BACK TO Business Resilience
ZINNOV PODCAST | Business Resilience
|
Can globally distributed software teams truly unlock their full potential through autonomy and collaboration? Peter George, Executive Vice President and Chief Technology Officer of HMH, joins Nilesh Thakker, President of Zinnov, to share his battle-tested insights on striking this delicate balance.
Peter dives deep into the transformative power of cutting-edge AI tools like Co-pilot and CodeGPT, exploring how they can augment human ingenuity in software engineering and content creation processes.
Beyond technology, Peter unveils the secrets to nurturing cohesive, high-performing global teams that thrive on shared vision, empowerment, and a culture of excellence. He dissects the intricate dynamics of distributed workforces, offering a treasure trove of actionable strategies for leaders, managers, and team members alike.
Don’t miss this invaluable episode packed with game-changing insights! Tune in now to revolutionize your approach to managing globally distributed software teams and harness the full potential of AI in your organization.
PODCAST SUMMARY
Nilesh: Welcome everyone to a new episode of the Business Resilience Series of the Zinnov Podcast. I’m your host today, Nilesh Thakker, the President of Zinnov. Today, we are delighted to have Peter George, the Chief Technology Officer at Houghton Milford Harcourt, also known as HMH, a leading provider of educational content, technology, and services. Peter brings over 30 years of experience in the software and technology field and is responsible for leading HMH’s technology vision and execution. Welcome, Peter.
Peter: Thank you, Nilesh. It’s a pleasure to be here.
Nilesh: Let me begin. Peter, you’ve been establishing global teams, particularly in places like India. You’ve mentioned multiple teams in Pune over the past few decades. How do you ensure global integration across teams worldwide, enabling them to work together as a cohesive unit?
Peter: That’s an excellent question. Over the years, I have indeed had the challenge of managing globally distributed teams tasked with collaborating on integrated solutions while being geographically dispersed worldwide, both in India and other countries. My philosophy revolves around establishing teams as equal peers with clear, independent missions, equipped with the necessary resources for self-sufficiency and autonomy. However, they must also acknowledge the need to collaborate with other teams to ultimately deliver an integrated solution. The initial step involves making the teams as self-sufficient as possible, after which mechanisms are implemented to facilitate collaboration across teams, operating at multiple levels.
Nilesh: That’s insightful. I understand your team in India was recognized as one of the best setups within a year. Now, a common concern among leaders is the difficulty in measuring productivity, especially in software development, where there is no universally agreed-upon metric. How do you approach this challenge with globally distributed teams? First, how do you think about productivity from a software perspective, and then, how have you measured it?
Peter: You raise an excellent point. Measuring productivity has always been a tricky endeavor, primarily because an excessive focus on a single metric can inadvertently lead to undesirable consequences on other crucial metrics. If individuals are aware that a particular aspect is being intensely measured, they may neglect other essential areas. Additionally, identifying a universal metric to quantify productivity in the software industry has remained an elusive goal. Our approach involves utilizing the velocity of story points as a productivity indicator for individual teams. We closely monitor each team’s progress and evolution over time, aiming for a consistent increase in velocity. If any breakdowns occur, we investigate the underlying reasons. However, to mitigate the risk of unintended side effects, we incorporate this metric into a more balanced scorecard, considering a range of metrics for each team. By analyzing the combination of these metrics, we gain a more comprehensive understanding of the situation and ensure we’re not overly indexing in one dimension at the expense of others.
Nilesh: How do you ensure that the teams are fully productive quickly? I think you mentioned some of that earlier, but are there any specific strategies that have worked well for you? Your accomplishments in Pune have been impressive, so please share some tips beyond the governance aspect you discussed.
Peter: Certainly. I firmly believe that the most crucial aspect is establishing a robust partnership between the former team and the new team, as this handoff is critically important for a seamless transition. Our process begins by assembling a new team that we deem capable of undertaking the work, providing them with all the necessary skills – front-end and back-end engineers, scrum masters, product owners, and everything they need to thrive. Once we have the majority of the team or key leadership in place, we initiate the knowledge transfer process. We engage the former team to provide comprehensive training, introduce the new team to the project’s intricacies, and ideally, facilitate face-to-face interactions. In my experience, there is no substitute for in-person meetings between the leadership teams to discuss challenges, build trust, and facilitate effective knowledge transfer.
Another crucial aspect is ensuring the former team feels comfortable handing over their “baby” to the new team, as they need to trust that the transition will be handled smoothly. Once we have these essential ingredients in place – a new team established, a partnership formed, and knowledge transfer underway – we move into the operational details. We implement a model where we gradually ramp down the work for the former team while simultaneously ramping up the new team’s workload. The new team might start with smaller projects while the current team handles the remaining tasks. Ultimately, we shift the balance until the new team becomes fully self-sufficient and takes full ownership.
This process can be time-consuming, and we’ve employed tactics like utilizing contractors to bridge knowledge gaps and accelerate the transition. However, once we build a strong partnership between the former and new owners, align everyone on our goals, and establish strong relationships, the process typically proceeds reasonably smoothly.
Nilesh: Understood. Now, everyone is talking about AI, so I have to ask – have you used any new AI tools for boosting productivity? What has been your experience, or is it still too early?
Peter: AI is indeed a broad and highly discussed topic these days. Let me address it in two parts. First, we’ve been leveraging AI technology, particularly machine learning, at HMH for an extended period, and we’ve witnessed tremendous results. Specifically, we’ve employed machine learning to entirely replace some manual, high-volume content categorization work that was previously part of our solutions, and the savings in those areas have been extraordinary. We’ve essentially eliminated highly manual, repetitive processes by automating them through AI solutions.
As for generative AI capabilities, the current hot topic, I would say we’re still in the early stages of our experience with that. While we are actively utilizing generative AI technologies, and we’ve received excellent qualitative feedback that they’re very helpful, we haven’t yet reached the point where we have quantitative analyses demonstrating a specific boost in productivity, for instance, a 15% or 25% increase. To give you some examples, we’ve implemented Microsoft Co-pilot for a couple of hundred HMH employees to aid with general productivity, rolled out CodeGPT to all our engineers to enhance their productivity, and introduced other generative AI tools to our content developers to assist them in their authoring efforts. The initial qualitative feedback from everyone has been enthusiastic and positive.
However, the key question remains – what does that translate to quantitatively? We expect to conduct before-and-after analyses of velocity within individual teams using these tools to assess the impact accurately. Unfortunately, I don’t have those concrete numbers to share with you yet, but perhaps we could revisit this topic in a future conversation or podcast once we have more comprehensive data.
Nilesh: It seems like the early indications from the people using these AI tools are positive.
Peter: Yes, absolutely. And our emphasis, like many companies and individuals, is on utilizing AI to augment human productivity, not replace people entirely. The idea is for our engineers, authors, and others to leverage these tools to jumpstart their work, accelerate progress, and generate tests or other outputs more efficiently. However, at the end of the day, they still need to take personal responsibility for the ultimate results and ensure their quality and accuracy.