Tech product management: Incorporating privacy, sustainability, and inclusion
In this episode of the McKinsey on Start-ups podcast, McKinsey executive editor Daniel Eisenberg speaks with McKinsey partner Martin Harrysson and associate partner Rikki Singh about responsible product management in the technology sector. An edited transcript of their conversation, which took place earlier this year, follows. To hear more episodes of McKinsey on Start-ups, subscribe on Apple, Audible, Google Podcasts, Spotify, or Stitcher.
Daniel Eisenberg: Once primarily focused on execution and on-time delivery, software product manager roles have been heavily transformed over the past decade or two. They are expected to make critical decisions and act as the glue that binds the many functions that touch a product, from engineering, design, and customer success, to sales, marketing, operations, and finance. And now, amid growing societal concerns about responsible stewardship, product managers in established tech companies and start-ups are beginning to incorporate privacy, sustainability, and inclusion into their already complex process. Customers, investors, and regulators increasingly demand this shift to responsible product management.
A recent McKinsey article
examined this increasingly important issue. Today we’ll be speaking to two co-authors, partner Martin Harrysson and associate partner Rikki Singh, who are based in the firm’s Bay Area office. The article and underlying research are just one part of McKinsey’s ongoing focus on tech product management, including previous articles on the evolving PM role
and the associated hiring challenges.
McKinsey has also recently rolled out its Product Academy, an open program featuring content from leading-edge product practitioners.
And now, let’s get to our conversation with Martin Harrysson and Rikki Singh of McKinsey.
Martin, Rikki, thanks so much for joining us. Martin, let’s start by defining responsible product management.
Martin Harrysson: Product management has experienced at least two significant waves of change in the last couple of decades. The first wave was the shift from on-prem infrastructure to cloud, which allowed product managers to shift from requirement-getters to product visionaries developing minimum viable products.
The second wave was driven by the consumerization of technology, which led PMs to be more anchored in design thinking, make data-driven product decisions, and respond to customer obsessions.
Right now, we think we’re on the cusp of a third wave that is about to start, with product managers beginning to incorporate inclusion, privacy, and sustainability into their products. That is how we define responsible product management and responsible innovation as a whole.
Daniel Eisenberg: What developments are fueling this third wave of product management? Is it primarily financial? Other McKinsey research has previously looked at how strong performance on ESG issues
can correlate with higher shareholder value.
Martin Harrysson: It’s a good question. These topics are certainly not new, but the COVID-19 pandemic has accentuated their importance. People have become even more socially aware. Organizations have very vocal employees, and there is increasing regulation. These things are accelerating the wave of change we’ve seen over the past couple of years.
Of course, the tech sector provides many tools and systems that companies and individuals use daily, so it’s very much at the forefront of these changes. And as you mentioned, McKinsey research has found that over 80 percent of C-suite executives strongly believe that performance on ESG issues is correlated with higher stakeholder value. So that is certainly another driver.
Daniel Eisenberg: And when you were thinking about examining responsible innovation, what drove the decision to focus on the role of product managers in particular?
Martin Harrysson: When it comes to developing products, the product management role is like the center of a spiderweb. They decide what gets built, and they incorporate input from all kinds of different stakeholders. They are at a natural point when it comes to thinking about responsibility. And so, when PMs think about building a good experience for the broader set of stakeholders, we believe that responsible product management will soon not just be a nice to have but will be a need to have.
Daniel Eisenberg: Rikki, tell us about the research you, Martin, and the whole team did in this area. What were the main goals?
Rikki Singh: When we first started looking into this space, our goal was to identify the common practices. We weren’t sure about the maturity levels for these responsible innovation dimensions, so it was very exploratory. We started by interviewing industry leaders, which allowed us to test our hypothesis on the maturity question. We realized quickly that the leading tech companies adding prominent environmental, social, and governance [ESG] job roles might be investing in some frameworks, but it is only happening in pockets. Once we realized that, we made an initial perspective and started hosting panels with product managers to understand better how these frameworks are being adopted at grassroots levels. And then finally, we ran a survey to quantify all the trends and frameworks we had heard about.
Daniel Eisenberg: What were the primary findings about how the three dimensions of responsible product management—privacy, sustainability, and inclusion—are being prioritized these days?
Rikki Singh: It was no surprise that performance came first as a strategic priority. What did surprise us was that privacy and sustainability surpassed usability in terms of strategic priorities. The other unexpected thing was that inclusion ranked the lowest despite how visible its impact is.
Daniel Eisenberg: Let’s dive deeper into each of those three dimensions, starting with privacy. What were the biggest challenges you found in incorporating privacy into product management?
Rikki Singh: Privacy as a dimension was the most mature. Privacy is most challenging because it has a trade-off against usability. Most people love products that offer customized recommendations or remember our introductory information, and that data is often needed to drive more usability. But we also don’t feel comfortable letting the product keep all that data. Increasingly, with GDPR and the focus on privacy, it seems organizations are now trying to prioritize privacy over usability.
Daniel Eisenberg: Can you talk about the frameworks that some companies are starting to use to improve in this area?
Rikki Singh: At the nascent stage, it’s fundamentally about doing data audits and understanding what customer data is being stored and its retention policy. What’s beginning to happen is that some standardized frameworks are now coming to the fore.
Three-quarters of product managers said they employ differential privacy frameworks, while 40 percent said they also use federated learning approaches. Both of these are more prominent when building an intelligent algorithm. In the case of differential privacy, that’s using stake tracking when pulling personalized information, and with federated learning that’s running on the device instead of on the server. Those are the two most common frameworks that we encountered in our research.
Daniel Eisenberg: And judging by the percentages you quoted, privacy may not be ubiquitous, but it’s starting to become quite prevalent as an issue that companies are thinking about.
Rikki Singh: Absolutely. I think it started with GDPR. People consider GDPR to be the norm when measuring their privacy progress. But they don’t equate that with success; they want to do more than that, which is pretty inspiring.
Daniel Eisenberg: Martin, let’s talk about sustainability. What are the key factors motivating companies to incorporate sustainability into product management?
Martin Harrysson: Well, from a basic business point of view, it’s become more apparent that customers are willing to pay for goods and services made sustainably. This is an overall trend we see, and PMs have certainly noted it. We saw from the research that among those who ranked sustainability practices high on their list, almost 60 percent cited consumer demand as the top reason. About half pointed to the ability to attract capital as a really important factor; a third point was regulatory pressures. The combination of these three factors has driven the motivation to look at sustainability.
Daniel Eisenberg: What barriers are there still in making sustainability a core dimension in the product management process?
Martin Harrysson: Two things came up consistently. The top one was, maybe not surprisingly, a lack of capability. These are still relatively new areas; even though they’re recognized, knowing what to do about them is still unclear. The second one was a standard definition of what good, sustainable design really looks like.
Daniel Eisenberg: You also found that either very few or no respondents have gone beyond using greenhouse gas protocol metrics to monitor carbon output. Can you talk briefly about those metrics and why they have been limited to that so far?
Martin Harrysson: That’s right. Very few organizations have mapped out clear policies and incentives related to sustainable design. Some respondents used scope one and scope two to measure carbon output. Going beyond that means that organizations wouldn’t just look at the direct emissions from the thing that they are working with, they’d also look at the secondary and tertiary effects, and that’s what’s difficult for PMs to do these days.
We think this is a missed opportunity, especially as recent regulatory news from the SEC has shown that it will become even more important to account for emissions throughout the life cycle.
A few companies in the tech space have now started looking at this. For example, SAP has begun incorporating carbon footprints into its software and financial accounting products. And Google has recently rolled out a comparison feature in its Maps app to pick the best route based on carbon impact. So, we are starting to see things happening, but it’s still pretty early days.
Daniel Eisenberg: Are other sectors already using this kind of life cycle assessment accounting? Is that something that other early movers are using?
Martin Harrysson: It’s debatable which one is more advanced, but consumer products have been thinking about this for a longer time, so I think that is an interesting space for tech companies.
Daniel Eisenberg: Rikki, let’s move on to the third dimension, inclusion. Can you give some examples of how persistent bias, whether race or gender, currently impacts how products in tech work or don’t work, for that matter?
Rikki Singh: There’s been much discussion around racial and gender inequity in recent years. However, even if you did an audit of products, apps, and hardware that you have out there and tried to understand if it’s all equitable, the answer is probably still no. For example, we know that if you look at smartphone biometrics and cameras, they struggle to perceive and render the skin tones of non-white individuals.
Skewed data in payments and banking products can result in Black applicants being denied credit at a higher rate than White applicants. And then, individuals with visual impairments or mobility constraints sometimes have to wait longer even to get access to products that have been released.
A lot of the bias that we see in algorithms, such as with financial products, can be attributed to the fact that they are typically trained on historical data that is often biased.
Daniel Eisenberg: You mentioned earlier that inclusion had been rated the lowest as a top strategic priority. Only 17 percent of managers cited it as such, compared with around 37 or 38 percent for privacy and sustainability. What is the primary reason it’s so challenging compared to the other two?
Rikki Singh: Product managers stated the primary reasons are limited tools, subjective metrics, and no direct link to performance. Underlying all these challenges is the inherent subjectivity of defining success in the inclusion dimension.
So, for example, companies could say the bar on being inclusive is having the application meet ADA compliance standards. It’s accessible, and a vendor’s narrator could be used to navigate the entire app. Or they may say they want to go further and look at being more than just accessibility compliant. The fact that there is a gradient scale, with good being hard to define on that scale, makes it challenging to address this dimension as a whole. Of course, the metrics used to assess whether a product is 100 percent inclusive or 50 percent inclusive don’t exist.
The final thing to say is that you can potentially expand your consumer base by being inclusive. So, drawing that direct linkage to performance metrics is critical to incentivizing PMs and holding them accountable for prioritizing the inclusion dimension.
Daniel Eisenberg: As you pointed out, inclusion is the most subjective in measuring success. That’s an incredibly complex issue to hand off to product managers if they’re not getting enough tools and guidance.
Rikki Singh: It’s also a risky one. Imagine going out and saying, “I’ve come up with a scale, and I’m 30 percent inclusive.” Nobody would want to say that. So, there has to be some sensitivity around defining how to measure and report inclusion.
Martin’s experienced this himself on the sustainability dimension front. Only when companies started measuring sustainability and seeing movement on the green line did everyone begin to adopt that dimension sincerely. Something similar has to happen on inclusion to get the recognition it deserves.
Daniel Eisenberg: In the article, you referenced some pioneering approaches that a few companies are starting to use in this area. Can you talk about them?
Rikki Singh: There are two aspects on inclusion. The first is how to do more inclusive design. That’s around shifting left the thinking around inclusion at the beginning of the product design stage to ensure that consideration is given to all the different types of user bases. Some popular methods came up in our research, such as Perspective Hats and Another Lens. Another Lens is the framework that Airbnb has put out there. It’s a tool with these cards, each of which addresses a point that allows you to balance your bias, consider the opposite perspective, and then embrace a growth mindset. It’s a tool that lets designers and PMs think differently upfront about users and their needs.
The second aspect concerns the historical data set driving bias in what a smart algorithm is learning. Microsoft and Google have each developed publicly available AI playbooks to give development teams guidelines on ethical and responsible design.
Daniel Eisenberg: On the topic of measuring success, is there an industry body or group of different tech companies pooling their ideas on measuring something like this? Or do you think one company will figure it out, make it publicly available, and then it’ll get more widely adopted?
Rikki Singh: It’s interesting. There has been inherent bias in many products we’ve built historically. AI, ML, or intelligent algorithms showed us a meter that the bias existed because they learned from the data and then reflected the bias in the system back to us. With that, companies started realizing that if they want to deploy intelligent algorithms at scale, they have to consider de-biasing them. On that front, I think that the larger data sets there are in-house, and the more companies can think about their bias, the more successful they’ll be at developing potential ways to de-bias.
On the design front, I think there’ll be experimentation regarding what different companies are trying. Setting the metric is where people will need to come together because metrics are typically only adopted when they become industry standard.
The strongest parallel I can draw is with sustainability. Until we had Scope 1 and Scope 2 measurements defined as the industry standard on how people would report, it was tough to know how best to measure carbon emissions.
Daniel Eisenberg: Martin, we touched on how finding qualified candidates with these varied skills and backgrounds was already proving difficult. What steps can organizations take to help product managers succeed at responsible product management?
Martin Harrysson: While this is difficult, we also think there are several simple steps that organizations can start taking today that will have a real impact. This is not something that leaders should sit back and wait on.
Elevating one of these dimensions, or more, as a strategic priority will make a huge difference. This means having executives focus on these areas in product reviews as a top-level outcome metric. That’s really where it has to start.
The next thing is to update the approach for selecting and prioritizing new products and features and how that gets incorporated into the product development life cycle. Prioritization metrics today are mostly around how customers will like the product or service, how the organization will drive adoption and engagement, and ultimately how it will monetize. Adding criteria around some of the topics we’ve discussed can have a significant impact.
The third piece is to start rolling out some training and frameworks. There are standardized frameworks and artifacts that we think can have much impact relatively quickly; for example, some tech companies use simple templates around where their data comes from and what kind of machine learning models they use. This is a major first step.
And then the final thing that I would mention is to consider exploring some of the off-the-shelf tools that can be adopted today, such as those that assess federated machine learning techniques. This will help product managers and their teams save time and effort instead of developing these tools from scratch.
Daniel Eisenberg: Martin, the article mentioned the importance of integrating responsible design criteria directly into the product development life cycle, starting at the lab phase. Why is that so critical?
Martin Harrysson: We believe that organizations may want to start thinking about this from the outset when they begin developing the product to ensure bias doesn’t creep into products. If you don’t think about them from the beginning, it gets hard to see them later in the process.
Rikki Singh: No engineer likes downstream churn. So, if a PM detects it later and wants to go back and talk to an engineer about how the requirement changed, I don’t think that’s a fun conversation to have.
Daniel Eisenberg: You talked about the lack of a definition of what good means in sustainable design. This may also have been an issue in inclusion and privacy. How do you think companies and the industry can work to address that going forward?
Martin Harrysson: Having companies come together and agree on what good looks like is one approach. I also think that as we get better and better at measuring things, we get more sophisticated. We saw this in sustainability when initially it was hard even to measure Scope 1, but we started installing more sensors on things we made, and the more things spread out in the supply chain, the more data is produced. This evolves the way we can measure something, making these measurements of what good looks less ambiguous. If history’s any indication, there are reasons to believe that we will get more sophisticated and measure more data around these topics. That will help us to figure out what good might look like.
Daniel Eisenberg: We discussed shareholder value and performance on ESG issues at the beginning. Is cost an obstacle for companies contemplating adding these dimensions to the product management process?
Martin Harrysson: It’s not so much about cost. We think it is more about capabilities and how to incorporate these dimensions. Many of the things that we talked about today can have a positive cost impact. When you make things more efficient in a broader sense, it usually becomes less costly. So, while it might require a little bit more thinking and time upfront, I think most would say that the cost barrier is not of primary importance. Rikki, would you agree?
Rikki Singh: Yes. The fact that we asked PMs and leaders to prioritize a dimension, and they prioritized performance, followed immediately by privacy and sustainability, is a strong marker of how they think as they weigh the impact versus the effort.
Daniel Eisenberg: I wanted to ask both of you to take a slightly broader view of the future for responsible product management. How far do you realistically hope and expect responsible product management to be in five to ten years from now? Can you envision a day when responsibility dimensions rank as crucial as performance at the top of the charts?
Martin Harrysson: I would turn that around. If you look at where the world is heading, it’s hard to imagine how it would not become as important as performance and usability. It’s like speculating on any new technology, whether self-driving cars or quantum computing. Getting the exact timing right is very hard. Does it take five or ten years? It’s hard to assess. I think it’s probably closer to five years than ten. But that’s perhaps an optimistic view.
Rikki Singh: Product managers love solving challenging problems. They cracked the one on customer experience and how to build products that customers fall in love with. So, they will be up to the challenge of figuring these out next. Given that they’re already thinking about it, prioritizing them in pockets, I can see them finding ways to scale this.
Companies may want to think about how active a role they want to take in enabling their product managers because the PMs are on that journey, regardless of whether the organization is with them or not.
Daniel Eisenberg: And it would seem logical then that the companies who take more of an active role may get to the answer faster and potentially reap the rewards for being ahead of the curve.
Rikki Singh: Indeed, amid the continuing war for talent, it may help attract some of the best talent. Because people want to work for companies that have a mission, that invest in responsible practices.
Daniel Eisenberg: That’s a good point. I know McKinsey’s done some research recently about the importance of purpose and meaning at work
and how critical that is in talent recruitment these days. I can see that being part of the PM value proposition in terms of who’s seriously addressing these issues and making them a vital part of the process.
Rikki Singh: At the end of the day, all product managers want to build products they can be proud of.
Daniel Eisenberg: Well, Martin and Rikki, thank you for taking the time to talk about responsible product management.
Martin Harrysson: Thank you very much.
Rikki Singh: Thank you.