While there are still technologies that are in the innovator or early adopter phase of the typical technology adoption life cycle (VR and bitcoin as examples), machine learning and AI are clearly in the early majority phase. And it’s very possible that they will reach the late majority even before some of those other technologies reach the early majority!
It won’t be long before the use of machine learning and AI is assumed for every application in every industry.
Already, VCs have a hard time taking software pitches seriously if they don’t include ML/AI as an underpinning. In the same way that SaaS is now recognized as a delivery mechanism and a business model shift, AI and machine learning will be recognized as enabling technologies that make every application (and every user of an application) smarter and more capable.
In the same way that there’s nothing proprietary about SaaS, we believe that over the long term there will be virtually nothing proprietary about the software that enables AI/ML. Already, Google, Amazon, Microsoft, IBM and many others make their ML frameworks readily available. And that’s because they recognize that the adoption of ML in software is largely to their benefit.
If all software relies on ML, there is one asset that becomes critical to the success of that software ─ data. The incumbent players enjoy a clear and significant data advantage. If you think you’re going to compete against Google, Facebook, Amazon or Apple in areas where they have deep data, good luck to you. And more than likely, good luck raising venture money, too.
So where’s the opportunity for startups? At Homebrew, we’ve focused on ML/AI investing in companies that have two key elements: 1) data that is largely hidden from the incumbents and 2) cost effective, scalable win-the-market strategies (plain old go-to-market ain’t good enough!). Our view is the combination of these two elements leads to long term advantage and a competitive moat that protects against both the incumbents and other startups that enter the market.
Finding data pools that are black boxes to incumbents usually means targeting verticals or use cases where the larger players don’t have access, experience or focus.
For example, pulseData works with healthcare providers to access medical records data, patients outcomes data and insurance claims data to predict and prevent acute health events amongst the providers’ patient populations. The company’s data is not only unavailable to other companies, it also requires the difficult work of aggregating, cleaning and normalizing the data, which comes from many sources and in many forms.
Hickory actually goes so far to create the data it needs for its ML models. Through its modularized training app, the company assesses how well customer service reps know the information that is needed for their jobs. And then it uses that data to build models to predict when each rep is likely to have forgotten specific pieces of information so that the right training lessons can be resurfaced at the right time to the right rep.
pulseData and Hickory are both working with data that doesn’t exist within the domains of the tech behemoths and isn’t easily accessible to them given that their priorities lie elsewhere.
But as you may have noticed, both companies depend upon getting access to data from their customers. Which is why we believe that win-the-market strategies are the second component of success for ML/AI software companies.
We think of winning the market as opposed to simply going to market because the latter assumes that just selling is important. We believe that selling is necessary, but not sufficient, for winning, especially in ML/AI-driven markets. As important as selling is delivering customer success.
In fact, we’d go so far as to say that in most software markets it’s not the best product that wins, it’s the best “sticky distribution”. When we evaluate win-the-market strategies we look at three specific components to gauge the potential for success.
The first is whether the company is addressing an acute pain point, what we often refer to as a “hair on fire” problem. If your hair’s on fire, you’re probably going to put that fire out before you do anything else! In the same way, if you’re a business customer that is dealing with many different challenges (and being pitched software solutions all day!), you’re going to focus first on the most urgent need and the potential solution to that problem.
The second component is whether the software requires a change in user behavior. Getting people to do anything different than what they’re doing today is hard. And that’s even more true in a business context when several people or teams may be involved in a specific workflow or process.
Further, it’s much easier to access a budget that’s already created for a specific problem than to create a budget for something that seems new. Fitting into a potential customer’s existing business and budget makes for a simpler sales process and a shorter sales cycle.
The third component is whether the software will make the customer a hero. It’s one thing to get someone to buy your software, it’s another thing for her to be successful using it. The power of ML/AI is it can help give customers speed, accuracy and insight that meaningfully impacts their businesses.