed the task they set out to do. I can see that only 2% of users are still using the old way to do something and that 98% have adopted the new way I have just designed. And that right there is the reward I’m seeing for my actions; another hit.
Most social media sites and apps create irregularly timed rewards which is a technique based on the work of psychologist BF Skinner. Skinner discovered that the strongest way to reinforce a learned behaviour in rats is to reward it on a random schedule, and it’s exactly the same for humans. Facebook is a perfect example; we compulsively check the site because we never know when we are going to get a little notification pop up telling us somebody liked something we did.
I can flick between Facebook Analytics telling me how many new users I have, to Medium telling me how many views my article has had, to Instagram telling me another 5 people have liked my photo all to get my ultimate high dose of dopamine.
What if I am a designer with no dashboards?
I should also make it clear that you can measure in a multitude of ways and it doesn’t have to be through a dashboard with a bunch of numbers on it.
Part of a designers job is validating ideas and seeing if one way of doing something is better than another. This is often through usability testing, surveys and observing behaviour, and can have the same effect as the irregular rewards graph.
This time the example would be that I’m going in to talk with some users about which design they would feel more comfortable using. Design A, where I ask for their personal details and then their bank details; or Design B, where I ask for their bank details first before their personal details. I have my hypothesis, the design I think will be more successful, but there is an element of uncertainty and once we have the test results, my design will be validated.
What if I’m a designer that isn’t using the build-measure-learn cycle?
For those who are not using or unable to use the build-measure-learn cycle, don’t worry, you can have some dopamine too.
Apart from the anticipation of rewards, dopamine also triggers the goal-directed-behaviour which is needed to earn rewards. When we have to complete a piece of work in order to receive a reward the mesocortical dopamine pathway activates and gives us the motivation to do the work to earn the reward. This time we get a gradual increase in dopamine that motivates us over time to complete work.
An example of this would be when a designer has to pitch to a client. We get the brief, start the work and as we get closer and closer to the deadline where we have to present the work the dopamine increases to get us across the line, right up until the point of presentation.
Designers have it good!
Strangely this may be one of the best things dopamine could possibly reward you for. Here I am doing my ordinary day job and I get a chemical release making me feel happy when I see it working successfully. If it’s not working I fix it, and if it’s only performing mediocrely I improve it. This is encouraging me to become a better designer, to explore my weakest designs and to develop my design skills in order for a quick release of chemicals, and we haven’t even talked about being rewarded with money yet!
Note: Everything written above is my own theory and I have not carried out any scientific testing on designers dopamine levels. There has been research carried out in order to write this, and I’m trusting those sources.
