OXAgile 2019 Takeaways

A few of my main takeaways from this year’s OXAgile conference in Oxford, England.  The annual conference is hosted by Elsevier, in their Tech HQ there.

Soft Skills – Culture and Psychology

The predominant theme of the day was on soft skills – “Using Neuroscience to build high performing teams”, “Agile around the World”, “How evolutionary psychology validates the agile mindset”, and “People are more complex than computers”.

The quote of the day for me was “There is no staging environment in Life – you are always testing in Production”.  Great – not only are ‘soft’ skills the hard skills to master, but you’re doing it without a safety net!  This just reinforces (in a particularly geeky way) is why it is soooo important to explicitly establish and constantly maintain psychological safety.

‘Evolutionary psychology’ focused on the difference between introverts and extroverts.  After a review of the brain structure, its chemistry, and a quick trashing personality models such as DISC and Myers-Briggs, David Michel (a self-confessed extrovert) talked about ways in which Agile coaches can inadvertently push extroverted modes onto hapless team members.  For example, by making team members brainstorm and think on the fly, and by creating conditions for conflict during retrospectives, extroverted coaches and team members can overpower their introverted team mates.  Because extroverts process information differently, they can find it difficult or impossible to empathize or even recognise that they are doing this.  Worse, they can take this natural inclination to caution and reflection as being resistance or passive aggressiveness.  

The team building, behavior reinforcing SCARFS Model by David Rock was covered by Elaine Sullivan, during her neuroscience presentation. It’s similar to most SDT models, and has Status, Certainty, Autonomy, Relatedness, Fairness and (Success), but these elements are supposedly tied directly into the more primitive parts of the brain (this review of SCARF taken from a summary of the YouTube video by David Rock)

Status : your perception of where you are in relation to people around you.

Brain science tells us when you experience a drop in status, your brain networks lights up just as if you were experiencing physical pain. And an increase in status it activates a reward circuitry. Giving critical feedback to people can be seen as a threat to their status, at an unconscious level. One way to overcome this is to have people give their own feedback through questions about themselves. This helps their status to go up rather than feel threatened.

Certainty : your perception of how well you can predict the future

Brain science shows that ambiguity can trigger a ‘danger’ response. The brain is a certainty creating machine always trying to predict what happens. When we provide clear expectations about what we expect from people it helps create more certainty. In times of change, letting people know when we might be able to communicate more information about the change can help reduce some uncertainty.

Autonomy : a feeling of having choices, being able to make choices

Brain science tells us that when we feel like we have no say, choice or control, our stress level rises. Giving some autonomy, choice or control dramatically reduces this. People need to know they have choices even when they feel that there aren’t any choices.

Relatedness : feeling safe with people

When we are surrounded by new people, our stress level typically rises. Just seeing some people we know can reduce the stress level dramatically. Bonding with people through a conversation or handshake triggers an oxytocin response. An implication for a leader is to determine if they need to help create these bonding opportunities for people – to feel more in the ‘friend’ rather than ‘foe’ state. Eg virtual teams or teams with people from different cultures may need help to connect and build bonds with others.

Fairness : feeling of fair connections and exchanges with others

A ‘fair exchange’ activates the brain’s ‘reward’ circuitry. An ‘unfair’ exchange activates a danger response. Leaders need to be conscious, open and obvious about treating people equally, more than may seem necessary at times.

Imagine a manager/team who does these things: 

  1. Shows you what is great about yourself, thereby increasing you
    status.
  2. Who provides really clear expectations, increasing your certainty
  3. Who lets you make decisions, increasing your autonomy,
  4. Who trusts you and there is a human bond between you (relatedness), and
  5. Who treats you fairly, and you know they are fair.

These things help generate the ‘reward’ state, literally making people smarter, more effective, more engaged and more productive in the workplace.

A somewhat cynical quote from Krishnan Mani, during a lightning talk.  When a developer says “Well, it works on my machine!”…they could be implying “…and that works just fine for me!”.  I think the message was that  developers may have their own agendas and motivations, and these may not actually include ‘shipping working software’!

Comparing different cultures using The Lewis Model (Richard Lewis) gave us all some self-depreciating giggles….

…and this:

Technically Speaking – Debt by Design

“This is heavy doc: lessons in just-in-time architecture” by Adrian Potter cautioned against being too clever.  Three examples of architectural decisions were given, where in each case the optimal answer was the simple, brute force one.  In the first example, a processing queue was designed with failure handling.  In reality, no automatic handling was required as any failures were rare, and unrecoverable from anyway.  Optimal solution was to ignore failures, and send them for manual review.  In the second, a monolithic component needed to be split.  Instead, the optimal solution was just to duplicate the component.  In the third example, automatic scaling to accommodate variable loading was designed.  It turns out the optimal solution would have been just to spin up a whole bunch of processes in parallel and let them run all the time, instead of trying to dynamically finesse capacity.

In Sean Moir’s presentation ‘How to handle tech debt better’, two gems stood out – from Michael Feathers ‘Legacy code is any code that does not have a test written against it’, and ‘software that can’t be [easily] changed, is actually hardware’.  The heart of the presentation was a framework based on a subjective risk evaluation by representative stakeholders from the whole system, followed by T-Shirt scoring of the risk factor of likelihood of change, impact of an error, and degree of difficulty.  As changes become easier to make, the likelihood of change should increase.

Sean gave us a reason to get motivated about Technical Debt….

It was a fun day, wonderfully hosted by Elsevier – the food was excellent, and the coffee bottomless.  Check out the Bacon Butties….


Paul Osborn

Leave a Reply

Your email address will not be published.