Innovation is like Waldo: hard to find

It might sound strange coming from me, but it seems like it’s harder to find real innovation these days. Take digital marketing as an example. I don’t claim to be a marketing expert, but I saw similarities in my thinking and John Battelle as I read his answer to the question, Has Innovation Died in Marketing?

John wrote: The history of marketing over the past 15 years has not been one of customer engagement, and as for supporting innovation in news – it’s been mostly crickets. Innovations budgets have all but disappeared – one senior media buyer responsible for billions in annual ad spend recently told me that they hadn’t had money for media experimentation for nearly a decade.

A lot of the big money in marketing is going to brain-dead strategies: throw your marketing budget at Google, Amazon or Facebook and pat yourself on the back for doing your job. It’s very similar to the old mentality that nobody ever got fired for buying IBM (and later, Microsoft).

It’s not that innovation isn’t happening. Obviously, we are using new technology in new ways (like, for instance using robots to peel bananas, which works 57 percent of the time. Okay, maybe not the best example of exciting innovation). I think my feeling on this is that, while our capabilities are increasing, maybe the end game for how these technologies will be used isn’t so well thought out.

Innovation for its own sake can get very expensive, as IBM has discovered. It is selling Watson Health assets for $1 billion. Nice payday! But that comes after spending $4 billion just on acquisitions to prop up this project.

Let me give you a few examples, where some of the most well-known companies in the world say they are trying to drive innovation - but are getting hung up when their technology works exactly as designed. It’s almost like the people who build technology should be thinking about strategy before they build it.

You might have been surprised to learn this week that Google, a company that relies upon machine learning algorithms to a huge extent, is not happy with AI. Specifically, people in Google’s search division are having issues with websites where the content is generated by machines.

Keep in mind, this Is Google, the company which built AlphaGo, 'the first computer program to defeat a professional human Go player, the first to defeat a Go world champion, and is arguably the strongest Go player in history.'

Google suddenly has an issue with people using the same technology against it. This is like saying 'we can build things which are far better than humans, but we still need humans to learn from.' #chokingonirony

Meanwhile, Salesforce’s research team is working on a system where developers will be able to generate code by talking to their computer. According to the developers this is how this is going to work:

Human: “I would like to create a red button.”
Machine: “Where do you want to position the button?”
Human: “In the center.”
Machine: “What happens if the button is pressed?”
Human: “Calculate a measure of center for Bitcoin's price over the past 10 days.”
Machine: “Mean, median, or mode?”
Human: “Mean. Please show me the code.”
Machine: <shows code it generated>
Human: “In line <X>, I was meaning to <state a correction or revision or new approach> … could you revise this accordingly, please?” (Do I really have to keep saying 'please'?)
Machine: <shows revised code it generated>

In a few simple sentences, you can describe what you want to do. Perfect. And now, for the rest of the day, you will have an argument with the computer about what you actually meant. (Seriously, this will not be as easy as making a nice, hot cup of tea.)

There are some areas of innovation where I’m not all that sad that a lack of strategic foresight is going to derail the project. While we are on the topic of AI, here is a new buzzword for you from the Orwellian category - One Person, One File. That's the latest requirement from governments when it comes to surveillance.

There are cameras, voice recordings, social media, SMS texts, emails. How do you reliably attribute all that to the right entity and build the profile with the right social score? That's the challenge.

I have been around CRM systems for keeping customer information long enough to know that this is an impossible task. CRM systems have been around for decades and nobody knows how to clean the data properly. Doing it at-scale with a billion people? Good luck.

Back to what exciting innovation is all about: not trying to replace humans, but helping us do what we want to do. I’m talking about augmented intelligence, not artificial intelligence.

People think. Robots do. Innovation flops, expensively, when you try to reverse that dynamic. And that’s the recurrent pattern.