I recently read the book “How Google Works” by Eric Schmidt and Jonathan Rosenberg. Additional to brilliant company and industry insides, they advertise the ease of product development in a digital economy. In short they see the availability of data and easy-to-rent compute power as key factors for every innovator to create new products. I agree, this has some serious advancements, but also raises the level of innovativeness and quality of products customers got used to and ask for. U.S. Tech firms deliver high standard online services and products which solve mostly basic problems with the help of technology (rent a car or apartment, find a good restaurant or your friends from high school, compare prices, sell old stuff easily etc.). Non-working or badly designed web-pages with technical difficulties lead to a bad customer experience and are therefore a no go. Additionally, modern tech firms base their business around platform markets (Spotify, Facebook, Google Search etc.). This implies you need to attract enough customers, but also sellers to your platform to become successful. There is a vast amount of literature on this topic in economics and management science (Rochet, Tirole 2003, 2006, Caillaud and Jullien 2003, Rysman 2004, or Andrei Haigu to name only a few), whereas Scholars deal with pricing, strategy and product design.
Adding the point of free software, like TensorFlow or Dash by Plotly, I explain the points Schmidt and Rosenberg make with the help of a little example. However, you still need the idea for the break-through product, which you can sell on a platform as a service 🙂
Give me some Data
It is true, the number of data collecting devices is increasing every day, but innovators need to obtain access to this information. Most APIs like Facebook’s or Google’s limit the number of API requests and their collected data. For the serious money business, you cannot dismiss the idea of not paying them for their services. Depending on your product, you then add some value to this data and further sell it to a customer. Two other options to earn money from data. Firstly, you draw traffic on your page/app and sell your impressions, or secondly collect as much as possible data by yourself and sell it afterwards. Either way, according to Schmidt and Rosenberg, you need to find a way to satisfy a basic customer need using technical insights.
It is clear however that this is a hen and egg problem. You can only sell data or impressions if you already have traffic. Your service needs to be extremely good to overcome this issue and gather momentum. Deep pockets for marketing are a huge plus, this is the reason why you need a business angel/investor.
Compute Power: Chose your weapons
In 2017 it is extremly easy, even as a private customer, to rent GPUs/CPUs or TPUs as well as storage to build prototypes and release your application for public use! Professional services by Google, Microsoft or Amazone make it very easy to to provide stable, reliable and most importantly scalable solutions to host your app. Load balancing even saves you costs compared to on-premise solutions (=buying your own servers). There are plenty examples like Earthquake Apps in Japan when holding expensive resources for a disaster is expensive, even if an earthquake occurs rarely. In this fashion your firm becomes manageable by only a few people, cutting out the bulk of IT managers and administrators. You still need experts on e.g. Google Cloud of course, but if the servers are down it is Google’s, and not your problem. Plus I cannot imagine a more professional environment when it comes to fail-safety, than Googles own infrastructure and team. This cost and firm size advantage is gold when it comes to trail and error, you can pull out of business at low fixed costs, plus you can shut down your servers immediately making decisions fast paced. This holds also true for enhancing your product portfolio. Additional services can be delivered by the same infrastructure as before, you simply pay some more servers.
Open Source: It is your turn
The Open Source Movement is huge in Silicon Valley. Even the big blue IBM offers their code for speech recognition and other Watson APIs. For you as an entrepreneur this is great news. People license their innovations mostly free of charge to any stranger. The value you create using this software or code belongs to you, how great is that?! In return companies have several advantages.
- They define the quasi standard in an industry. When they sell their services build up on their code, they have the experts right in place to do so.
- Other people debug, enhance, or improve your code in other ways.
- You gain some public awareness of your companies expertise. Think about TensorFlow, or Watson. These terms occur in news papers together with open source.
This concept is so successful, companies like Github exceed revenues of more than 200m $. For you however, this means a lot of work. You need to filter the libraries relevant to you. And do not forget to keep it manageable. Did I mention the cost savings compared to commercial software and the endless meetings you would have buying this?
Conclusion PD in the Digital Economy
Rosenberg and Schmidt make a great point when talking about the new product development. All the services named above reduce your costs to fail. Even small ideas can be set-up quite quickly without the need for many employees managing your data ware house, your servers etc. You have more time to focus on customer needs and further push your product. The downside is dependency, when Googles servers are down your product is, when the libraries you use are no longer maintained the backbone of your product may suffer. But on the other hand, this can also happen with commercial software.