Machine learning is under the spot light for investors, strategic, tactic or commercial. While at the same time, you may not feel so empowered to make an investment decision because it is an early-stage industry with many changes and risks involved. And especially if you are an angel investor, maybe the guys in the garage will create the next Facebook or Google, maybe not, but how can i tell? So what are the questions I need to figure out when investing in machine learning and what tips I can use to evaluate a company?
The first key question you need to figure out is really what do you want to use the investment for. Are you making the investment so you can
- Use its product in your company and/or serve your clients
- Eliminate a competitor and improve your competitiveness
- Obtain a constant cash flow for a certain period
- Exit with a good return later on
- Acquire other benefits you want
If the investment is fundamentally for self-consumption. Then the focus is on how the target’s technology implementation is tailored towards your intended use. Ford recently announced they will invest US$1 billion into Argo AI to boost its capability on self-driving car. Saleforce.com acquired MetaMind AI which using an artificial intelligence technique called "deep learning" to help businesses crunch through data and make better decisions to help its clients make better decision. In the case of strategic acquisition or investment, readiness of the target technology to be transformed to end product is the most important criteria.
Joining forces to target the same market is also common reason of strategic investment or acquisition. IBM acquired deep learning startup AlchemyAPI and consolidated its offering into the Watson Developer Cloud. Basically it is about the resource consolidation, while IBM obtain new capabilities for making inferences on images and text using a form of trendy artificial intelligence called deep learning, Alchemy API are able to reach a large developer and consumer base through the Watson platform.
It is less intuitive but nevertheless common that an acquisition is made to generate a cash flow in the future. For example, Google’s acquisition of DeepMind and the subsequent promotion strategy towards DeepMind (AlphaGo bets Lee Sedol in the Go game) indicates its intention to develop the DeepMind product and service line and thus create revenue stream in the future.
Sill another group that invest in machine learning area look for an exit with good returns. This group do not know knows all the sophisticated models and algorithms on machine learning. So what they are trying to do is instead fall back to their comfortable areas which is market size, consumer sentiment etc.This group can utilize 3 simple techniques I discussed in … to identify good companies and teams without knowledge the details the industry.
And finally, there is also investment that were made to obtain other benefits. Many investment to university science parks and government science labs are primarily based on these reasons, to be more favorable in front of the government so the local tax (and other) bureau can be more favorable to them, or to help create “achievement“ for the local government in exchange for cheaper land; favorable lease terms etc.
You may think, the list looks like it is for big corporates only, not for me. But it is not true, especially for small and medium sized company that face labor cost pressure, expertise gap, finding specific machine learning technologies that can help your business is very meaningful for the longer term.
Regardless what the reason was that you start to consider investing in machine learning, there are 5 things in common for any promising machine learning startups:
Team can Explain its Business in an Easy Way: being about to explain one’s business in common language is the fundamental and important criteria. One company, no matter how sophisticated it is, cannot live. While working with clients, developer community, and also your own staff, the team cannot keep using terms and logics nobody understands, they need to translate the expertise in layman terms and probably even set the strategy how it will be rolled out. Test it on your team first, see if they get the point.
Team can Explain its Technology and Risks in an Easy Way: personally, I believe in the demystify and extraction of all technologies into common language and even translated into the other industry terminologies. After all, our industries originally formed as a group of people who shared common traits and thus terms developed to simplify a set of common work flow ; to identify a unique tool developed to accomplish a semi-common task and to create belonging.
Company is Up-to-date on the Market Trends: they do not need to do what it hottest, but they should know what is hot and give a good reason why others should or should not use their product but not the competitors. Knowing which battle to fight is a fundamental knowledge for a startup as it means the company is not stiff.
The Company has an Executable Strategy to Obtain Talents; they need to be familiar with where the talents are coming from, have contacts with some of them and have a strategy to acquire the talents and retain them. The talents can be from universities or labs or even freelancer, it can even be public services that can achieve the same purpose (for example, IBM, Microsoft etc all provides extensive public services for image recognition and natural language understanding and sentiment analysis) as long as the team can effectively evaluate the service and tweak it to their specific purposes.
Founders Have a Mature Relationship with Their Ideas: be passionate but not to be ideal; be flexible to work with other ventures or not but have a valid idea; know their value on the market and be open to transformation.
Financial investors look at the companies from a different perspective, you care on the market direction, the consumer segment, the team’s professional knowledge and also many times, how you can help the venture through bringing sales lead, attracting additional investment, but eventually, how to exit. When the deal is on the table, many of the times it is a good time to think about if we should exit and transform, and having a mature relationship with the company play an important role on that.
For some of you, the personal connection with the founders are also important. It helps avoid future conflicts when the time comes for additional investment, or when you plan to cash out and exit in some way.
Machine learning is absolutely a popular topic, more and more venture companies try to associate themselves with this concept. The industry starts to have a mix of serious players and trojans. It is guaranteed many of the investment will not have a good return, while on the other hand, it is also considered as one industry that will produce the next Facebook, Google etc.
At the end, I want to quote what Nathan Benaich stated in his article “Investing in Artificial Intelligence”: the open sourcing of technologies by large incumbents (Google, Microsoft, Intel, IBM) and the range of companies productizing technologies for cheap means that technical barriers are eroding fast. What ends up moving the needle are proprietary data access/creation, experienced talent and addictive products.