Want to launch a platform? It’s all in the mix

One of the wake-up moments for me,” reveals London Business School Associate Professor of Strategy and Entrepreneurship Gary Dushnitsky, “was when I realised that, if you want to launch a platform today, there are many ready-made vehicles out there. Not so long ago, writing the code was a critical element, but now there are white-label providers who offer ready-made websites. The real challenge in launching a successful platform now is understanding that the mix of strategy options is key to unlocking the flywheel of indirect network effects.”

Investigating the mix of strategic choices and performance of transaction platforms: Evidence from the crowdfunding setting, co-authored with Evila Piva and Cristina Rossi-Lamastra of the Milan School of Management, looks at crowdfunding platforms from the start of the industry to 2018 (a population of 788 crowdfunding platforms in EU-15 countries).

Gary reveals a direct professional motivation for the research: “Pretty much every student at LBS doing an entrepreneurial course now is launching a platform – a dating platform, a delivery platform, an investment platform – yet most of the academic studies tend to focus on the large platforms. What does Amazon do? What does Tencent do? There are few large-population studies that look at how to start a platform in the first place.”

Pricing strategies

The researchers therefore set out to answer the question “at the top of almost every student’s agenda these days”: What strategic choices should your platform undertake? Underpinning this enquiry is how the mix of different strategic choices unleashes the indirect network effects. When Gary asks his students how they would grow a platform, they invariably cite pricing strategies – the need to subsidise vendors to attract them first, which then attracts customers, or suggest subsidising customers to attract them first, so that the vendors follow.

The literature has much to say on pricing strategies, particularly with regard to fee allocation. “Theory holds that, if one side is more important to grow the marketplace, that’s the side you subsidise, so you allocate all or most of the fees to the less-critical side,” Gary explains. “This kickstarts the flywheel to stimulate the indirect network effects.”

These are crucial insights, but the researchers wanted to build on them by examining the available mix of all strategic options. Gary says: “To truly understand how to grow multi-sided platforms successfully, you need to realise that you have more than just one or two strategies at your disposal.” The paper actually identifies six different strategies, considers how they are combined, and analyses how the mix affects performance.

Three non-pricing options

Whereas business schools tend to focus on pricing strategies, non-pricing strategies may be more available and appropriate for startups. There are three major choices here: level of accessibility (variety of the participant pool); degree of inclusivity (scope of offerings that the platform accommodates); and bundling (additional services provided by the platform).

Variety, in terms of the participant pool, determines who will join. Do you only run your platform in Italian, or in French? Or do you translate it into multiple languages? Gary notes that a lot of studies focus on the US and look at platforms where English is the only language option but, anecdotally (it wasn’t part of the study) many successful platforms in the US are bilingual and run in English and Spanish.

The paper also found that there are many different ways in which you can control how many different people will join the platform; for example, “you could mandate that only accredited investors can participate.” Another set of decisions concerns participant offerings. In this dimension, a platform might allow people to trade only 18th-century French furniture, or any furniture, or all goods.

The third major non-pricing strategy option, platform services, is about the extent to which the platform should operate as a ‘barebones’ marketplace, as opposed to providing add-on services, such as shipping or certification of authenticity.

Six strategic dimensions

The paper also looked at the three pricing strategies that have been studied extensively in the literature: subscription, transaction and fee allocation.

Subscription relates to how much people are charged to get on the platform; transaction is how much they pay per transaction once they’re on it; and fee allocation is the subsidy element – how does the platform allocate fees between the different sides of the marketplace?

Identifying these strategies is not new. What is new is analysing their combination in this way, which reveals the six different dimensions that the platform needs to take into account. For example, it might be more prudent in terms of cost to increase platform scope by widening participant variety – simply translating into English overnight may be a better strategic option than to start playing around with the fees structure, which may take longer to take effect and will inevitably cost money because the platform is either subsidising one side of the marketplace or spending money on social media.

Go wide versus go deep

In theory, based on the strategy choices that can be combined in any way, a huge number of mixes is available. (The example in the study actually looked at all six dimensions; assuming each can have three different levels gives 6^3 =729 choices.)

In practice, however, the researchers found that crowdfunding platforms tend to cluster into three strategy mixes. This suggests that platform pricing and non-pricing strategic choices are interrelated. The question then arises, are those options chosen randomly, or is there a way to mix strategies to optimum effect?

The researchers found there were two dominant strategies: ‘go wide’ and ‘go deep’. A go-wide strategy is mostly seen in financially-oriented platforms, such as equity and lending platforms, because the rationale here is to allow anything on the platform that has a financial basis; hence a broad variety of offerings is appropriate in a go-wide approach.

At the same time, because such platforms need to deter entrepreneurs whose business ideas are unlikely to attract funding, they use pricing strategy to levy a significant upfront fee on both sides, thereby screening out the less serious funders.

This approach is very different from what the researchers found on donation- and reward-based platforms, where a go-deep strategy was more prevalent. “These platforms are driven by communities of interest,” explains Gary. “People are interested in a certain topic, so you narrow the range of offerings; be it about antiques or sci-fi books or a region in Italy.” Here, platforms charge a relatively low subscription fee because they want people to join and list offerings and other people to browse what’s available, so platforms want users who are motivated by a genuine interest in a specialist field. And from those interested parties, the platform can charge a higher transaction fee.

There are also all-embracing platforms, which try to do everything all at once. These support many different types of participant and different types of project and offer multiple platform services.

The conventional thinking about sectors with strong indirect network effects is that the winner takes all – more vendors come, so more customers come, so more vendors come and so on, until you eventually attract all the participants away from all your competitors, with the result that one platform is home to everyone. “Why would you go on platform.com if everyone is on amazon.com?” Gary asks. In fact, during the time period of the study, 58% of platforms remained active and only 39% dissolved, suggesting that there wasn’t a strong winner-takes-all dynamic during the period studied, although this may have changed in recent years.

Key takeaways

The key takeaway for founders and investors is that the six strategies that are available to startups from the earliest stages need to be explored. “Play around with these levers,” Gary advises. “Don’t just focus on pricing strategies. Nor should you treat the levers as independent from each other – see how you can combine them into two distinct strategy outcomes in terms of go wide versus go deep.”

For a broader audience, the size of the study is significant. While much has been written about the big, successful platforms, research is lacking, the paper argues, if it doesn’t consider the full population – including those marketplaces that were launched but failed. As Gary says, “The study used a large population over a long period of time – two decades – which is unique in itself, and included those platforms that had failed, which brought other useful insights.”