Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage Articles Peopleware 2015 - An Interview with Bradley Scott of Xero

Peopleware 2015 - An Interview with Bradley Scott of Xero

At the recent Agile New Zealand conference Bradley Scott gave a talk titled “Peopleware 2015” in which he explained the management structures, policies and approaches Xero has used to support its agile transition.  He discussed how they worked and presented some ideas on the future of management. 

The slides from the conference presentations can be found here.

After the talk he spoke to InfoQ about his experiences and ideas.

InfoQ: Good day this is Shane Hastie for InfoQ we are at the Agile New Zeeland conference, talking with Bradley Scott. Bradley you are the General Manager of Practice Products at Xero. What does the General Manager of Practice Products do?

Bradley Scott: Xero builds a number of products, and some of those products are targeted at small businesses but we also make products for accountants and bookkeepers to help make it easier for them to do their work. That is my focus area. My job is to look after the product managers who are working together to come up with the roadmap, and also all the people that come together to make the product real.

InfoQ: With that focus on “looking after the people” you just gave a talk here at Agile New Zealand, talking about “Peopleware 2015”.  Why peopleware?

Bradley Scott: When I started managing developers early in my career, one of my colleagues gave me this book and I loved the perspective that it came from- that a lot of the problems aren’t to do with your processes, your methodology, the programming language or the tools that you use. Actually, they are people problems.

You need to think of solving people problems and not solving technical problems. I wanted to spend some time taking that theme and applying it in the modern day context in 2015. In doing so, I wanted to use a lot of examples from Xero, my own thinking, and also referencing other organizations that are doing interesting things.

InfoQ: One of the things that you drew out of the book was that the things that they were finding way back in 1987 about productivity across individuals and teams still apply, and the fact that actually the tools, technology, all the technical stuff make very little difference to productivity. So what does make the difference?

Bradley Scott: In that original study the author looked at an exercise called the coding war games. They found that people from the same organization tended to perform alike. There was something about the organizations themselves that was determining how effective these organizations were and how effective the developers were.

Throughout the book the author goes through many different theories about the cause of this relationship and their particular interpretations of what was important. What I talked about is my thinking that culture is the single most important factor that makes you successful or not.

You need to think about and invest in culture and I went through a number of examples of things that we do. Such as: trying to eliminate incentives that are counter-cultural or disappoint effective people; how to change salary reviews that so they don’t get in the way; how you can empower people by allowing them to arrange in a way that just best gets the work done.

InfoQ: Let’s tackle these things. Incentives – a big hairy problem from any organization, people want them, do they? And we want to recognize and acknowledge good performance. How do we build an incentive structure that is going to work?

Bradley Scott: I think you are right to say that people like incentives, I think it’s probably more that they want to be recognised for their good work. The proposition I talked about is- doing that on an annual basis isn’t the right way. One of the models that is becoming more and more prevalent is to take the focus away from the person or the organization giving people rewards at the end of the year and instead create a system where peers can reward other peers, in a really timely and meaningful way. You can make that happen financially as well, so you still have a bonus scheme but that bonus scheme works by the wisdom of peers and people’s observations day to day rather than have a manager trying to determine that on a pseudo-objective-subjective basis.

InfoQ: So it’s not just your peers it’s also constant. 

Bradley Scott: Yes, that is right, it’s regular and it can happen at any time. And there are good case studies on how this is being done. But the idea is that you create really meaningful feedback, and you actually has a bonus scheme that ties into that.

InfoQ: And then what about that horrible thing the annual review?

Bradley Scott: So annual performance reviews, I am glad to see, are starting to die.. Xero has just moved away from them and we are moving to a culture and a habit of more regular chats about performance- a continuous conversation about what’s working and what’s not. If you can harvest the wisdom of everyone to reward what’s working, you have a system of good performance feedback on a continual basis. That’s a direction that we are trying to head, which is a more ongoing, timely discussion rather than a very big, heavy process that tries to traverse the organization’s memory over a whole year, which is of course inherently patchy. 

InfoQ: One of the things that organizations are going to struggle in there is how do you then deal with the underperformers?

Bradley Scott: In many respects it’s the same problem, hopefully though these types of habits should make it far more transparent. You should be able to tell who is not being recognized by their peers. That’s not to say it’s case in point, but it’s certainly something to look into. I think that the best that you can do is create more transparency of problems. Managing the problems I think is still a difficult process. But certainly in my observation the first challenge is shedding light on problems and making sure that your organization isn’t one where people can hide, and where problems aren’t only dealt with at the end of the year.

InfoQ: Another thing that you spoke about was self organization and you mentioned about some pretty radical things that you are doing at Xero in terms of self organization.

Bradley Scott: I wouldn’t say they are so much radical so much as they actually work. One of our challenges was to grow as an organization to double our team size. We gave that problem to the team and we talked to them about the things that were hard: that we wanted to do it in a way that respected and kept the culture alive and positive, and we wanted to do it in a way that was effective. In the end it’s about delivering value not just spending more money on product development. We challenged the pods (teams) to grow and grow and create another pod, another cross-functional team. Exactly the mechanism whereby you do that is really up to you but that is one of your objectives in addition to writing code

We also had instances where pods self organized and merged, because they had work that was dependent on each other and the best way to get the work done was to rearrange themselves.

In another example we had a very difficult task to get some deliverables completed by the start of tax filing season. We needed all sorts of different skills and it was a hard optimization problem to gather the right team. So we actually threw that to the teams and facilitated a conversation with them to get them to come up with the best way that they could arrange themselves to get the work done within the periods of times that I needed it to be done by. I think we got a better outcome than what we would have gotten as a leadership team. But more importantly, we had buy in and empathy with the problem that they were trying to solve and they understood the constraints and they were on board to try and get the work done within these.

InfoQ: The people owned solving the problem.

Bradley Scott: That’s right. We did ask if it was a valuable process to do again and the consensus was yes, so it might become a more deliberate and regular part of how we do software development.

InfoQ: The other thing that you spoke about was a role that sounded really interesting the Happiness Engineer, what does a Happiness Engineer do?

Bradley Scott: That role has evolved in a really positive way; it was initially a role that we hired to get some momentum around some things that were important in terms of induction and training schemes to help us with the scaling problem that we had. It was either lucky or very fortunate that we hired the person that we did - they had a genuine knack for organizational culture and had a really good background to engineer culture and happiness around the organization. The role has become a very important part of our organization they are the catalyst to make sure that things happen around the office,.And they reinforce not only really good culture but also continual learning. They are also very good at hunting out any concerns that we need to be aware of. For example, people that are having a hard time that we need to give some support to. One of the challenges that we are working through is how do you find more of these people. The right people to deliver that kind of value.

InfoQ: Yes, because that skill set is so nebulous or difficult to define I would imagine.

Bradley Scott: Yes, it’s difficult to define; once you see it you can notice it, once you notice it you see it. It’s a question of how do you screen for that? So if you are bringing someone into your organization, is it possible to even know that they are going to be a right fit? Or is it that you have to create this opportunity internally for people who are really interested in it.

InfoQ: And the last thing that you spoke about, and I will say was you acknowledged that it was a little bit of a rant,  employee engagement surveys, but they’re so important today, aren’t they?

Bradley Scott: Yes, but the way that most organizations seem to do them is bad – certainly the way Xero does it seems to be really unusual. A very long lengthy survey that people have to endure filling out annually, that takes a long time to then process and get results back in any kind of a useful state. Using an analogy to software- you wouldn’t run a software organization like that- big batch sizes, a long feedback time, with an inability to act on the learning that you’ve made. But we seem to do our engagement surveys that way. So I am proposing this needs to be reformed and we need to do our engagement surveys like we do software retrospectives- very regularly, make changes and then watch to see if those changes improve things.

InfoQ: So constant feedback.

Bradley Scott: That’s right, yes. And actually create tools,habits and cultures around the organization that are actually designed for that rather than annual processes. There are products coming on the market I think that are starting to solve this. None of the products available today have completely solved the problem yet, but there’s certainly a lot of innovation happening around that particular employment engagement space as well as cultural innovation support tools generally.

InfoQ: Bradley thank you very much for your time, it’s been great to talk to you and enjoy the rest of the conference.

Bradley Scott: Thank you.

About the Interviewee

Bradley Scott has spent the majority of his career working in fast-growth technology companies. He is currently General Manager of Practice Products at Xero (an NZX-listed cloud accounting software company recently rated the world's most innovative growth company by Forbes magazine) and was formerly GM of Product Development. Before Xero, Bradley held other product management and development roles in the USA and NZ, including at mobile banking and payments company M-Com (now Fiserv).

Rate this Article