Challenging bias in your agile practice

Let’s challenge the myth that you are capable of discovering objective truth and think without bias. You cannot be unbiased. Every decision you make and every metric you choose will have some measure of bias. Your brain is a bias making machine.

An article by Andy Cleff summarises some of the work out there on bias such as The Cognitive Bias Codex. The Codex breaks down the amazing array of biases into four principal causes:

  1. Too much information
  2. Not enough meaning
  3. The need to act quickly
  4. The limits of memory

I won’t attempt to summarize all of it. I would strongly recommend you look at the diagram. It is pretty amazing.

We find bias mentioned in any work that involves people processing information. In fact, even monkeys display a clear bias to in group versus out group and associate outsiders with bad things like spiders. Academics plead for removing political bias from their works. Reporters are called out for the bias in their work. It is still a favourite debate among philosophy students whether truth can be objective. Even one of the masters weighed in:

The truth is rarely pure and never simple. – Mark Twain

Bias is an insidious little beast that can lead us to make sub-optimal choices. This doesn’t always mean a bad choice though. Our biases are like stereotypes, a short cut to quicker decisions with no respect to right or wrong. They are not always evil of their own accord, but if we lack awareness of them, we cannot make necessary course corrections in our thinking.

Without deliberately remaining aware of why we are make the decisions we do, we will make decisions based on faulty assumptions, heuristics and emotionally driven reactions to situations and issues that need a more nuanced approach.

How bias affects us

Congratulations, if you are successful in your life. You have gotten to where you are based on the decisions you have made combined with a bit of luck. The higher up the ladder and the more successful you are, survivor bias leads you to disregard the luck and place more emphasis on your own knowledge. This is fine until it leads to overconfidence and you step too far and fall. This is just one of many biases.

Biases also surface in other ways. An extreme example is the systematic implicit racism that  exists in many policing systems. This is one end of the spectrum. The other is when we hire for “fit” alone – we can easily end up with teams that look just like us and lack diversity.  We rob ourselves of smarter teams.

In a recent study, researchers provided a bunch of project managers with randomized data that showed a number of metrics on agile implementation. The results were that people who strongly believed in agile saw proof that agile was delivering positive results. Not a problem, unless you are trying to win over someone who doesn’t share the same bias and they see no correlation in the data. You rob yourself of legitimacy.

A slightly less obvious example is when we implement practices that are “agile” simply because someone said they were a good way to go. This is the heart of “cargo-cult agile”. The bias to lean into anything agile without thought is one we have to fight. It may look cool, but let’s stop and make sure they fit our team, and our organizational culture.

When we pretend our biases don’t exist:

  • We look for reports that confirm our opinions
  • Look to our own echo chamber of news
  • Place weight on things that are recent
  • Disregard prior learning
  • Use metrics that lie to us
  • Hire sub-optimally
  • Place greater weight on pre-planning than is warranted
  • Go for the home-baked solutions over an existing (free and completely tested) solutions
  • Do things ourselves that others could do
  • Go for quick solutions and not long-term payoffs

Dealing with our biases

There are some great resources out there for dealing with specific biases. I have a few that help me.

Pre-define your success metrics

How do you know if you have reached your goal? How about by simply defining success before you begin. Defining success metrics ahead of time, you focus your thoughts and evaluate what is important to you.  It also affords you the opportunity to narrow in on exactly what you are trying to do. “Am I trying to get 1 conversion for every 100 impressions or am I defining it as a retention rate of 30% after sign up?”

The idea isn’t a new one. The Centre for Open Science has in fact developed the Open Science Framework to aid in the reduction of bias and cherry picking of results to give a greater level of transparency. Don’t be scared to define your bounds.

This doesn’t mean not reacting to information that the market is telling you, but simply knowing what you are aiming for with the experiment.

Define re-evaluation metrics

Not everything you do is gold. Not every experiment will work. Rather than push forever, consider the bounds that would trigger a re-evaluation.

A very useful exercise for an experiment can be doing a future retrospective as if your experiment has failed. The activity should show a variety of assumptions you had and touch points to watch for. Use the exercise to define bounds for when you should re-evaluate your path. We are trying to be agile, so know when to pivot can be as important as when to push ahead.

Don’t be afraid of it, face the biases ahead of time and make visible the things that are otherwise hidden.

Critically analyse new practices

Many people that have been around development for a while will joke about the latest cycle or fad. We are as guilty of this in the agile space. There really is nothing new under the sun.

When you see a new practice that is the latest hottest thing, the regency effect comes into play. Put simply, human being add more value and weight to things that are new. So when some new framework comes out we want to jump on the bandwagon.

Pausing before the leap and really looking at what we are about to start is invaluable. When we take a moment to move it from the go space to the slow space, we give time for our brains to catch up with the heart. It can be exciting to do the next best thing, but look through the hype to the heart and make a considered decision.

Know your biases and react accordingly

Mindfulness is one of your best weapons in this world of go go go. If you know you have a particular tendency (mine is cool toys), being aware of the effect can be enough to slow you down when you approach a decision or action. You can put steps in place to avoid the monkey brain from kicking in.

An outworking of this can be seen in orchestral recruitment method that uses blind auditions – the practice of having a musician audition behind a closed curtain so that selection panel cannot see anything of the person playing. The musician is judged only by the quality of their music. The result is a much more diverse (both race and gender) orchestra composed of musicians of higher caliber working together.

We can compensate for our biases by working to find our weak points and creating systems or algorithms to address them. These don’t have to be complex. When a developer says we should rewrite using “xyz framework”, simply ask “is it because it is cool and new, or will really solve issues ahead of us?” the space gives you time to test before going boots and all.

Your biases are personal to you. Know them and act accordingly.

A word of warning

Don’t get bogged down in analysis paralysis. We still have to move, our organisations operate and our teams function.  The systems around us are often riddled with implicit bias. Challenging the bias in our organisations, our teams and our own lives is an ongoing journey.

It is about self-awareness at all levels and creating both little hacks and sometimes larger system change that lead us to better results.

Biases aren’t the devil, they are shortcuts that kick in to help our brains get to a solution sooner. We are more likely to be affected by them when we have high cognitive load (stress, anger, overload etc) or influenced by poor systems and culture.

It is, however, possible for us to find ways to prevent them from allowing us make the best decision:

  • Pre-define your success metrics
  • Define reevaluation points
  • Critically analyse new practices
  • Know your biases and react accordingly

Let’s continue the discussion. What am I missing and how can we do better?

More importantly, how can we impact the underlying system conditions that lead to bias practices in our workplaces and lives? How can we enable our teams to challenge status quo or lazy thinking? What can we do to help individuals have a greater level of impact in spite of implicit and explicit biases around them?

How can you make a change today?