"At the end of the day, people are not there forever, but the culture is. And the culture describes the company’s values, and those values describe the skills and behaviors of everybody in the building." - Gib Biddle
Show Notes
The theme of this week’s Product Thinking Podcast is culture, and Melissa Perri’s guest is Gibson Biddle, author of the “Ask Gib” product newsletter. Gibson is dedicated to teaching product strategy and culture both in and outside of the classroom. He joins Melissa to discuss the importance of having a good organizational culture, and how that affects strategy.
Here are some key points you’ll hear Melissa and Gibson talk about in this episode:
The product leader’s job is to delight customers in hard-to-copy, margin-enhancing ways (DHM).
How Netflix developed their high-level strategy for personalization.
Gibson’s experience with managing the product team as VP of Product Management at Netflix. Every product leader who worked for him was expected to have clear, defined strategies for their teams.
The failure of Project Griffin.
Building a culture where everyone is on the same page about making the best choices for the business is difficult, but not impossible. One key practice is context over control; ensure that everyone understands the context that the business is operating in.
“At the end of the day, people are not there forever but the culture is; the culture describes the [company’s] values and the values describe the skills and behaviors of everyone in the building,” Gibson tells Melissa. “At our quarterly meetings, we would discuss strategy, but we would also learn to form good judgment. This is what the culture is about; helping individuals make great decisions about products.”
"Culture is all about who you hire… who you promote- every time someone was promoted to director or VP there was a celebration because it was about them being a culture carrier, them living the Netflix culture. And [when you would] let people go, they might demonstrate amazing results, but they were living outside of Netflix's values,” Gibson shares.
Strategy is important when starting a product organization, but it’s ok if half of it fails. It’s a process and a bunch of hypotheses that you need to learn from.
Resources Gibson Biddle on LinkedIn | Twitter GibsonBiddle.com
Episode Transcript
Intro Music & Voiceover: Creating great products isn't just about product managers and their day-to-day interactions with developers. It's about how an organization supports products as a whole. The system, the processes, and cultures in place that help companies deliver value to their customers. With the help of some boundary-pushing guests and inspiration from your most pressing product questions, we'll dive into this system from every angle and help you think like a great product leader. This is the Product Thinking Podcast. Here's your host, Melissa Perri.
Host: Welcome to the show, Gib. I'm so excited to have you on.
Guest: Hello, Melissa. Thank you for having me.
Host: I have been waiting for this for a long time because when I was writing Escaping the Build Trap, I read so many of your things about product strategy. I was really obsessed with learning more about how you did it at Netflix, and I actually used it as a case study in the strategy section of the book. So I'm really excited to dive into that more with you. But to get our listeners up to speed with who you are, can you introduce yourself a little bit and tell them how you got to be this amazing person at product strategy and all the wonderful stuff you've done?
Guest: Sure. Thank you for flattering me. I'm not sure if it's warranted. You know, at the highest level, I joined Netflix in 2005 as the VP of Product. In 2010, I joined Chegg, which is a textbook rental and homework help service, as the CPO. And then five years after that, I committed myself to the life which I live today, which is very flexible, but I spend a lot of time focused on teaching outside the classroom. I focus a lot on product strategy, on culture, on helping product managers all over the world.
Host: That's great. And a lot of time on skiing, too.
Guest: It is true that one of my metrics is I try to ski my age in days. So this year I will be 59 on March 28th, and I know I won't make my goal this year because COVID really has slowed me down on that front.
Host: Oh, that's a shame. But happy early birthday.
Guest: Well, thank you.
Host: Well, hopefully next year you'll get to do all 60, and that would be cool.
Guest: Very good. Yeah.
Host: Yeah. So, I want to talk to you a little bit about strategy. When I was reading about Netflix, I thought that the strategy you presented and the way that you really communicated their progress, right, from getting big on DVDs to leading streaming, to expanding internationally, and then going into original content, it makes so much sense to me, right?
Like I'm reading that going, what, wow, yes, that is a strategy. Like that is a clear-cut strategy. And you've come up with so many of these frameworks, you know, with wonderful acronyms like GEM and GLEE and DHM to really define your product strategy. Can you tell us a little bit more about how do you think about product strategy?
Like what does that mean to you? And what are some of the strategies that you advocate for?
Guest: Sure. So it is true, as a teacher, I'm always thinking about tools and models and frameworks that give folks a way to think and talk about stuff. And as companies get bigger, we're full of humans, and communicating clearly with each other becomes increasingly hard, which is why I use these tools, models, and frameworks.
Probably the most important one for me is this notion of DHM, which is a product leader's job is to delight customers in hard-to-copy, margin-enhancing ways. When I approach new products or companies, I'm usually thinking about that first. So to give you an example, at Netflix, one of the key strategies was personalization.
And the theory, which we proved out eventually, was that personalization would delight customers. It would make it easier for them to find movies they love. It would be wonderfully hard to copy. So think about the tools, the technology, the algorithms that they're using with 200 million members today.
And then the third component is margin enhancement. And that's simply create a profit, create a business that works. So at Netflix, the retention, in the old days, it was 10% canceled every month. Today, 2% cancel every month. So that's a huge improvement in the business. The lifetime value went from 100 bucks to almost $300 a day.
The other thing that's cool about personalization is it enables Netflix to be really smart about how they spend their money on content. So for instance, they'll predict that Stranger Things will get 100 million people watching it, and they'll spend 500 million bucks on that. They'll predict that 20 million people will watch BoJack Horseman, like me, I'm a freak.
And because of that, they'll spend 100 million bucks on that. And then they'll predict that 1 million people like me will watch Everest climbing documentaries, and they'll spend 5 million on that. And I call that right-sizing, which is enabled by personalization, but it's one of the ways that Netflix has managed to make a business that works.
So I just gave you an example of how I use the DHM model, the notion of delighting customers in hard-to-copy, margin-enhancing ways, plays out when you think about developing a product strategy.
Host: So when I start thinking through product strategies and when I come into organizations to help deploy this, there's usually not a product strategy. There's like this lofty vision of make more money or transform into this company. And then there's a bunch of stuff on the bottom that's with the teams where they're like executing on features.
We usually have to start from scratch and we have to figure out what's the right level of product strategy for each part of the organization, right? Like what is the CEO going to be setting or the C-suite? What is the product leadership team set? What does the product director set? What do the teams set?
How do your frameworks kind of fit into, like which level do they kindically fall in when you're thinking about delighting people through hard-to-copy, margin-enhancing ways? Like if you're putting that guideline in, who's coming up with that? And how does that get translated into what you're going to do inside the company?
Guest: Yeah, I'll answer the question sort of three ways. So first, I want to describe how I think about a product vision, which is really the closest thing to helping define a company strategy. I'll give you an example of how I apply that DHM model, for instance, at my level, and I'm always the product leader.
But then I'll demonstrate what I expect of product leaders who work for me in their work to define the product strategy in their swim lane. So I'll start at the top. This other model is called the GLEE model. And I use it because I want product leaders to think long-term, just for a moment. So for instance, at Netflix, the way I articulated the GLEE model, the G stands for, what are you going to get big on initially?
The L sort of says, hey, five or 10 years after, what's that next thing that you want to lead? And then five or 10 years after that, how might you expand even further? And that's why it spells GLEE. The idea at Netflix, we were going to get big on DVD. At some point, we were going to lead streaming.
And then once we were a digital service, then and only then, we would expand internationally. And then you can see this continues forever. There's another E, expand further into original content. I mean, I can even articulate a guess on another E, five or 10 years in the future, which might be interactive stories, for instance.
But that GLEE model, what I'm trying to do is get people to think in waves. If you can ride a wave, good things happen. So the first wave that Netflix rode was essentially the DVD player was born. The second was the growth of video on the internet. That was the streaming. The internationalization, the idea that you can push a button and all over the world, you can watch the same film that's been dubbed into 40 different languages.
And then, you know, in perfect faith, these are always hypotheses. At Netflix, we experimented with an original content strategy, if you will, in circa 2007. It failed in a DVD era, but it worked extraordinarily well in 2013 with the launch of House of Cards. Anyways, that GLEE model is my way of encouraging the product leader and the CEO and the whole company to think big about how the work that they do will play out over 10 or 20 years.
You know, people tend to look down at their toes and think very incrementally. And this GLEE model just nicely encourages people to think big and to think long-term. I usually will craft that. I'll typically get the CEO and the exec team to buy into that. And then when it comes to articulating the product strategy, that's for me.
So I shared at a high level that one of the product strategies at Netflix was personalization. And there were other theories and hypotheses. We had a social strategy. You know, we had a theory that the simpler and easier the experience, the better it would work. We had a theory about building a network effect, which happened.
You know, there's a huge device ecosystem today. You can watch things anytime, anywhere. But for each of those areas, I defined a swim lane. So for instance, Todd Yellin was the director of product who worked for me. His focus was personalization. And his job was to articulate the strategy for personalization.
And he would use these same tools, models, and frameworks that I use, with exception of GLEE. There's only one GLEE model for the company. But, you know, Todd was able to say, okay, the job with personalization is to connect people with the movies they'll love. What do we need? We need explicit taste data.
We need their ratings so we understand what they like and don't like. We'll have implicit data. You know, we know if somebody watches a movie for two minutes and quits, that's a sign. We need all the data about the movies. So who's the director? Where was it filmed? And then there's this thing called the movie genome project where everything is tagged.
So you know it's steamy romance or you know it's got a strong female character as lead or you know it's quirky or whatever else. And then the last step there was to create algorithms that connect knowledge of people's movie tastes with the movie data. And do that with algorithms and also presentation layer tactics.
So at a high level, I've given you the high-level strategy for personalization. And then for me, every time there's a strategy, there's a way of measuring if that strategy is working. So for instance, the proxy metric for personalization, and this was Todd's job, his job was, it's kind of a wild proxy.
This is a proxy for retention because retention is so freaking hard to measure. His proxy to prove that personalization was going to work in the long term was the percentage of members who rated at least 50 movies, 5-0, in the course of a month. And I'll let you think about what was that number. Most people guess 2%.
The reality is that we drove it up to like 30%. So that was a proxy. You know, the reason that people were rating movies was because they knew that the more that they rated, the better Netflix would do at helping them to find movies they'll love. So that's an example of a proxy metric for personalization.
That was Todd's job to drive that one metric. There were a number of other metrics as well. But I just took you a little bit into the world of personalization, that swim lane, that was Todd's area. He could articulate his personalization strategy, the proxy metrics he used to measure whether these strategies were working, and then he had projects or tactics against each.
I mean, the big dog in getting all that rating data in the first two months of a member's life was something called the ratings wizard. We put in the tab, it said, movies you'll heart. There was a little red heart that the design team described as fugly, which at the time I didn't know what that meant, but now I do.
But people would click on that and it would open up and it would say, the more you rate, the better we'll do at finding movies you'll love. And there were rows of movies and people would go on these rating jags. And that's how we collected like 10 billion pieces of data to understand people's movies tastes.
I could do the same thing I just told you about Todd's life in personalization. Brent Avery launched streaming. So he had a specific strategy and proxy metrics in his area. Megan Stern was focused on creating a simple and easy experience. Crystal Chincuti Trexel was focused on essentially DVD merchandising.
So I would have a high-level product strategy and then each of these swim lanes would have its own well-defined strategy, articulating the strategy, the proxy metrics to measure if things were working, and then the projects or tactics against each of those strategies to measure if it was working.
Host: Awesome. So, here's my high-level summary. I would propagate the GLEE model, how to think big in these 5, 10, 20-year chunks. It would feel very close to a company strategy. I would have a high-level strategy, and then I expected every product leader who worked for me to be able to articulate their product strategy in their swim lane.
Host: Yeah, it sounds very similar to the way that I think of the levels when I talk about product strategy. Because I do so much deployment and a lot of my work in the in the past six years consulting has been trying to get leaders to think bigger, right? And not just have them dictate a bunch of features down to teams.
I usually talk about things in strategic intents, which sounds a lot like how you would come up with the GLEE model. The GLEE model is very much the meat of it. Like it's your framework for thinking through what is a strategic intent. So in my world, it would be, you know, expanding into streaming would be a strategic intent.
It would be like, okay, that's the way that we want to go with the metrics associated that you talked to. And then the product initiatives would be like your personalization initiative underneath it. What are we going to do to actually look at that? And then the options I call them are the projects.
So I love how this all ladders. Yeah.
Guest: Well, so and you're really getting to one of the key challenges, especially as companies get bigger, which is there's a phrase, to tame it, you need to name it, right? So the importance of having shared language, which is why I write a lot. So I write on Medium about these different tools and models and frameworks.
So I want people to develop shared language so they can talk about this stuff together. I'm going to jump on a question I know you want to ask me, which is, you're the product leader, a big company, and you're thinking about what's the right amount of resources to put into each of these areas. I mean, you use the phrase deployment, right?
Host: Mm-hmm.
Guest: I'm cheating. I'm looking at your Google Doc.
Host: It's okay. You're anticipating my questions. It's perfect.
Guest: Yeah. So here's the cool thing. So I do these quarterly product strategy meetings where each of these product leaders that runs a swim lane is coming forward with, you know, their theories and hypotheses, their results and learnings, the AB tests, and then these future hypotheses. And that's helping me at a high level to figure out how much to invest or not invest in different areas.
So for instance, in personalization where that proxy metric went from 1% all the way up to 29%, we kept investing more and more in personalization. There was another theory, the theory that a social experience would improve members' lives on Netflix, that if you and I connected with each other, we could share movie ideas with each other. And we wouldn't want to leave the service because we wouldn't want to leave our friends behind, right?
And that we launched something called Friends on Netflix. We launched it at 1% of members were using it, and after like four years of pounding our head against the wall, we drove that metric up to 5%, which is not enough if you're really trying to improve retention. So I just remember one of these quarterly product strategy meetings, we're looking at the growth of personalization where things are working really well, and then realizing that that we really weren't proving out our hypothesis that social and friends would create a better experience for Netflix members.
And there was one quarter where we just killed the friends effort. We took all those resources and put it against personalization. By the way, the consumer insight on why social and friends didn't work at Netflix, there's really two reasons. One is your friends have sucky movie tastes. And we thought we could solve that through algorithms.
We couldn't. The other is you don't really want everybody to know what you were watching last night. Last night I binge-watched Cake Boss, right?
Host: Oh, yeah.
Guest: You don't really want to share that with the world, right?
Host: There's some weird stuff on Netflix, too. I don't want people knowing I watch it.
Guest: There is some weird stuff. Anyways, these quarterly product strategy meetings, one of the big part of my job there was just to figure out the right level of resources to each of these swim lanes. You can imagine, you know, Brent Avery, we launched streaming in January of 2007. At launch, 5% of members watched at least 15 minutes in the course of a month.
By the end of that first year, we were at 20%. Like, okay, let's keep doubling down on that area, right? Or Megan Stern was able to prove that a simpler, easier experience improved retention. Like, okay, Megan, go to town, figure out anything on our site experience that's complicated and just reduce it.
So these are the kinds of conversations that would happen at these quarterly product strategy meetings to figure out the right level of resources in each area. And for me, that was a big part of my job.
Host: And it sounds like you're also, you know, you're talking about hypothesis over and over again, which I love. So lots of learning baked into what you're doing. But a lot of those hypotheses don't come to fruition, right? You were just talking about moving stuff over to personalization. One of my favorite stories about Netflix is actually Project Griffin, which was when, and I want you to actually tell the story, but you were just telling me before we jumped on this, there was even another failure story that happened in 2005.
Can you tell us a little bit about what were some big failures at Netflix and how did you use product strategy to decide to kill it?
Guest: Sure. So I gave you the the GLEE model. At Netflix, we're going to get big on DVD, and then we're going to lead streaming, and then we're going to go international. But it didn't exactly play out that way. To your point, that's a high-level hypothesis, right? We weren't really sure what the right year was to launch streaming, for instance.
To answer your question, a big failure circa 2005 was we got ahead of ourselves. And we invested like a year's worth of energy in launching a DVD by mail service in the UK. And one week before launch, a bad thing happened, which was there was a rumor that Amazon was going to launch a DVD by mail rental service, which, like at the time, we were like at a million or two members.
Like, we were still a yipping chihuahua, and Amazon was already a big dog, right? So this was totally scary for us. So one week before we launched in the UK, we canceled it. And you can imagine it was devastating. People had moved over to London, etc. But there were two ideas. One, we couldn't fight a war on two fronts, right?
We'd gotten ahead of ourselves. It's really hard to implement a system that ties into local mail systems, whereas if you think about it, it's really easy to stream internationally, right? So that was like, okay, duh, we were stupid. But the second was that Amazon was obviously a huge threat. So we needed to stay focused on our efforts in the US.
In that case, we took all of the resources from the UK and we put them back against personalization and also the work that we needed to do to enable the launch of streaming. So that was a huge reallocation of resources. It was devastating. You can imagine if you were on the team, you know, how you felt.
But at Netflix, we built this culture where you spend a lot of time helping people to understand what it would take to make the business effective. And in this case, people understood that, you know, the right thing for the company was to cancel the launch and reallocate the resources. Okay, so now I'll get to the story that you were curious about, Project Griffin.
What happened was we did launch streaming in January of 2007. You know, I gave you the metrics at launch, 5% watched at least 15 minutes. We only had 300 titles, and our members described them as steamy romance. Steamy romance, which is code for something else that's bad, which is to say we just didn't have very much good content.
But after a year, we got to maybe 20%. When we launched, it was only on a PC laptop, essentially. That's where people watched. We didn't have on Mac, we didn't have mobile devices. And our insight was we needed to get to the TV because that's where everybody wanted to watch. So we would knock on the door at Xbox, and they said, no, you're too small.
We'd knock at PlayStation, you're too small. We'd talk to the Wii. These were the big game systems at the time that were connected to a television. And they kept saying no, because at this point, maybe we were up to maybe four or five million members. So nobody would play with us, so we built our own box.
And that was, that's what you call Project Griffin. And we knew that it was going to be a big success because we had done a painted door test on our site. We said, hey, members, for $70, would you like a magic box delivered to you that will let you watch on TV? And people clicked the button. We said, sorry, it's not available in your zip code or whatever we said.
I call that a painted door test. But we had a really strong signal that a lot of people wanted this. So we built the thing. And one month before the launch of this, this will sound very familiar, our CFO, Barry McCarthy, came back from New York and said, hey, the investment community is not going to buy into this.
They think that we're going to become a hardware company. And they know that hardware companies have notoriously low margin. They don't make a lot of money. And they know how hard it will be for us to get good at the thing that Apple is good at, like figuring out how to ship boxes across the Pacific Ocean on time, you know, all the operational stuff.
And so, you know, we had a heart-to-heart with ourselves. And Reed Hastings, he's the CEO of the company, he called me on a Saturday. He said, hey, Gib, I think we're going to cancel the launch of this box. And his question to me was, how do you think Paul will feel about this? Paul Asker was the product manager who worked for me, was focused on this.
And Reed's question was, should I do it via phone or should I show up at his house? I said, you know what, Paul's a well-formed adult. I think you can pick up the phone. He's got enough context to understand why we don't think this is a good idea. You know, here are the new learnings, etc. Which is exactly how it played out.
But you can, you can sort of feel the ethos at Netflix about understanding, doing the right thing for the business. This one, I didn't, you know, Paul, frankly, he doesn't feel that stupid about it either. This box became an amazing prototype. So we showed Xbox and PlayStation and Wii what we could do.
Yes, we threw 100 engineer out of the building. And that's the formation story for Roku, essentially. And then Roku launched their box maybe a year or a year and a half later. In the meantime, Xbox finally said yes. And in their first month, they had a million of their members on the gold membership streaming Netflix to the television.
And because the Xbox went so well, they were the number three platform. PlayStation knocked on our door and said, hey, will you do that for us? And we said, yes. And then the Wii came around
