Whenever one of the tech giants stage an event for one of their announcements there is a degree of hyperbole that we have come to expect from these companies. They are, after all, showing us what they believe to be the next big thing in the industry.
But in among Google's I/O 2017 developer conference last month there was a piece of information slipped into presentation about machine learning by CEO Sundar Pichai that struck me as monumental, despite the understated way in which it was delivered.
“We are excited about creating better machine learning models, but it’s time consuming. We want developers to get to use machine learning, so what better way to better our machine learning than to get the neural nets to design better neural nets…it’s learning to learn.”
Learning to learn. Machines learning how to learn better ways of doing things themselves? This is pretty wild.
But it got me thinking about the importance of learning to learn, and how often it is overlooked as a step towards change.
Digital transformation is a hot topic in the lives of many clients today, and that’s great but what if you haven’t gone through the process of learning to learn how to transform?
On a recent episode of Harvard Business Review's HBR IdeaCast podcast, Microsoft was held up as a great example of a business that had a huge opportunity to transform but in this particular instance, failed spectacularly.
Back when online advertising was in its infancy, Microsoft had everything in place to own the paid-for search market. They had made acquisitions, they had browser dominance, and their MSN portal had considerable market share.
And so they rolled out their solution, a series of pilots to test the efficacy of their solution and it failed.
What transpired was that there were many across the business who didn’t believe in what they were trying to achieve, who felt that their business was software and that was that, and the core business subtlety plotted for the failure of the experiment and it was killed.
And Google won. Microsoft hadn’t learnt how to learn, it assumed that people would just jump on board. The company's leaders had failed to notice a lack of enthusiasm across core parts of the business, which destined the project to be mothballed.
Listening to a panel discussion about modern marketing last week led by some of the UK’s most prominent CMOs, from Paul Davies at Microsoft to Jennelle Tilling at KFC, there was a unified desire for agencies to help their brands prosper at a time of tumultuous change in the industry.
And a universal, thundering warning that agencies simply weren’t up to the task. Agencies were branded slow, expensive, uncollaborative, culturally insensitive, and technologically inferior.
This is largely, I believe, unfair on many agencies built on precisely the agile principles sought by those on the stage, but woven into this is a truth that change is inherently difficult.
For agencies to change, to even have a sniff of a chance to transform, requires a fundamental re-calibration of legacy cultures, systems, roles and responsibilities. And unfortunately, you can’t just jump to that.
Before any of that can begin, agencies have to learn how to learn in order to set themselves up for success, such things as…
- bringing together cross-disciplinary teams to establish the best way of learning
- creating feedback environments so progress can be shared
- sharing the long-term vision as to why progress is needed
- involving clients in their plans to ensure that they’re compatible
…among a host of others, each dedicated to ensuring that the ways we set about learning new ways of doing things are iteratively improved upon time and time again.
The machines are, in many ways, already way ahead of us.