The first time I used Waze was a revelation. This was not because of the network effects it generates, but more importantly, because of how transparently the application’s user interface (UI) exposes them.
This means every time I open the Waze app, I’m implicitly participating in the system. In doing so, I help improve every other user’s driving experience, and as an added bonus, all the other Waze users participating around me improves mine, too.
This example shows that when you’re an active participant in the learning loop, everyone benefits – in this case, because everyone reaches their destination faster than we could by ourselves.
Two important things have changed in the last few years to accelerate both network effects and the feedback loops that they support. The first was the commoditization of what’s commonly referred to as “machine learning” technology.
The second major factor was the increased transparency around the intrinsic relationship inherent between using a piece of software and contributing to it through your behaviors, much like the Waze example we’ve already discussed. ‘
Together, these two trends will not only coexist but become increasingly intertwined as they power the next generation of technology: learning loops.
I know what you’re thinking: “what the heck is a learning loop, exactly?” This is of course an excellent question, and fortunately one with a relatively simple answer. You see, learning loops combine machine learning with large scale data sets provided through social networks and human behavior.
What you, the end user, provides in terms of social data has nothing to do with your status, your photos, or the books on your reading list. Instead, you’re providing a direct feed of your behavior in real time, all the time.
The idea is that in learning loops, everyone in a given network will collectively benefit from the programmatic experience and analytic insights generated by everyone else in the network.
Just like Waze is designed to get me to my destination faster in exchange for me allowing it to track my location in order to do so, learning loops in general are designed to generate better results for everyone than what any individual could ever get on their own.
This makes learning loops faster than previous technologies – a whole lot faster, in fact. Within the new wave of apps built around this core concept of learning loops, the lag time is not just zero – in most instances, there is in fact no lag time; they’re predicting what’s next, not responding to what’s now.
And if you think about it, that’s pretty profound: learning loops can not only react to behavior, but anticipate it, too.
A quick look at the consumer software landscape reveals the increasing ubiquity of learning loops, which have more or less taken over the category – Waze, Yelp, Tinder, Netflix, Amazon, Facebook, you name it, if it’s tech, its business model was likely built on the basis of leveraging machine learning based on user generated information to provide better experiences for those users.
Whether you’re figuring out what book to read, what music to listen to, who to date or which restaurant to eat at, someone has conveniently created some form of learning loop to support you.
The experience you’ll get as an end user is completely unique, and yet, it’s been built upon millions of data points collected and captured from previous users who are more or less just like you. Which is pretty cool, right?
The only thing is, you’ve got to live up to your side of the bargain: to use the product and contribute your individual data to the greater good of the network and the learning loop this information creates, in aggregate, at least. In fact, ask any VC firm how many consumer tech products they’ve funded in the past year that don’t have some sort of learning loop, you’ll see how pervasive this phenomenon has truly become for the consumer market.
When it comes to enterprise software, however, things are just getting going when it comes to machine learning.
As this concept steadily takes root, as is already the case at many forward-thinking enterprises and companies on the cutting edge of what’s new and what’s next, this new category of software presents the promise of not only transforming an organization’s ability to compete (and win) when it comes to their business and bottom line, but to do so seemingly overnight.
All For You.
One business segment where learning loops have historically been highly operationalized is in security and fraud detection, where they’ve become a pervasive and powerful tool in companies’ abilities to control and protect propriatary networks and data.
Consider the case of Area 1 Security, an innovative cybersecurity startup in which network participants send a constant stream of sensor data back to the core platform, which, in turn, uses this information to enhance its capabilities and more effectively deter security risks for everyone in the network.
Anyone who participates directly benefits from the experience of everyone leveraging that learning loop (and the more participants, the more data generated, and the better the app works. It’s a pretty simple, pretty cool phenomenon, and one that’s becoming a huge part of our everyday lives and interactions.
While this seems like a no brainer, remember that only a few years ago, for any company to share any sort of data about what happens within their firewall to anyone outside it was a seriously big deal, and a relatively rare phenomenon. This was seen as information that needed to be protected, propriatary information that had the potential (it was thought) to put both the bigger business and its employees at risk.
A few enterprises, however, had the foresight to recognize that a closed network, particularly when it comes to data privacy and information security, provides far less security detection and protection capabilities individually than can be achieved by benevolently (and, selfishly, given the increased efficacy) being a part of a network, like Area 1, which is powered on the premise of learning loops and the economics of the economy of scale.
In a totally different domain, we see analogous effects in the application of learning loops to recruiting and hiring. For example, in almost every multi-tenant, talent related SaaS instance, companies are contributing their unique hiring data to their recruiting systems or talent solutions providers, who are in turn starting to use this data to look at anonymized information such as job posting and recruitment marketing performance, applicant stats and recruiting baselines like time-to-fill and cost per hire.
In turn, these talent-specific learning loops provide companies with a predictive engine that looks at massive amounts of historical data collected from the broader network and gives them concrete guidance that’s proven to be successful in filling roles with increasingly diverse and better qualified candidates than competitors who don’t use similarly structured software solutions (and by a pretty significant amount, too).
As in the cybersecurity example, the capability gap between those participating in learning loops and those using their own data to act independently is only widening every month as these networks’ data sets continue to grow at a rate far faster than any individual company or enterprise could ever conceive of generating by themselves.
In both cybersecurity and recruiting, companies in the learning loop realize concrete advantages that have a tangible impact on their competitive standing, often almost immediately after initial adoption.
We’re talking days, in many cases. The days of waiting three years for a version update are long gone; learning loops are exponentially faster, and getting faster and more efficient as each network gets larger. This means as learning loops get exponentially more effective, the impact of not participating in these networks continues to get exponentially more painful for those outside of it.
If you take a step back and look beyond security or recruiting, you’ll quickly realize that pretty much every other part of your business is likely ripe for upending by learning loops.
Finance, for example, can now leverage software that considers your previous spending patterns in the context of your coworkers and competitors and tells you how to budget. A CRM can tell you when the optimal time to pitch a lead might be based on when in the fiscal companies with similar profiles generally purchase products or services like yours.
These opportunities are one opportunity cost no company can afford to pass up, period. So, here’s a brief lesson plan on how to make learning loops loop in with your policies, processes and people, and how this emerging category is one platform every business will need to stand on if they want to survive (and thrive) in the ever changing world of work.
Now, I’d like to point out that learning loops already integrated into our consumer software today goes way beyond discovering some cool band or figuring out what movie you’re in the mood to watch or which online match is going to be the most compatible in a long term relationship.
No, that’s just the beginning – and as big as the impact of learning loops has been on consumers, the social implications of this revolution as it takes hold of enterprise software are equally (if not even more) profound.
You see, if you’re in the learning loop – any learning loop – you’re going to hire, sell, market or build better, more beautiful and more profitable things than anyone else outside the loop ever can.
That means that everyone inside the learning loop wins, because they all benefit from the collective experience generated from an entire community, whereas anyone outside the loop is going to lose, because they’re using data that’s just too limited to effectively compete with the network effect affected by these networks.
If you’re building or developing your own learning loop, know that formerly prohibitive obstacles like machine learning technology and artificial intelligence are no longer barriers to success; similarly, they’re also no longer competitive differentiators, either. Machine learning has become increasingly commoditized as bigger technology companies are trying to enter and compete in a category that, while it’s the future, very few seem to have any capabilities around at the present.
This has turned the whole conversation and concept, sadly, into something of an amorphous buzzword or tired cliche, which is too bad, considering it’s one of the most exciting technological developments we’ve seen since the invention of the internet – and every bit as disruptive.
With machine learning, business as usual is anything but – and so too is the concept of success.
Any Time, Any Place.
With this new revolution in technology, winning means applying learning loops and similar machine learning models to new and novel combinations of social data, which can only be generated if users within a network’s learning loop are provided with an experience – and results – which clearly and unambiguously outperform their expectations by far surpassing any outcome they could conceivably generate on their own.
It’s a basic precept of human behavior that we work better together than individually, and across domains as diverse as sociology or software engineering, our collective experience consistently solves problems better than a single brain working in the absence of experience or the vacuum of isolation.
I know when I’m trying to get home during rush hour, or how my job ads sound to my target candidates, for example, learning loops provide infinitely more insight that takes me vastly further than I could ever get on my own, and that data is predictive, not prescriptive, something consumer products have already largely embraced.
For enterprise recruiting and HR technology, the writing’s already on the wall – and if you choose to keep your business outside the learning loop, there’s a good chance you’re going to be going out of business. No one can stay viable by staying disconnected anymore.
In the world of work today – and tomorrow – winning at business is no longer about individual competition, but interpersonal collaboration. If there’s one lesson no business can ignore, it’s that learning loops are the future of enterprise technology, and, my friends, the future is right now.
And you’ve got to admit, it’s pretty cool.
Read more at the Textio Word Nerd Blog.
About the Author: Kieran Snyder is the co-founder and CEO of Textio, a recruiting technology startup based in Seattle. Kieran holds a PhD in linguistics and has held product and design leadership roles at Microsoft and Amazon. She has authored several studies on language, technology, and document bias.
Kieran earned her doctorate in linguistics and cognitive science from the University of Pennsylvania and has published original research on gender bias in performance reviews and conversational interruptions in the workplace over the last year.
She participates actively in Seattle-based STEM education initiatives and women in technology advocacy groups.
CEO and Co-Founder of Textio. Long-time software product leader, accomplished data writer, recovering academic with a PhD in Linguistics and Cognitive Science from the University of Pennsylvania. Deep experience in product management, product marketing, pricing and licensing, and SaaS across the board, with specific technical strengths in natural language processing and data science.
Weekly news and industry insights delivered straight to your inbox.