Keith McDonald: tech blogger and podcaster

31. Waiting For The Dinosaurs To Leave

31. Waiting For The Dinosaurs To Leave

Dimitar Karaivanov on Agile Atelier, Claire Lew on The Product Experience, Eric Willeke on Agile Amped, Mike Bugembe on The Product Experience, Colleen Esposito on Hired Thought I’d love for you to email me with any comments about the show or any suggestions for podcasts I might want to feature. Email And, if you haven’t done it already, don’t forget to hit the subscribe button, and if you like the show, please tell a friend or co-worker who might be interested. This episode covers the five podcast episodes I found most interesting and wanted to share links to during the two week period starting February 17, 2020. These podcast episodes may have been released much earlier, but this was the fortnight when I started sharing links to them to my social network followers. DIMITAR KARAIVANOV ON AGILE ATELIER The Agile Atelier podcast featured Dimitar Karaivanov with host Rahul Bhattacharya. Dimitar is an expert on scaling Kanban. Dimitar thinks of Agile as a company sport rather than a team sport. At the team level, scaling is horizontal. The more interesting kind of scaling to Dimitar is vertical scaling. If you have a hundred or a thousand teams, the real challenge is the coordination piece on top of those teams and the strategic piece on top of that. If you don’t have an optimized coordination layer that reduces the number of things the organization is working on, your organization is spread too thin. He explained the importance of teamwork and coordination using the metaphor of a band of musicians. Scaling Kanban starts with a single team. What Dimitar likes about Kanban is that if you follow the basic rules, it always results in some kind of improvement. Next, we want to connect the teams to a management layer that performs the coordination activities. People often perceive Kanban as a visual board with some sticky notes on it. Actually, if you go horizontally, then vertically, it is more of an instrumentation facility for your organization. Like a performance profiling tool, you connect Kanban to your organization and it provides entry points with time stamps and starts collecting data. With this profiler, you can dig in and find out what the slowest part of your organization is. Rahul asked about roles in scaled Kanban. Dimitar says there are only two specialized roles called out in Kanban: the service delivery manager and the service request manager. Because one of the principles of Kanban is to start where you are, you do not have to change a lot about roles when you start using Kanban. The service request manager role just means having someone who is responsible for requesting work, such as product manager. The service delivery manager just needs to be someone who is responsible for ensuring the work gets done. This could be a Scrum Master or maybe just a team lead. If the organization is adopting Kanban as a whole, you will need someone on the strategic level that is connected to the Kanban system and has a say in what gets done and when. Rahul asked about failures Dimitar has seen. Dimitar has seen problems in which training just the teams and expecting this to lead to business agility failed. Another route to failure was relying on tools to do all of the work of creating agility. He says you need people with personal agility. You need to find these people or stimulate your existing people to grow themselves so that they become agile in their mindset. Apple Podcasts link: Website link: CLAIRE LEW ON THE PRODUCT EXPERIENCE The Product Experience featured Claire Lew with hosts Randy Silver and Lily Smith. Randy started by asking Claire if she’s ever accidentally been anybody’s worst boss. This was the question that Claire herself had asked at the Business of Software conference in her talk, “The Accidental Bad Manager.” She says that, based on her data, the answer is “probably.” She says that 85% of the time companies are choosing the wrong manager or promoting the wrong people into the role. They’re choosing for a manager those individuals who were showing excellent skills and outcomes as an individual contributor, but those skills don’t transfer over when they become a manager. Claire cited a Gallup study that found that there are five to seven traits that characterize the best managers and, yet, only one in ten managers possesses these traits inherently. Claire created Know Your Team because she herself had a really bad boss and he had no idea. The first thing that made this boss so bad was that he didn’t follow through on his commitments. Looking back, she sees this as a classic case of failing to build trust because he made promises and didn’t deliver. When leaders think about trust, oftentimes their minds go to likability, team-building, and images of trust falls and happy hours. None of those things have to do with real trust, which is the ability to show people that you will do what you say. A second thing that made this boss so bad was that he lacked an ability to communicate and share vision. This is a common problem because most of us get the definition of vision wrong. Claire says that vision is not what you do and it’s not how you do it; it is where you’re going. Vision is the strongest motivating force in a team and the most clarifying force for decision-making. Neither motivation nor decision-making were tenable under her bad boss because the vision wasn’t clear. Lily asked how Claire designed Know Your Team. Claire says that the number of conceptions of leadership is as large as the number of people who have attempted to define the term. She believes the reason there are so many definitions of leadership and the reason that there is no agreed upon best approach to leadership is because, most of the time, the right thing to do is highly dependent on many factors: your own disposition, the team’s disposition, team dynamics, the market, the task at hand, etc. So the best thing to do is to compile as much data as possible and determine the two or three best things to focus on. The best managers, she says, tend to focus on three things. First is trust. Second is honesty. Third is being able to create context in a team, that is, being able to understand and share where you are trying to go and what progress is being made along the way. Lily asked how these areas of focus compare with the traits in the Gallup study Claire mentioned earlier. Claire says that the Gallup study identified temperamental characteristics like positive thinking, good judgment, and empathy, and Claire’s areas of focus represent the skills you can build and the things that you can do to make your team run better. But there are connections between the Gallup characteristics and Claire’s areas of focus: you need empathy to build trust, and you need good judgement to create context. Randy asks why managers are the last to know that they are bad at this. Claire says the psychological reason is that we create a narrative for ourselves that fits with a coherent positive self-image. More practically, we are complicit in being the last to know for several reasons, including the fact that we don’t create an environment for people to tell us. As a result, people don’t speak up in the workplace and this is because of fear and a sense of futility; they believe that nothing would change. To resolve this, we need to be able to ask for feedback in the right way and we have to act on that feedback. To ask for feedback in the right way, we need to be vulnerable. Tell people you are struggling. When you go first and you come from a place of vulnerability, you give the other person permission to be vulnerable themselves and you defuse the element of fear. You also need to be specific. You can’t ask, “How’s it going?” Instead, ask something like, “What is one thing that we could have done better in the past quarter?” or “When is the last time you felt frustrated with your work?” or “Have you observed any micro-managing tendencies from me in the past few months?” or “Have we been all talk and no action on anything lately?” Next, you need to act on the feedback. If asking questions is all about defusing fear, acting on the feedback is all about defusing futility. When you show people that their feedback is not in vain, that helps people to speak up. Some people think this means having to implement every single piece of feedback. Not at all. Acting on feedback can be as simple as thanking someone for their feedback or explaining why you are not doing something. As leaders, we often explain why we are doing something but we forget to share why we are not doing something. The best way to modulate and calibrate the other person’s expectations so that they don’t think speaking up is futile is to say, “You’re not likely to see a ton of progress on this in the beginning but I will give you regular updates on the progress.” And then make sure you give those updates. Another best practice for creating an environment where you are not the last to know is to ask people what their preferences are around feedback. They may want an email, a slack message, or a phone call. Another preference we often forget to ask about is how quickly to give feedback. They may want it right away, or scheduled for the next day or the next week. A third preference is their orientation toward conflict. Do they believe that conflict is healthy and necessary to be productive in a team or do they much prefer a low-conflict environment? A manager should not just be looking to be a great manager or leader but to be the best manager or leader for each particular person and to know that this is going to require customizing your approach to every individual. Randy asked what lessons people can learn about leadership if they don’t have direct reports but need to be able to influence without power. Claire says leadership is not about your title or the number of direct reports you have. At its most core form, leadership is about modeling the behavior that you want to be true of your team. Say you are so annoyed that your entire team is always late for meetings and late on deadlines. Instead of thinking you need to speak to someone or to manage up, one effective way of exhibiting leadership is to turn to yourself and ask, “To what degree can I model the behavior I would like to be true of the team?” A second way to exhibit leadership is to consider how you, as a teammate, can create an environment for those around you to do their best work. Apple Podcasts link:  Website link:  ERIC WILLEKE ON AGILE AMPED The Agile Amped podcast featured Eric Willeke with host Leslie Morse. The first and most critical thing Eric learned about WIP, or work in process, is to pay attention to how WIP cascades and multiplies in an organization. A single piece of strategic WIP equals hundreds to thousands of pieces of individual WIP. A lot of good work comes from corporate strategies, but there is too much of it. Eric gave an example of a VP of product management whose work he helped visualize. They discovered that he had 38 initiatives that he had to report on for his eight teams. When you look at that kind of flood, there is little wonder that we are creating an inability to focus and limit work in process. Eric no longer looks at the executive ranks and says they are to blame. He owns up to it and says that we are all to blame. He now feels empathy for the powerlessness that senior leaders feel in spite of their titles and maybe even because of their titles since those titles carry with them a kind of trap. Eric has three strategies that he uses at organizations to reduce their WIP problems: 1) Start with alignment. Make sure people understand intent and purpose. Eliminate the excess WIP that comes from the “Am I in the right direction?” question. 2) Practice reduction in depth. According to Michael Porter, the essence of a good strategy is what you’re not doing. Help people learn what is not part of the strategy and generate focus. You may have to repeat yourself because, as Patrick Lencioni says, “You only get one message per quarter and you need to say that message hundreds of times.” 3) Create permission and safety as part of how you decentralize. When you decentralize, people need to have all the permission to take responsibility and the safety to try things, learn, and experience the associated failures that come with learning. The conversation with leaders to get them to limit WIP is difficult. The leader starts with the best of intentions. If you come in too strongly with a message that they are doing it wrong, you are saying, “You were trying really hard in the best way you know how and you failed.” They didn’t. They were not responsible, necessarily, for all of the different pieces of WIP or how it cascaded, yet they have to take responsibility for helping people set things down. One leader Eric is working with understands this and uses a quarterly message that says, “You may put things down, but you need to put them down gently.” A lot of people look at WIP and say, “We just need to throw away half the items in process.” But that hurts people, hurts initiatives, and hurts business leaders. So we need to know how to carefully set down the things we’re not going to do yet and bring everybody else along. Apple Podcasts link: Website link: MIKE BUGEMBE ON THE PRODUCT EXPERIENCE The Product Experience podcast featured Mike Bugembe with hosts Randy Silver and Lily Smith. Mike is the former Chief Analytics Officer for, which uses machine learning to try to increase generosity in the UK. When Mike joined, the company had in excess of ten years of data on people raising and giving money for causes they cared about. They used their technology to get a signal about what people were passionate about and used this to match people to causes. They started by using a collaborative filter approach like Amazon’s “people who purchased your item also tend to purchase these items”, but charitable giving is so personal that collaborative filtering doesn’t work. Instead, using Nicholas Christakis’ research, they could see that connections between individuals could help them understand the flow of generosity and the flow of the things that are important to people. Mike says that a lot of large companies talk about the fancy things they do with machine learning and data science like Facebook’s EdgeRank or Amazon’s and Netflix’s recommendation engines, but sometimes there are use cases that are unsexy but deliver a huge amount of value. For example, when people put a fundraising page on JustGiving, they have the option to specify a target. Only 30% of fundraisers were specifying such a target, but Mike found that this behavior led to much more money raised. So Mike created a machine learning system that predicted how much a fundraiser was likely to raise and pre-populated the target field. This was a lot of work to deliver one number on a screen, but this feature delivered an additional 7% on a 400 million pound business.  His approach to understanding where AI can deliver business value is to look at every business as a system of people making decisions, whether it’s marketers, product teams, or users. When you look at a product this way, the machine learning use cases float to the surface. You see where machine learning can make a decision more efficient, more automated, or more predictable. You then add a metric to each decision and see how decisions relate to each other or how they relate to key metrics you are trying to move. Your data is quite unique to your business and your product. It acts like a fingerprint. One of the risks of data science is that it is an experiment every time you do it. Even if somebody else has done it before, you have no guarantee that when you do it you will get a successful result. Product management teams that work with data science teams need to be aware that data science is not the same as delivering a feature with a software development team. It is an experiment. You have a question in mind and you have no idea whether or not the research will produce the result you’re expecting. Lily asked Mike how he recommends people hire a data scientist. Mike says he is very much against the idea of hiring a data scientist just because they have a PhD. That’s a massive risk. You could get a PhD-holding job candidate who only understands regression and is not numerate enough to try a lot of the different algorithms that data scientists use. Mike himself looks for data scientists who have real life experience. This doesn’t mean they’ve worked in a lot of companies before. It means they’ve got things that they’ve produced. You can read and study, he says, but if you’ve never done it, you won’t know the gotchas and foibles that come with working with data. Randy asked if there any guidelines or cheat sheets for people to educate themselves about bias in data collection, in algorithms, in assumptions, and in interpretation. Mike created a non-technical course for executives, product managers, and founders because they know their business better than the data scientists, in most cases. They add a layer of domain knowledge that helps reduce risk due to bias. Apple Podcasts link: Website link: COLLEEN ESPOSITO ON HIRED THOUGHT The Hired Thought podcast featured Colleen Esposito with host Ben Mosior. Colleen loves helping teams get started, helping them understand that Agile is about uncovering better ways of developing software, and helping them identify what Agile means for them. She also loves helping the managers, directors, and VPs to interact better with the teams so that the teams become empowered. Colleen is a fan of invitation-based coaching and making sure the people understand the whys behind the change, what it looks like in the anticipated end state, and what steps the team may take to move towards that vision they’ve identified. Colleen says that the first thing she does when she comes into a brand new organization is she tries to understand the whys behind their decisions. “Why did you choose to use Agile?” If what she hears is, “Twice the work in half the time,” then she knows that she might have to reset expectations. Before starting an Agile adoption, Colleen gets everyone to think about the answers to some common questions like, “Why are we doing this? What’s in our way? What’s in our favor?” She says that if the leadership makes changes without involvement of the people, they are going to miss out on a valuable perspective. Colleen says that the people who hire her often think that she is going to come into their organization and make huge, sweeping changes. Instead, in the very beginning, it is often small changes like connecting what a development team is doing with what an operations team is doing. Apple Podcasts link: Website link: LINKS Ask questions, make comments, and let your voice be heard by emailing Twitter: LinkedIn: Facebook: Instagram: YouTube: Website: Intro/outro music: "waste time" by Vincent Augustus

Duration: 21 min

Release Date:

Share part or all of the audio of this episode