Remember digital transformation? How “software was eating the world”? If you are reading this, you are probably painfully aware that we have been (and you likely participated in) spending our time building millions of applications over the last 20 years to digitise our businesses and change how we interact with the customers.
And we’re about to rewrite them all over again.
Over the past few years there have been true breakthroughs in using data science to create predictive and autonomous decisions and experiences. Machine learning and AI (deep learning in this context) have begun to permeate our world. From the mundane (Netflix’s unfailing ability to recommend the next show you’ll love) to the sublime (Google and Amazon assistants opening up entirely new ways to interact with the world), leaders in every industry have embraced using data and models to fuel their businesses. The implication? We will spend the next 20 years rewriting every application and reinventing every process, or risk being a relic of the past, just as Blockbuster, Kmart, and countless others were. Digital transformation keeps rolling, and now it’s a data revolution.
What this means for developers
This scenario raises some critical questions that every enterprise should be asking itself as AI and the model-driven applications it powers expand rapidly. This “app transformation narrative” will be driven by answers to questions such as: What does this mean for your technology portfolio? What does this mean for how you build applications? What are the changes ahead for the ways in which you develop, deploy and manage applications across your enterprise? What does it mean for the most critical asset in this – data? What will be required to bring these new applications to life and operate them at scale?
If you look at the innovation and resulting disruption that's occurred across industries in just the last five to 10 years, what you consistently see is that the organisations that are starting to transform their categories all have one thing in common: they are leveraging AI and ML to turn their data into fuel for model-driven applications that, in turn, automate their businesses and/or unlock entirely new products and services. In simplest terms; model-driven applications are rapidly replacing logic-driven applications.
Extrapolating this out a bit further, these breakout organisations are leveraging data to not just understand what's happening in real time and react instantly, but to also predict the future and automatically learn and tune their products and business to respond to a changing market. This trend will only continue.
Fare thee well, digital transformation
What we’re experiencing now – industry by industry, category by category – is a true data revolution. And, yes, I mean “revolution,” not “evolution.” The same approach that drove the breakthroughs in digital transformation is now driving breakthroughs in data, powered by the growing, AI-powered wave of model-driven applications. It’s also starting to separate winners from losers, just like digital transformation did.
“So, how did we get here?” That’s a question I’ve heard numerous times in some way, shape or form with regard to the data revolution in our midst. Here’s how I answer it:
In the last 20 years, we automated our businesses and transformed our product and customer experiences to leverage broadband and ubiquitous devices. And we did all of these things by building millions of applications, which changed how we interacted with each other – be it colleagues, customers, partners or vendors.
Those applications were all built on decision trees and described logic. We automated processes based on the established protocols, behaviour and rules that we determined to be most effective. But the next wave? That is now being built on models.
The truly transformational businesses of today have taken the mindset and the lessons we learned from digital transformation to look forward into the near-term future. From what I’ve seen and heard, these are the critical questions they’re asking – and answering – about model-driven applications instead of logic-driven applications:
- What would I have done if I didn’t have to take a best guess at what the logic or the experience should be?
- What happens if I then have a model that can improve on itself (machine learning) attached to that decision?
- What happens if I can use all of the available data to accurately predict outcomes?
- What happens if I can leverage AI to replace or augment human interaction to do some of those things?
- What, then, would my business look like?
If you’re thinking Amazon and Netflix...you are on the right path
The digital revolution leaders have already answered these questions and implemented solutions at scale. Think Amazon’s personalisation engine or Netflix's recommendation engine; these are some of the most basic model-driven applications we’re all familiar with (and now take for granted). Said simply, they are machine learning models that are designed to remove the friction from your personal shopping experience. They are, relatively speaking, early examples of ML either replacing or augmenting a once-human task.
Now, let’s take that full circle. With the data generated by these models and the resulting application usage, we can start to predict what products people will buy. From there, you can use those predictive analytics to actually fill out your supply chain. Supply chain for Amazon is what they stock in the warehouse, and supply chain for Netflix is what shows they produce and license.
Another example that many of us are probably a little too familiar with is Uber Eats (or Postmates, DoorDash or others like them). Supply chain for Uber Eats is what food is searched for, browsed and sold via their app. On the user interface, Uber Eats is basically saying: “Here's what's available. What do you want? I'll go get it for you." But at the same time, Uber Eats is going to restaurants and saying: "We believe, based on our models, that there is a demand for this kind of a product. And, if you change your menu and produce it, this is how much money you can make. Will you start selling it?"
Innovation and customer demand are leading the charge
What's driving the change toward a data revolution is the collision of entrepreneurial innovation and ruthless customer expectation. If you think about this from a customer perspective, what all of us now expect is a better product, the best possible experience, and we want it RIGHT NOW. Apple, Google, Netflix, and countless other data-fuelled brands taught us that. Netflix didn't initially “win” the streaming market (though their competition is much fiercer now, of course) because they were digital. They won the market because they actually were able to produce better content and push the content you would most likely love directly to you. That’s the kind of transformation we are undergoing right now.
The last thing I’ll add is that this transformation is inevitable. Yes – it’s that stark, though not without hope. In fact, it’s only getting more possible to become a successful participant in the data revolution because the models, the data tools, and the surrounding infrastructure to unlock the potential of ML and AI is available to every business right now, and they’re only getting more powerful every day. Time to get busy.
Peter Guagenti, CMO, MemSQL