How To Innovate Like A (Google) Pro
Alberto Savoia, founder of Pretotype Labs and former Google innovation executive, says pretotyping – a method for deciding whether or not to pursue a new idea – can save companies millions by helping them avoid costly flops and get the right products to market
Innovation is all the rage these days, with companies and contributors racing to rock the market with the ‘next big thing.’ Unfortunately, most new products and innovations fail – not because they are poorly executed but because the idea wasn’t the right idea.
“The biggest challenge in innovation isn’t coming up with new ideas, it’s identifying which of those ideas will be successful in the market,” says Savoia. “Pretotyping helps organizations test their innovative ideas to make sure that they have the right ‘it’ before they invest in building ‘it’ right.”
I sat down with Savoia in Silicon Valley earlier this month for a lively Q&A. Savoia explains why a seemingly great idea goes bust while an outright crazy one makes bank, and argues that pretotyping, often seen as a ‘just a start-up thing,’ is relevant to companies of all sizes.
Jennifer Lakheim: What was the inspiration behind pretotyping?
Alberto Savoia: I’ve had the good fortune of being an early employee of two companies that went on to become industry giants – Google and Sun Microsystems – but in my heart of hearts, I’m an entrepreneur. After my first stint at Google from 2001-2002, I had a successful start-up. Then I had (a second) start-up. We felt we were doing everything right. We raised a sufficient amount of money, had great people. Everything should have gone perfectly, but when we launched the product, not enough people wanted it and wasn’t successful. I asked myself, what went wrong here? I went back to Google (in 2008) with the intention of studying failure: Why is it that eight out of ten start-ups fail? Why do 80 percent of new product introductions flop?
What is the crux of the pretotyping approach?
Make sure you’re building the right ‘it’ before you build ‘it’ right. This doesn’t just apply to technology; ‘it’ could be anything. If you’re an author, before you write a book you want to make sure there’s a market for it.
Aren’t companies doing this already? Researching whether there’s a market for your widget seems like a no-brainer in business.
They think they know the market, but really they’re just guessing. All entrepreneurs fall in love with their great idea, so they invest a lot of money, launch it, and nothing happens. Google Wave, which was supposed to replace email, is a famous example. It was done by two of the smartest people at Google. It had the pedigree; it had the people. Email was ready for an improvement, and yet – after a significant investment – the launch fell flat on its face. Why? Because, instead of actually collecting data on whether or not people would use the product, they asked for opinions about the product and based the development on that.
What’s the difference?
I’ll give you an example. Remember Webvan? They presented the idea and asked people if they’d like to order groceries online. Everybody said, “Yes! Of Course! The Internet is cool; it’s a convenient service; and I like to shop!” So, on the strength of that positive reaction, they went for it and launched the company, all out. In just a few years, Webvan raised nearly a billion dollars, built huge refrigerated warehouses full of stock, and acquired a fleet of vehicles. Well, it turns out that even though nearly 100 percent of the people surveyed loved the idea of Webvan, only about two percent of them actually used it. And the company went out of business.
Apply pretotyping to the Webvan example and instead of spending over $100 million to launch the company based on surveys, they would have started small and watched what happened. They would set up the website, advertise it one market – say, San Francisco and its suburbs – and see how it goes. They would not buy warehouses and trucks; instead they’d work out some kind of deal with a supermarket or supplier to fulfill the orders. Had they done this, Webvan would have seen that the number of people who said they would use the service was dramatically different than the number of people who actually did use it. They would have seen that the majority of users were in the city. It would have been apparent that they needed to scale back the concept, tailor it to city-dwellers and forgo the suburbs. And they would know this without having spent enormous amounts of money. The Webvan story might have ended differently.
Webvan illustrates one innovation blunder: over-investing in a concept that turns out to be nonviable. Are there any other common mistakes companies make?
The second error, even more painful, is killing the right idea too early. Take Twitter. Get a focus group together and ask them if they would like to communicate to the world, in 140 characters, that they’ve just gone shopping or done laundry, and in turn follow other people posting the same kind of information – well, it sounds ridiculous. If you relied on just that, the idea would have been killed. And today Twitter is worth almost $10 billion.
How can innovators use pretotyping to help determine what is actually a viable idea, and not one that just sounds good in theory or, on paper, seems patently absurd?
The challenge is, until you build something, you’re dealing only in ideas, and ideas are abstract; the most you can solicit from an idea is an opinion, which is also abstract – and subjective. To get objective data, you have to get out of this ‘thought-land’ as fast as you can, and that’s what pretotyping is all about. It is the process of figuring out what you can do, short of building the product or a prototype, to see if there is really a market for your innovation.
My favorite example is IBM and speech-to-text technology. Thirty-five years ago, it was thought that professionals and managers could not be expected to type; back then, only secretaries and programmers typed. IBM felt the only way they could succeed in the personal computer business was to invent speech-to-text, so that users could just speak to a computer and not use a keyboard. This was going to cost a fortune. IBM did preliminary research that indicated people would indeed buy a technology that allowed them to talk into their computers rather than type, but they (wisely) decided that before making a huge investment they needed to find a way to test this hypothesis.
They did it by making what I credit as one of the earliest pretotypes. They built nothing – no hardware, no software. They simply put a person in a room with a screen and a microphone and told them to start speaking. The test subject would say something and – thanks to a behind-the-scenes stenographer– their words would appear on screen, simulating speech-to-text.
Guess what IBM learned? Just like with Webvan, what people thought they wanted was very different from what they would actually use. The test subjects said they wanted to talk to their computer, but after five minutes of actually doing it, their throats were getting sore and the room, with everyone talking to their computers, was noisy and obnoxious. They were over it. So IBM learned that speech-to-text technology was, at the time, not worth developing. That is the essence of pretotyping.
You also cite the example of Jeff Hawkins and the Palm Pilot as an early pretotype. What’s that story?
Legend has it that before committing to develop the Pilot, Jeff went into his garage and cut a block of wood that would fit in his shirt pocket. He carried it around for months, pretending it was a handheld. If someone invited him to lunch, he’d pull out the block of wood and tap on it as if he were checking his schedule. If he needed a phone number, he would pretend to look it up. He even tried different design faces and button configurations by gluing different pieces of paper to the wood block.
Both of those examples involve pretotyping-like activity in large companies. Isn’t this the stuff of start-ups?
Actually, the companies that are the most interested in pretotyping are the largest companies. I admit I was surprised by this; I also thought it was a start-up thing. But if you look at the existing ‘waterfall’ model for new product development in the enterprise, it makes sense. Traditionally, a company decides they’re going to launch a new product, spends two years in development, a year in testing and refinement, and then they launch it. People are dissatisfied with that, because it takes a long time and if the product falls flat on its face – which happens all the time – they’ve just wasted an enormous amount of money, time, and resources.
Do you find that large, established companies are receptive to the idea of pretotyping? Or are they stuck in their (perhaps outmoded) ways?
It’s true. The bigger and more established the company, the harder it is to get into the frame of mind to do experiments.
How does an innovator in a large organization combat that to be successful?
Whenever there is innovation, there are two constituencies: one is the bright-eyed, bushy-tailed engineers with the great idea. Then there are the people on the business side, asking about headcount and resources. In a conventional, waterfall scenario, the first camp only gets people to buy-in to their idea if they are powerful and charming and have a history of success as an innovator. More often, you’re just a guy with a crazy idea, like Twitter, and you get shot down.
At Google we say ‘Data beats opinion’ and ‘Say it with numbers’. Pretotyping does this: it moves you from the world of ideas to the world of data. The discussion between the innovators and investors is no longer about opinion. You don’t just think your product is going to be great; you have data to back you up. That changes the conversation. I like to say: ‘Innovate like a start-up; go to market like a grown up.’
Beyond that, it’s really just a matter of time, given the rate of change in the market, that concepts like pretotyping are widely adopted. If you look at mega trends in business and technology, it’s not just optimal, but essential. Things are moving too fast to start with big plans. If you have something with a two-year development plan, it’s going to be obsolete before you launch it. It just doesn’t work. You need to be very nimble and do things quickly.