We all know the subscription model is invading the world. We subscribe to everything these days because that’s much easier than paying for each event, activity or transaction. We have subscription businesses in B2C as well as B2B. We have tons of examples for both. Some of the popular ones are Spotify and Netflix in B2C, Slack and Intercom in B2B.
In case of B2B particularly, the widespread thinking has been around customer engagement. Everyone wants to know if your customers are coming to your dashboard regularly or not. How much time are they spending there? What can you do to increase the time in-app because that’s tied to the outcome for the customer and you?
It’s all about to change with machine learning invading the world of SaaS. Most SaaS founders I have spoken to across India, Canada and the US in the last few months think of machine learning as a new layer they can add on top of their product and start generating some extra insights about their customers. Nobody I’ve spoken to recently seems to understand that their very product needs to be rethought because human expectations from software and machines are about to take a big leap.
Your customers use iPhones and Android phones. Both of them have AI assistants built-in so that some tasks can be passively executed without you being involved. Every year, Apple and Google announce newer features to help you get more from your smartphones by doing less. That’s affecting people’s expectations in more ways than you can imagine.
I believe machine learning is still in early stages but people have figured out interesting ways to solve tons of narrow problems using new deep learning and deep reinforcement learning technologies, now that both compute and data is available to solve problems acrosss industries. Technology is evolving slower than the adoption of the use cases powered by it. I use Google Home at home and I can only wish it was a bit smarter. It doesn’t stop me from using it every day.
I’m expecting a similar kind of shift is about to hit the world of B2B SaaS. The users of your software are the same humans as the rest of us and they use the same smartphones, and in most cases, a very similar set of consumer apps. Maybe you have been optimizing for more engagement all this while for your SaaS software, but they must be wishing if they could skip a day or two and still got the job done. Let me give you an example.
Let’s consider an email marketing app that your customers use to send and measure the performance of their email campaigns. In a typical scenario, a marketer would come to the app every other day, create a few emails, schedule and then send. He knows his success rates with his campaigns and he has been doing it for a year now. He makes basic customizations based on the type of customers, geography, and timing. What if the app stopped requiring him to come and do all this every day? What if he could just tell the system his expectations and basic template and then ask it to make plans and customizations based on what it learned from you (and maybe from users like you)? What if the marketer used the app only to know analytics — which could be sent to her automatically and regularly through SMS, email or whatever medium she prefers— and didn’t log in for a month in a row but still got results? You will probably have zero engagement and probably more perceived value prop. Would they consider your SaaS product useful then? Of course, yes!
You will have to reimagine the success metrics for your SaaS software once you take this route. Instead of trying to increase how many times a customer is coming to your dashboard, you will try to reduce that. You will do the same for the time they spend in your dashboard per session. And so on. As you may have noticed by now, your product and growth teams would need a bit of rewiring of their brains to adjust to this paradism shift in building and growing SaaS products. Depending on how you have defined your value metric, your business model may require a bit of tuning too.
The thing with disruption is that it impacts everyone unprepared. I see the overall SaaS community is figuring out vertical SaaS and similar ideas to see if large companies can still be built in this model as many of the well-known problems have either been solved or are being solved by existing SaaS giants. My guess is that there is still an enormous opportunity to solve specific problems across industries in a better way using a better design, technology or distribution model. You can pretty much disrupt and eat the lunch of any large SaaS company by taking a self-serve approach to product and growth because most of the big companies have removed their pricing pages and have gone upmarket with a mostly sales-driven model of customer acquisition. While you are at it, give a thought to the impending arrival of Invisible SaaS which can catch you completely off guard.
We at Artifacia are making a bold bet on the Invisible SaaS model as part of our Artifacia 2.0 vision so that we can drive even more value for our existing and future customers using our machine learning technology that we have been building for the last 3 years. As a pioneer of the model, we will have to figure out a lot of things to make it work for our business without compromising on user experience and business growth. As we learn more and build a framework of execution over the next few months or years, we would love to share it with the wider community. In the meanwhile, I’m open to helping other product startups figure out their machine learning strategy in whatever capacity I can.
P.S. I started my career as an ML Engineer/ Data Scientist before taking the leap as a startup founder in 2015. I learned SaaS only last year as we pivoted our startup from an API to a SaaS delivery model with the same underlying machine learning technology as the foundation. It’s interesting how one can connect the dots only looking backwards!
P.S.S. We are hiring aggressively across product, marketing and machine learning for our Bangalore and Toronto offices. Write to careers@artifacia.com, if you are interested.
Comentarios recientes