Greymatter Blitzscaling | Going Global with Reid Hoffman & Chris Yeh

Hi everyone, I'm Reid Hoffman, a partner at Greylock and host of another scale.

Essential podcast, Masters of Scale.

Welcome to our new Blitzscaling series, a part of the Gray Matter podcast, elaborating on the book that Chris Yeh and I published, Blitzscaling. Chris and I have been fielding a number of great questions from entrepreneurs and others on on blitzscaling.

So we decided to answer some of.

Them here on Greylock's Gray Matter podcast. Today we will focus on globalization.

So let's get started.

So our topic for today under the heading of strategy is globalization. And of course, as always, I am joined by my co author Reid Hoffman to talk about this topic. So, Reid, I guess let's start with the basics. How does a startup know when it should start thinking about globalization?

So part of the thing to realize and to think about from the very beginning is how much does your product have to be global? Because frequently in consumer Internet there's big differences between being, or to put it more precisely, multinational, because there's huge amount of, you know, greater than 50% of Google's revenue and traffic comes from outside the US et cetera. You have to think about like, okay, how important is that? And then what's the cost of that in terms of its timing? All entrepreneurs inherently want to be the I got my product service, I want to be the global provider, everyone for it. You know, I generated this idea, I want it to be there. But part of the key judgment is what is your sequencing? Because investing in other nations, other languages, other cultures is frequently very expensive. Frequently the question of product market fit can be fairly different for your product. Sometimes you're lucky it's the same, but frequently it's fairly different. Certainly there's different expectations about how dense the UI is, how the words work, how visuals come together. So there's a whole bunch of things there. And usually once you begin to do multi language, that that creates a broader set of UI constraints and programming constraints and everything else for doing all that. So the typical pattern is to say, okay, if I'm a consumer Internet company, I should be global English as fast as possible, then it's all right. What other markets are really important either from playing in that market, dealing with startup competition, dealing with figuring out product market fit, getting the right scale to work, how do those things work? So this is what we did at LinkedIn. We launched English only. We were surprised at how quickly we moved into the rest of the 200 plus countries because with great rapidity we were going over to countries that we didn't even know were countries. You know, maybe that's our American geography or our American non global geography problem. But we quickly moved into them and then the question became is, okay, well, when do we go into key languages? China is obviously huge German, French, Italian, Russian, et cetera, et cetera, et cetera, Arabic. And so what we decided was that we were working to critical mass density, we were working to the features that would get us to critical mass density and that we would try to be good for global English, but we would take some time getting other languages. Which meant that a variety of competitors who said, oh well, we're going to be LinkedIn, but in Chinese LinkedIn, but in French LinkedIn, but in German, they could all essentially be going for their market, perhaps even discovering some very unique things about their market to really kind of solidify that in. And we might get there late, we might be a distant second player, you know, that kind of thing. And we decided that was okay because.

It'S always a trade off of risks.

Part of the thing about talking about LinkedIn is the trade off of, well, are actually going to get the right product market fit for getting the right engagement to scale is a much more important thing to get right than are we going to be in all these other languages. Even though all these other languages and all these other nations are ultimately super important because it's a global professional network and business happens across the globe in the world. And so that's I think the key set of kind of trade offs. Now this can also get harder on regulatory regimes, on go to market strategies that are, that are country specific. Doesn't even have to be like obviously hardware, it tends to be country specific. But it could even be the, oh, I'm doing educational software. Well, educational software is wildly different and in a lot of different countries. And so it may be that. And so you have to factor through all of those in terms of what you're doing. But presuming that you're doing kind of the classic we have a new technology for the entire world, the set of conditions that I just outlined, those are at least a core set of them.

And I think something you said is very important, which is that globalization is in fact a trade off. There is a trade off between how quickly you globalize and, and how many resources you have to devote to other things like continuing to refine the product, et cetera, et cetera. One of the things you mentioned is that at LinkedIn you made a specific decision saying it would be okay to allow these competitors in other countries in other languages to compete for a while, because the most important thing was to actually focus on that global product market fit. Eventually LinkedIn did localize and move into different languages. Can you talk a little bit about the decision process you went through to decide the which languages to prioritize and did you pursue the same strategy for each language or was it different for each language?

Well, so first, just to emphasize one other thing in your summary is it isn't just there's a trade off between globalization of feature and other features you do, but also as you put in the globalization feature, the speed at which you iterate on other features tends to be slower. And that is an important thing to say, okay, we're going to make this now part of our acceptance test and new features and new UIs across the whole thing, that then makes that whole loop slower. So you only want to do that at the right time. So then when LinkedIn went global, English got a lot of use in just about every country. In English you have a little bit of innovators dilemma problem, because when you start introducing the language to broaden within a country to some degree, the people who are there were there because the English or expat community, etc. Is their own little special community. And so you get some blowback from that. That's a classic, as we put in blitzscaling. When do you ignore your customer as a key thing to getting to scale? And that's an example of it. We were looking at kind of an index of where was there already a natural product market fit for the kind of product LinkedIn is? Where did we have a user base for doing it? Where do we have an addressable market that's extremely valuable and how much work was it to get there? And we were essentially triaging across all of that. So on one side you had kind of high natural use within the Central and South American countries. And so moving into Spanish was kind of obvious and straightforward to do. And to the far other side you had, well, we were only in the English expat market in China, but actually, in fact there's obviously so much going on within Chinese global economy, we knew that we needed to be present in China too. And you know, one relatively straightforward work along with doing the work for French and German and Italian, et cetera. And then the other one, we ran a dual track, both globalizing LinkedIn and experimenting with a local product LinkedIn variant called Qi2 to learn from that and develop that. And so that's your whole range. And then you, you add based on do you have the bandwidth to do it? Not just technologically, but also launching it, making sure that works, et cetera, et cetera, and getting those things going, you obviously have to design in a global network, how do you handle language profiles that might show up in searches and so forth and how people go, okay, I speak only language X, usually English, I speak only that, and so I'm disoriented if I see an Italian profile. How do you handle all that? That's all the work that you had to do. Although once you of course do that work, generally the incremental languages tend to be more the specific set of user interface, specific set of cultural norms, specific set versus the. Well, now we know how it is that people who are disoriented by things other than English mostly don't see anything that isn't only in English.

And as you're going through this process, at some point in time you hire a head of global or international, how should people think about that? When would they want to hire someone to run this business? And what are the key leading indicators to watch for that tell you now is the time.

So generally you want to hire your kind of first global international language cultural leader, minimum 6 to 12 months before you go because you want to do all of the long term light prep. You want to be modifying some of your key systems that are the long poles in the tent early. You want to be teaching the company what kind of multilingualistic multicultural, multinational really means and what the issues are. And so it's generally speaking in advance. Now, if you're like, oh look, we're three years of tech development, we're not even there yet, then you might be still talking to some people to make sure your product has a higher chance of product market fit on this. But you might go, okay, I'm not going to hire anyone. You know my small group budget yet until then. So that's some parameters.

And how large are companies typically when they take this step? Would they take this step when they are just 20 or 30 people? Are they going to wait until they have a thousand? I know it's going to vary, but what are some rough guidelines?

Well, it does vary a lot on are you a consumer company, are you an enterprise company, are you a hardware company, are you something else company, pharmaceutical, et cetera, et cetera? All those vary tremendously. And I don't know if I put them exactly this order, but roughly speaking, the lightest is we're a consumer Internet company. And you might do that as soon as Instagram having 13 employees, it's probably the extreme end. And then it might be thousands when you're dealing with medical regulation or you're dealing with a completely unique go to market and payment strategy in medicine or something else as a way of navigating it. So it can really range. But you should be paying attention because again, first to scale Cadillac, first to scale gets surprised. And first to scale frequently is first to global scale.

That brings up one of the interesting topics you mentioned, that you made an intentional decision to focus on going global in English first, knowing that there would be local copycats in international markets. So how should a blitzscaler respond to the emergence of local copycats? How did you respond to the emergence of local copycats?

What I realized at LinkedIn was that what really mattered was the density of the global network and its ability to deliver value in growing ways within each country, each local network, each local industry even, and so forth. And so we went, all right, what really matters is acceleration of that loop. And then the question where globalization plays into is the cost benefit for the acceleration of that loop. And I realized that we had enough to do with global English that even if you said, okay, we're going to seed all the rest other competitors, then a, we'll still have something super valuable and useful. But B, part of the question that I had about a variety of the competitors that were in Europe was, well, are you going to get to enough density? You say, oh, we have a feature, we're in English, German, French and Italian, Croatian, etc. We're in all those languages. Okay, but you know, is your feature set going to get you to that compounding loop or not? Because you put in so much work to that. And we were broadly correct that it was far more important to get to critical mass density, far more important to get to the feature set that allowed the key business use cases. Obviously, people tend to associate LinkedIn with job searching and hiring, which is obviously a major use case, but also business development, finding experts, I mean, a whole stack of things and to make sure that those were all well provisioned. So our decision was speed to scale is what matters. Globalization fit in there as a. Where does it fit within the speed to scale?

Right. So the primary focus was, as it always is, in blitzscaling, being the first to scale and in this case being the first to scale at a global level. And despite the fact that these folks had some local strengths, they couldn't match your density, they couldn't match the network effects you achieved.

Exactly.

So let's now Think about some of the listeners out there. Everyone's listening to this in English obviously, but for many of those folks, those listeners, English is not their native language. They may be speaking French or German or Spanish or Mandarin or what have you. Excellent. I would attempt Mandarin as well, but I think my Mandarin might actually be worse than yours. In a grand irony. But let's imagine the entrepreneur thinking about I'm going to begin my company in my native language. How should I think about internationalization and globalization from a non English perspective?

The analytic framework for these things tends to be the one that gets to the breakout scale first is one that really works. So wherever you are, you're trying to work your way back to that. So one of the advantages that we have in Silicon Valley is that Global English is a very important market. And so by being simplified and launching global English, that tends to give us a strong first mover effect on what the setting the standards, the technology, the use case is. And then we can back up into going into other languages, you know, French, German, Italian, et cetera, you know, Arabic. Now China is a bit of an exception to this. The Chinese tend to be very focused on their market, but also limits their expansion out because they get it all in Chinese and then they don't get to any of the rest of the global market. Whereas in global English, well actually in fact we have some folks in Nigeria using the product, we have some exposure to it, we have some pattern of how we're going to extend into other languages, other cultures. So when you're not native in English, you have to be thinking about the same problem. Now you may say, well we have a local captive market that's really dedicated to, that's very easy to address, country X, language Y. And we're going to do that. But you have to be thinking, is that really the whole story or do we need to be bridging urgently to the rest? And sometimes that answer is bridge to another country. Like if you're going to do France to also Spain or you're going to do France and Spain also do Germany, and that depends, like it's a medical product, it's an education product, it's an enterprise product with a go to market strategy of certain ways. All of those things could lead that way. But if it's like a consumer Internet, what you want to do is say, okay, well probably if I'm language X, which is not Global English, my second one probably wants to be Global English to play that battle. And you may want to do that very soon because part of what you do is say, well, if somebody is fighting for the global scale and they're coming at me with another language, do I care? So if I have English and you're coming at me with Croatian, don't care. English is right. But if I have Croatian and you're coming at me with English or Chinese or something else, then the high likelihood is that that will matter a lot for me. And so that's part of the calculus of prioritization decisioning.

As I think back on what you've shared with us today, it really seems like there is this core tension as we try to fit globalization into the concepts of blitzscaling. Globalization can be very helpful for blitzscaling because obviously you're reaching a broader global market. But the act of globalizing, of working in a multinational framework with multiple languages, tends to slow down the company because it increases the overhead for changing the product and innovating and doing new things. And so there really is this grand tension. You have to look and see, does globalization or internationalization accelerate our process, accelerate our progress towards being the global leader at scale, or does it not?

Exactly. And obviously part of the reason why the default tends to be is okay, it's much harder but also important to do to get to the global scale. You tend to lean a little bit on the entrepreneurs to take it more seriously because it's uncomfortable and discomfort. But on the other hand, that doesn't mean go do it immediately. It depends.

Do you have any final words of wisdom to the entrepreneur sitting in their garage? They're at the very early stages of the company. How should they be thinking about globalization at that stage?

So you should be asking the question about are there any low hanging fruit, are there any easy ways to do it? And you should be hopefully lucky in your product if you can be, and recognizing if you are lucky and then taking advantage of it. So an example of luck is Twitter with 140 characters, worked very well in Japan and is going gangbusters in Japan. And so it had a very natural fit and actually for many years and I don't know what the current state is, but it took a lot of effort from Facebook to also be interesting and considered in Japan as Twitter had dominated the market so much. But you take a look at it, you say, okay, are there ways I can get to global sooner and easier and cheaper? And then what are small things that I can do that when I do get to it means that it, as opposed to climbing Everest, is like climbing a local mountain.

As always, Reid, thank you for taking the time to share these insights. We'll be back next week with another look at the deep techniques of Blitzscaling.

Thanks for listening. This is Reid Hoffman, partner Grelock, host of the podcast Master the Scale and co host of this series on the Gray Matter Podcast. To get this podcast every week, subscribe to the Gray Matter Podcast on itunes or wherever you prefer to listen to podcasts. If you have any feedback on any topics we discussed today, or any questions you'd like to ask in the Future, tweet us reylockvc and reidhoffman with AskRead or, of course, post on LinkedIn.

Chris, Shay and I go through the.

Questions to select questions we will answer on future versions of this podcast. Thanks to the team that produces Masters of Scale and the team that produces this Blitzscaling series, and most of all.

To you, our subscribers.