As a VC, I get asked fairly frequently what I think will be the next big tech trend or market, especially regarding software as a service (SaaS). Truthfully, SaaS has come a long way since I first started investing in what I called “XaaS” businesses about a decade ago. I’ve seen numerous trends come and go, and many supposedly hot markets fail to deliver much of anything. The one thing that’s certainly clear is that SaaS is here to stay — and it’s disruptive nature is only likely to gain momentum.
So, what’s next?
Clearly, SaaS companies will continue to improve the effectiveness and efficiency of certain industry verticals, business functions, and processes, or attack specific stages in the customer journey (prospecting, onboarding, customer success, etc.). And I also suspect more SaaS solutions for things like hiring and leveraging big data will continue to emerge. Beyond those trends, mobile SaaS has enormous potential (see OpenView’s recent Mobile B2B SaaS report), cloud computing remains ripe for SaaS innovation, and social will continue to get a lot of attention.
In this post, however, I’d like to focus on how the SaaS model might empower the development of services or products that replace or automate human brainpower, actions, and decisions.
“Hello, Scott — What Can I Help You With?”
By no means is Apple’s “Siri” a new idea or technology. It (or she) has been around since the debut of the iPhone 4S in 2011, and it (or she) has evolved and improved significantly since then. And Siri certainly isn’t unique. In fact, an argument could be made that Google’s machine learning/virtual assistant technologies are much more advanced and useful.
But that’s not the point.
The point is that, lately, I’ve found myself using Siri more and more to replace tasks or actions that used to be manual. With a simple prompt or voice command, I can now set an appointment, place a call, play music, type messages, analyze my schedule, and perform a search for virtually anything. Over time, Siri (and Google’s machine learning technologies) also begin to learn which apps I use most, which sports teams I follow, which contacts I’m most likely to communicate with at specific times of day, etc. As a result, there are times I don’t have to issue a command to get the information I need. It’s just there.
For an individual user like me or you, that kind of machine learning is valuable on a relatively small scale. But for big enterprises with massive amounts of untamed data, the opportunities that type of automation presents are massive.
Leveraging Machine Learning to Help Enterprises Harness Big Data
Increasingly, SaaS companies are providing more and more ways for enterprises to save, protect, and store away Big Data. What continues to be a challenge, however, is converting that data into something useful or meaningful.
This is where machine learning comes into play. From my view as a VC, really sharp SaaS entrepreneurs are exploring opportunities to apply machine learning to help enterprises harness Big Data, optimize specific processes, and clean up inefficiencies — but not by simply automating another task. They’re doing it by using machine learning to make decisions and take actions for us. For example:
- Machine learning and email: Many of us receive hundreds of emails per day, but only a small percentage of those messages are important. Still, to assess the importance of each message, we have to open, read, and do something with each message. With machine learning, a SaaS application could be built to intelligently decide which messages are actually relevant and how those messages should be handled, without the user ever having to do anything. Gmail’s Priority Inbox does this to a degree, but users must still do some manual filtering for the service to work well.
- Machine learning and prospecting: One new service — Persado — is addressing a piece of the email problem by helping enterprises deliver highly personalized, persuasive email messages, without the need for human brainpower or creativity. The company’s software does this by digesting huge amounts of customer data to identify the key phrases, language, emotions, and product features that are most likely to resonate with individual buyers. This allows Persado’s customers (which include Verizon and American Express) to build and deliver email campaigns without the typical creative fuss.
Ultimately, every SaaS product has opportunities to incorporate intelligence and machine learning, and most modern SaaS products have already succeeded in using some degree of intelligence to remove (or “disrupt”) traditional “paper-pushing” activities. CRMs and marketing automation are perfect examples of this form of automation, but examples also exist in customer service, product dev, etc.
The next step for SaaS, in my opinion, is finding a way to leverage machine learning to automate away decisions and actions that historically require human brainpower (instead of just deciding if an email is important, these SaaS applications would decide precisely how to respond, when to send it, and who to send it to).
This isn’t some sort of wild, futuristic capability. It already exists. And the SaaS companies that win the next decade will be the ones that find out how to best leverage it in their products.