Platform Strategy w/ Marshall Van Alstyne
My colleague Seth Joseph and I sat down with Marshall Van Alstyne, Questrom Chair professor at Boston University, co-author of bestseller Platform Revolution, and Digital Fellow at the Massachusetts Institute of Technology. As two Questrom alumni, it was our honor to talk with Marshall, a global expert on network business models, about trends in platform thinking.
Summit Health (SH): When your book “Platform Revolution” first came out, the term platform was just starting to come to prominence. Now the term is often used. How do you define a platform?
Marshall Van Alstyne (MVA): We use a very simple definition, but each component really matters. We define a platform as an open architecture with rules of governance to facilitate interactions. The open architecture is what allows third parties to come in and add value. The rules of governance motivate them to come in and add value, and then the interactions create value. Without the interactions, you don't get the ride, the tweet, the auction purchase. You need all of the interactions that actually create value. You use the governance model to motivate more healthy interactions and to mitigate unhealthy interactions.
Again, a platform is an open architecture with rules of governance to facilitate interactions.
SH: A lot of the research we’ve seen on platforms focuses on B2C platforms. What are some of the key differences between platforms that serve B2C vs. B2B needs?
MVA: I haven't seen a lot of good work on this subject. As a matter of fact, the chapters added to the next generation of “Platform Revolution” will be about this distinction. We will be doing that this summer.
There are quite a few differences between B2B versus B2C. Most of them have to do with the organizational structure of the partnership, but also the firm itself.
The first thing you need to think about is the interaction for B2B. The frequency is lower, but the value is hot. So if you think of Google searches, for example, there are trillions of them where the value of each individual is tiny. If you think of B2B purchases, there are not nearly the same numbers, but the value per interaction is much higher. So to think of that differently, you are still trying to increase the frequency of the value of the interactions, but you are high on the value or lower on the frequency.
In the B2B context, you're often doing it in one-way matches instead of a two-way match. So B2C platform matching a person to a ride, a tweet, or what have you. Just two parties. That's much easier.
Suppose it is something like SAP or Salesforce or project-based or other B2B interactions. You may have multiple parties and buyers putting together a servicing one so they have to match a part to a plane or to a staff member. There are three or more parties involved in the interaction: so there is a higher chance of interaction failure. So again, it becomes more important to take the friction out of those interactions.
Another element is that it's more important to get cooperation among the parties, where maybe there is competition. This was the case for SAP, as an example. A lot of independent software vendors are competitors. They set up some knowledge marketplaces that actually enabled sharing knowledge across the different B2B partners. And there's even excellent statistical evidence that participants are more productive. They have higher IPOs, and they have stronger IP rights.
And so that cooperation was actually extremely important to the knowledge sharing elements. You also need organizational design elements. If you place the responsibility for the platform underneath the business unit, there's a tendency that if partner products compete with your products, the business unit kills them off, bad idea, because that's bad for your client.
Your clients want the best of breed. Maybe a competitor's product is better, and it needs to be allowed to thrive in the ecosystem, and you need to learn from that.
SH: You already touched on this a bit, but we know that companies building platforms face different startup challenges than those building pipeline businesses. What advice do you have for startups that are looking to build out a transaction platform?
MVA: It's harder to get critical mass when you're starting from scratch because, again, what you need is network effects where users create value from one and not for one another, but if you don't have users, they can't create value for one another, and if there's no value you can't get users, so you've got a chicken and egg problem.
It's more important at launch for you to find some compelling hook or create some compelling value proposition on your own. That gets folks in initially whereas if you have a product, the real challenge is thinking more clear-headedly.
A traditional B2B company views sales as a "hit and run" activity. But for a B2B platform, a sale is just the start; the company needs to focus on facilitating ongoing interactions.
In B2B context, a good example is the difference between selling software on-premise, where it goes behind their firewall, and you never see it again versus operating it in the cloud, you observe all the behaviors you know what's working, what's not working and you can improve it over time.
That's another critical difference if you're launching from scratch; it is more important to get your network effects going. If you are launching from an existing product, you have some other options to tap into an existing customer base that is easier to launch on its own.
We can give you 8 to 12 different strategies, but all of them focus on getting interactions, going frequently. A couple of things you need to watch for is a tendency to try to build a broad platform initially when what you need to do is to focus on a niche where you can create some killer app first.
Numerous companies have made mistakes by trying to go too broad, which narrowed their value proposition to a level that didn't attract interactions.
SH: I am curious. If we then flip it and think about established firms who have a market footprint and seek to "platformize" portions of their business. How should companies solve for that?
MVA: We have large EHR systems that received the benefit of federal funding for doctors and hospitals to adopt EHR. There was a considerable period of growth for them in 2011, though, in 2016, growth stalled.
I expect that many of them are now thinking, what is our next growth option? They may still be thinking in terms of pipeline business, what's the next vertical or the next killer app? They can face challenges if they consider converting their overall business from a pipeline business to a platform.
Although that could be very advantageous for them, they tend to think with a pipeline mentality. Which means if a third-party developer starts having success, they may point to it and say, Oh, well, that's what we should be doing. And we're going to kill them off and build that.
SH: So there are challenges concerning conflicts of interest in governance. I wonder if you've seen that and how do established companies seeking to platformize solve for the problem that you described earlier, whether it is at a business unit level or just a pipeline versus platform mentality, how do they successfully platformize and solve for the conflict?
MVA: That happens frequently and requires a mindset shift. It is not articulated as much in "Platform Revolution." It's a little bit in the Harvard Business Review article, but to do the shift in mindset properly, you need to think of platforms as inverted firms. That's a characterization. It's not a definition. By that I mean, the value is created outside, not inside. The reason once you see it is straightforward, you cannot scale network effects inside pharmacies, outside of the more people, in resources outside the phone. Once you see that if you want to scale, then you have to use network effects. You would have to scale network effects outside of the phone. So you need value-creating activities and human resources in marketing, strategy, and finance.
So you have to get that mind shift from you are creating all the value to enabling other people creating value. And this goes to the shift in the mindset on, okay, we see someone else creating value. Let's go kill them off or create our own because if you do that, what you're doing is you're telling people who are creating value for you that you don't want them to do that anymore. You're going to take away their toys, their investments. So the proper answer to what you're doing is you buy them out at a fair price.
You reward them, and then you showcase them so that other parties do the same thing. So rather than simply quashing them, you reward them for what they have done. What this does is it signals to other parties if I invest in this platform that I might be rewarded too, Now you have to negotiate a fair deal to do that.
But you have to be careful about not killing the goose who lays the golden eggs. That is, third parties creating value on your behalf, the governance model is what motivates that, and you need to tie your own hands or to reward others in ways that then lead other parties to trust that when they invest, they will be rewarded too.
It is because their product is successful, and you leave them alone, or you buy them out, or you connect them; you help them do business in the two-sided market. That's what you're using your recommender systems to cause more of the interactions to happen. Then you take a slice of the interactions that occur, and so you're facilitating their value creation and participating in it.
SH: Narrowing specifically on healthcare, what areas of healthcare do you see the most opportunities for platforms in, and which assets do you think will be most important?
MVA: It's going to be different if you are a healthcare provider or insurance company. Each of you has a different set of assets. And so you're going to be trying to control the interactions with very different angles. Now, who is going to win these battles at this point? I honestly don't know.
Each has different leverage points. And each has different powers in this marketplace, but it's almost like Apple competing with devices and Apple and Google competing with services. Each is coming at it from a different angle. There is every reason to believe that healthcare will transform because of the high proportion of value created from information.
It's easy to create network effects out of non-rivalrous information. The parties that do that more successfully will probably be the parties that succeed at the moment. I'm sure you're familiar with the failures of Google Health and Microsoft Health Services. Those are companies that know how to do platforms, and both failed for various reasons. One of them was that they were not party to the interaction themselves, unlike search or email or Office or other things where they're already part of the relationship. They could not insert themselves in the information flow. They didn't figure out what they were doing, and so they failed. Different parties are going to have different access points. Insurance companies have a lot of the data, the doctors and the hospitals have a lot of the data. The patients technically own it, but they really do not have much access. We're starting to see things like interoperability, but quite frankly, that's still friction-filled and not very useful data. I don't see those consortiums actually being worth the rent. Is it going to transform eventually? Absolutely.
SH: There is a lot of what you just said that resonates. Really well put.
Should venture capitalist investors putting money into early-stage firms who are evaluating platforms be thinking differently from the way you think about managing traditional technology companies and pipeline companies?
MVA: They are fundamentally different. You cannot use the old metrics. You'll get it wrong.
SH: Do you have counsel for investors who are evaluating platforms in terms of what key metrics they should be thinking about?
MVA: Look at the user base. The most obvious examples are Facebook buying Instagram, Amazon buying Twitch, Google buying Kaggle. We will have an academic report on how to do that hopefully by the end of the year. The current investment community doesn't have good metrics for measuring the value of platform firms. We're launching a research project on how to do that, and we should have results within the year.
If you are on the other side and you're a founder seeking to raise capital, in particular, something that I'm aware of, having lived this, monetization can be a barrier to network growth. It is not in all cases, but it frequently can be.
SH: If you are a founder pitching to investors, building a platform business, what is the story or the angle, you would advise?
MVA: I tell firms in the management process that there is a very different sequence in platform businesses, relevant to product businesses.
Initially, you use monetization to drive network effects. After you hit critical mass, you use network effects to drive monetization, and it's in that order. And not any other order. If you screw that up, then you are putting friction on your interactions, and you do not grow.
SH: Can I clarify, because you said monetization both ways, but you should use capital to drive network effects, and then in the second stage, once you've flipped and reached critical mass, you use network effects to drive monetization?
MVA: You're spending resources to drive network effects initially before you have them. Remember, there's this chicken and egg problem when you're starting.
How are you going to solve that? You spend some resources to solve it right until you hit critical mass. If you don't answer that problem now, you also need to describe after you've hit a critical mass to place your taxes in the interactions stream.
To take a share of the ecosystem’s revenue, you then have a vision for what you will do to make that clear, so that is how you tell the story. Then you could also run some estimates as to how much you are likely to make based on estimates of deal flow and value creation and your ability to tax post-critical mass.
SH: Before we let you go, can you speak about a new area of research you're currently excited about?
MVA: I've got three or four. The most immediate one is how to value platform firms. That is one; actually, that is three months away. I'm also heavily involved in platform antitrust. These platforms are likely to be regulated. I’m working with the European Commission, and that's more than forthcoming legislation. I'm also working on solutions to the fake news and misinformation problem.
SH: We look forward to seeing your work in these areas once it is published. Marshall, it has been a fantastic interview. Thank you again for taking the time to talk with us.