Agile Vancouver: Trust and Team-Building

I went to a very interesting talk at Agile Vancouver last night and thought I’d share my notes. I love human psychology experiments and this was packed full of them. Combine that with software development and you’ve got a winner.

The talk was given by Linda Rising and titled “Who Do You Trust?”

Trust is the most important factor in implementation of the Agile software development methodology because social dynamics & interaction are the biggest cost driver in software development – bigger than better tools and methods.

A previous talk of Linda’s has covered the Agile placebo effect… does Agile software development work because we THINK it’s going to work? We go into an Agile project expecting things to work.

She started with a disclaimer that, amongst other things, this is a “Presentation of a disturbing nature”. A nice introduction that got everybody’s attention.

In Computer Science, we never do experiments… proper scientific experiments with a hypothesis, observation etc – seeing results from a change in development methodology doesn’t prove anything.

She introduced the Robber’s Cave Experiment – conducted at a campground in Oklahoma in 1954. Two teams of 12 year old boys, same backgrounds, balanced teams. Both teams were transported to the campground separately, didn’t know the other team existed.

The experiment was conducted in three phases:

Phase I: The 1st week was spent in isolation… each team quickly “became an us”: gave their teams names, designed a flag, places became “our” swimming hole, “our” firepit etc

Study team gradually let them know the other group existed – but not yet seen. This resulted in immediate division, saying things like “I hope THEY don’t use OUR swimming hole” – even though the makeup of the other group was identical, there was nothing different between the two sets of children.

Phase II: The two groups were introduced to each other and staff organized competitions & events with a winner & loser, prize trophies and money. Deliberately designed to create friction between the teams.
The trophies were displayed in the common mess hall where both teams could see them and talk about their victories.

As phase II progressed, one team burnt the other team’s flag – then the other team retaliated and burnt their flag… eventually staff had to intervene.
Then they started raiding each other’s cabins at night. More retaliation led to rocks being stockpiled for use defending themselves against attacks. Again staff intervened and had to bring phase II to an end.

Phase III was about having fun together. They scheduled non-competitive activities eg watching movies, eating together.

(At this point Linda compared this action to corporate team building exercises – you bring together different parts of the business, different teams, and just expect them to enjoy mingling)

Phase III failed – the teams were still far apart – there were food fights and yelling. If you’ve ever been on a corporate team building exercise, this probably isn’t a surprise to you.

Linda pointed out that you might expect behaviour like this if the team divisions had been based on things like religion & politics.
But we also see it in development teams and other relatively trivial settings.
For example, during software project recaps, you hear people talking about “us”, “them”, “the others”.
eg something trivial like maybe some of the developers got a T-shirt, others didn’t. Maybe this was caused by a distributed team. The T-shirt is not the actual problem but is a symptom of thinking about the distributed team differently. – “oh, I forgot to order enough for the guys in Victoria”.

We see the “others” as the enemy. And yet we think of ourselves as unbiased & rational when it comes to decision making.

Psychologists see this as a hard-wired reaction – so it must have an evolutionary benefit. Our ancestors had to answer lot of classification questions quickly: is this food edible or not edible? Is this person a friend or a foe? The decisions are made very rapidly.

Another experiment: conducted by Jane Elliott, a 3rd grade teacher in Iowa in 1968. She was trying to give her class of white kids a feeling of what it might be like living in the US as a black person. She separated the class into blue eyed and brown eyed children and told them that it had been scientifically proven that blue eyed children were better. There’s a TV documentary you can watch about this experiment here: http://pbs.org/wgbh/pages/frontline/shows/divided/etc/view.html

The part of the outcome that’s interesting to us right now is that the kids were all as good as each other and yet the kids who were set aside “knew” that they weren’t as good as the others – they bought into the labelling.

Research says managers sort employees into winners/losers as early as 3 weeks after starting to work with them. At an Agile conference Linda asked a table of managers if this was true. They said “well yes, but we are always right”

If you get two groups of people who disagree on an issue and give them the same paper on the subject to read about it, both groups will say that it supports their point of view. We see what we want to see. The managers alter whatever facts they see about the employee to suit the decision that they’ve already made.

Everybody makes mistakes. However we forgive our own behaviour but not others. “MY intentions were good even if I missed the deadline” – people’s judgment of their own performance is “contest sensitive”, but their judgment of other people’s behaviour is absolute.

So people get stereotyped. But people are complex. When we label people we lose appreciation of their other talents. BUT interestingly we also do this to ourselves.

Eg: take a group of maths students and give them a hard test. When the test starts with a male/female tick box, men outperform women – as this is the gender stereotype. Remove the box and give the same test and both genders perform the same. Similar research playing video games has shown that both genders will give up easier if they’re playing as a female avatar.

Another example is the Solomon Asch experiment:

A group consisting of one test subject and a number of actors are asked simple questions related to line lengths on cards. The test subject is always asked last.
For the first couple of rounds, everyone agrees then the 1st actor gives a clearly wrong answer & the others agree with them.
A significant number of test subjects go along with the incorrect answer too.

Recent research shows that if you put the test subject in an MRI machine you can see that there’s no debate going on in their head – they’re not thinking “should I say this – I don’t want to look out of place”. They actually see the wrong line as correct. The actors set the filter for the line lengths & our brain believes it.

So if other people can lead us to believe that about line lengths, how about more significant things?

So what have we got now…

  • We’ve shown that stereotypes change our behaviour.
  • And we’ve shown that our behaviour affects other people’s behaviour.

And so we have a self-fulfilling prophecy: we judge our employee as not being very good and treat him accordingly and so that’s how he behaves. Hence the managers at the Agile conference ARE always right about their judgments.

So how do we change this – how do we use these things to our advantage.

Rule #1 of good management: catch your team members doing something right and praise them for it.

Linda mentioned she’d been to project retrospectives where they always started by repeating Norm Kerth‘s Prime Directive: “Everyone is doing the best job they can”. It seems cheesy but after a while the team members start to believe it’s true & eventually they create that in the team – everyone IS doing their best.

Back to the boys in the campground: the study team created problems that the groups must work together to solve.

They cut off the water to the campground and set the boys to search for the “leak” along the mile long pipeline. Required all the boys to work together. They discovered a clogged valve… they said “we” found it & celebrated together.

Other tasks followed that required everybody to work together to accomplish, culminating in a ‘last night of camp’ where they sat around the same camp fire and alternated singing songs for each other. Both groups insisted on going home on the same bus and the team who’d won the most prize money bought everybody milk-shakes.

Linda suggested maybe this experiment wasn’t surprising… after all the boys were all very similar. The experiment was repeated in 1963 in Beirut with a mix of Christians & Muslims. During phase II there was serious fighting and three group members threatened a member of the other group with knives that they’d stolen from the camp kitchen. The study team intervened and cancelled the experiment – there was no Phase III. To be expected?

However the groups were NOT divided by religion. The three group members who had the knives were all Christians – and the opposing group member that they threatened was also a Christian.

In this case, group membership trumped religion even though the religious divisions were hundreds of years old and the groups had only been together for a week.

So there are two responses here that are hard-wired. One is that we quickly judge people into groups as “us” and “them”. The other is that we like to work in small teams and we like to collaborate.

So to resolve conflict it’s necessary to cooperate on shared goals.

Similarly to the campground, this must combine the entire organization – it can’t just be development.

Agile practices help with this:
– face-to-face communication increase cooperation. Strongest effect of any variable
– Stand-up meetings
– pairing – produces better result than either individuals could achieve individually
– short iteration timeboxes means everybody gets frequent goals to work together on
– project retrospectives

Liking someone is not required for these activities to pay-off
Liking someone is different from “respect”

Social interdependence requires:
– common goals
– outcomes affected by actions of others
– individuals only reach goals if others reach goals

In a collaboration, nobody succeeds unless everyone does. Therefore efforts must be coordinated.
The coordination produces respect.
We all like being trusted and respected.
It follows then that Agile teams get trust and respect all day and are therefore happy 🙂

This has a positive impact and results in effort improvement in both individuals and the group.

Linda has a book out called ‘Fearless Change’ that documents design patterns for introducing change into organizations. People tell me I’m lucky at draws but last night was one of my unlucky nights… in the door prize raffle the guy immediately to my right won a copy and the guy immediately to my left won a copy and I left empty handed. Although I did leave with a head full of interesting information!

4 comments to Agile Vancouver: Trust and Team-Building