A great portion of the literature on digital business models highlights Platform Business Models, which are therefore very often technical platforms where several different parties meet and add value to each other (in other words multi-sided business models). As platforms, it is also possible to imagine cities or even event arenas, where experience producers, service providers, customers and many other parties meet. It is important to understand when to talk about business models in the platform economy and when it comes to technical platforms that solve the challenges of the data and API economy, for an example. Technical platforms are very often considered to be all integration platforms, artificial intelligence and machine learning platforms, IoT platforms, data platforms, data harmonization platforms and nowadays also low-code and no-code platforms. Rarely can you meet all your business needs with one and often it wouldn’t even be the best solution.
In the data economy, technical platforms are very often used to collect data, build data models and provide or share data. Indeed, many know these by name or at least the brand of them. The most well-known global technical or service platforms are certainly Microsoft Azure or Amazon AWS. In addition to these, there are also many technical platforms that are popular in various industries, from which certain industry specifics, such as legislation, have been resolved centrally. Platforms can be Pay-as-you-Go, based on transaction pricing, or based on a fixed monthly fee, just to mention few main pricing strategies.
Let’s start by considering these first, whether it’s worth starting your own platform for your own use? According to many sources, the technical implementation of your own full-fledged platform will take about two years and cost up to 3 million €. In addition to this, more and more platforms are being created all the time, so is it all of the company’s core business to make one new platform among the others? Very often if the technical staff of a company is asked what kind of work it would be to build one integration on the platform of a global cloud service provider and the answer is that it is not big and takes a couple of weeks, for example. Soon you will have the next need for new API, then the next need and many more. After all, you’ve driven your company into a situation where you’re switching costs from self-made solutions to commercial ones are significant and over the times, you will leave behind the ones who use the most advanced commercial platforms. While the cost of doing it yourself was remarkably small, is it your core business, how do you make sure you don’t accumulate technical debt, and how do you make sure you have enough know-how for yourself throughout the life of your business? Many do this, so you get a better start based on your business needs and choose the right platform for you.
Should you then choose one player focused on your industry or someone else? Wherever the choice of one artificial intelligence solution or one data transfer standard makes you a vendor or technology lock, so does your choice of platform. Especially if the platform vendor has used some non-standards-based data model (e.g., self-made) or does not use universal and generic technical standards in all of its solutions, you are building these for yourself by using a higher replacement cost. What if, for example, you use an organization-wide platform for production facilities? Do you even have to customize your financial management and customer information systems because of the industry-specific solution? Perhaps a better solution would be to deploy at least two platforms, one so general that it works for your general data needs and the other then for those specific needs in your industry, with little customization. Another advantage of a two-platform strategy is that there are always at least two different options to implement business use cases. Of course, the best solution would be two generic platforms interconnected to serve your industry needs.
While we like the Pay-as-you-Go (PAYG) pricing model very much, it also has its downsides. The start-up costs are very tempting and on a small budget you can try it out, which is of course its purpose. At the same time, the cost of real business cases are very difficult to predict and rarely are these solutions completely transparent in pricing. Therefore, I constantly recommend keeping an eye on the capacity consumption, actual costs, and the development of these. Many companies today provide an optimization service for platforms like this precisely because inexperienced hands can easily build expensive technical configurations, leaving costs out of hand.
From these three starting points, when you start thinking about your platform strategy and infrastructure so that management understands, even at the top level, what kind of platform architecture is being formed for you, then you are already far from speaking the same language.