Startup diary: Developing Facebook-style network effect and startup holy grail of the viral loop are a must
Getting your users to invite new ones can give you a significant advantage over competitors, writes the voxgig founder
Last week I discussed all the ways we won't be trying to compete: on service quality; on user experience and design; on geography; on depth of features; on price; and on being the first-mover. If we're not going to use any of those to compete, what are we going to do?
You want to compete on something defensible. It has to be something that is hard for others to replicate. Generally this comes from something that you have exclusive access to. You can compete in a geography if you are the exclusive distributor for a product in that region. You can compete on features if you have a massive war chest, or strong patents (both quite hard to get!). You can compete on price if you've figured out your economies of scale and got internal processes up and running to make them work.
For Software-as-a-Service businesses like voxgig, there are three primary things you can build defensive moats with: network effects; platform effects; and data effects. We're going to use all three.
The network effect: when the value you get from a service increases the more users it has. The classic example is the telephone. The more people have them, the more valuable they become, because you can call even more people. Microsoft has benefited greatly from network effects - they could almost be called the foundation of their business. The more people that used Microsoft Office, the more it made sense to learn it, because your future employer would probably be using it. Even now, the ability to use Microsoft tools is a basic skill for most office jobs.
Network effects are very powerful when they take hold. Facebook obliterated all competition once they got one going. That is the trick though - how do you get the network started? In the early days you don't have the user numbers, so the network, you service, has very low value. You can overcome this problem in two ways: you can provide some initial value, and you can focus on smaller communities.
If your product provides utility independent of network effects then you can get users on board just by selling that utility. That's why we are building tools for people in the events industry. We're providing tools for speakers, organisers and exhibitors that make it easier for them to do their jobs. We are also focusing on helping speakers in particular, and making that the first community that we want to build - that's why we put so much effort into a newsletter for conference speakers. This approach is directly inspired by Facebook's initial strategy of targeting third-level colleges.
The key to getting the effect to take hold properly is that holy grail of startups - the viral loop. You need to get existing users to invite new users, and those new users to invite yet more users. You can create a natural incentive structure to make this happen. Dropbox, the cloud storage and backup company, is the best example of great execution when it comes to this strategy. In the early days of Dropbox, when you invited new users onto the system, you got extra storage space, and they got extra storage space. You could get loads of free storage the more people you invited. Simple, and wonderful.
Network effects have some very desirable business consequences. User acquisition becomes very cost-effective. Instead of a fixed acquisition cost per user driven by advertising costs, your acquisition cost per user keeps falling as the network grows - people seek you out because they want to join. Switching costs become very high, so your users stay with you. That gives you high retention, meaning that you keep most of your paying users long-term. You develop a strong defensive position, because it is much harder for new entrants to take your customers. No matter how good their features, your platform is the place where you can find and collaborate with everyone. That beats features any day.
The second competitive advantage that we want to develop is a platform effect. If you can become a platform where others make money, then you create a self-sustaining ecosystem where your partners generate business for you. Sales force is the model here. They provide a robust cloud-hosting platform, and even their own programming language. We'll definitely be building our own Salesforce app so that we can integrate with them. But, in turn, we will do something similar - so that anyone who has online services offerings that are valuable to the events industry can use our platform to reach customers.
The competitive advantage here also comes from switching costs. In this case, partner companies that invest in building software integrations with your system, that invest in marketing efforts to sell those integrations, and invest in training to learn your system, are not going to switch to a competitor that easily. Cheaper prices might get them to abandon you, but they would have to be much cheaper to justify new investment. And you're still offering good old networks effects, which your competitors are not. The platform effect helps to monetise the user base that you have created, and sustain value in the long term by making your system ever more useful.
Seeking a platform effect does have an immediate impact on our business. It takes longer to develop a software platform than it does to develop a software application. Platforms are more complex because they are the host environment for applications. Choosing this strategy has slowed us down. In many cases, this is a bad idea, especially in new and growing markets. However, that is not the market that we are entering. In our market, event management software, there are over 300 vendors. It is a large and mature market. Luckily, it is highly fragmented with all those other vendors, so there is lots of space for new entrants. Almost all of them are positioned as applications, not platforms. It is a safe play, but you end up being just another vendor because you don't have any significant differentiation from all the others.
There is a question about the event management software market that is worth asking - why is it so fragmented? Surely some of the vendors are better than others, and they should be attracting the majority of the market share? Event organisers should gravitate towards the higher quality offerings. I think they don't because the industry suffers from data lock-in. This occurs when you have entered so much data into a cloud system that it's very hard to extract everything and move to another one. Lock-in is a form of competitive advantage, and it does work, but it is justly and rightly despised by users. It is also unstable. It's not so hard to automate data extraction, nor so expensive to do it manually if you really have to. Nor are resentful customers going to be much help in generating network effects. Data lock-in is not a competitive advantage that we will be using.
Lastly, we will use data effects as a competitive advantage. No startup worth its salt these days is without a slide on their investor pitch deck extolling the virtues of artificial intelligence. If you're not helping to make Terminators a reality, you're not in the game. As with most fashions, there is a kernel of sense. Let's use the term machine learning instead, as it is both more accurate and less dramatic. This is the use of modern high-volume data processing algorithms to generate useful outputs and behaviours. In particular, this includes previously hard problems in reading and sensing inputs, such as image and speech recognition. There's no magic here, and nothing to indicate we are any closer to AGI - Artificial General Intelligence (ie Terminators).
What there are, are very effective mathematical techniques for building prediction models from very large amounts of training data. If you have large amounts of data, you can build these models. If you have lots of users, you have lots of data. That's how Facebook can read your mind, and freakily show you ads for things you talked about yesterday.
You can see how this becomes a competitive advantage. You end up with valuable proprietary machine learning models that deliver very hard to replicate features. We'll aim to develop high-quality conference recommendations for our users, and make this a key feature of the system.
These then are the three ways we will compete: network effects from a growing user base, platform effects from third party integrations, and data effects from proprietary machine learning models. They are not the only strategies we could adopt, but they do fit the dynamics of our market, which is large, mature, and in need of a little disruption.
It would be a shame not to mention the GDPR regulation, which goes into effect this week. What effect has it had on voxgig? A slight decrease in our newsletter subscriber numbers, which are now 1,756, and reduction in our open rate, which is now 10.1pc. We're going to call this a 'GDPR Week' aberration, since everybody's inboxes are full of emails requesting permission to keep sending emails. Storms like these are just part of the game.
Richard Rodger is the founder of voxgig. He is a former co-founder of NearForm, a Waterford technology consultancy firm.