In 2004, Blockbuster employed 84,000 people and operated in 14 countries, renting out videos and games to millions. Six years later, after the advent of video-on-demand and streaming companies, it was forced into administration. The story is not too dissimilar from that of camera giant Kodak. After an employee invented and patented the digital camera in 1975, the company shelved the device because execs believed that customers would never want to look at photos on a television. Too slow to switch to digital, the business filed for bankruptcy in 2012.
Both are cautionary tales of what happens when businesses fail to embrace technological innovation. It’s a mistake that Vancouver company Unbounce is not willing to make.
“This is straight out of one of my favourite books, The Innovator’s Dilemma,” Carl Schmidt, the company’s CTO, tells the Georgia Straight at Unbounce’s annual Call to Action conference. “If you don’t keep pace—if you don’t adapt your business model to changing technology—you will die. It’s that simple. Generally your customers will be solving today’s problems, and generally you need to be investing in the problems that they’re going to face a couple of years down the line.”
For the past few years, Unbounce has chosen to focus on how artificial intelligence can be used to figure out which factors make up the most attractive pages on the Net. A business that allows companies to easily design webpages and A/B–test which variants perform best in front of customers, Unbounce has created a machine-learning system that asks computers to rate landing pages—typically, a site’s homepage—on how likely users are to engage with them.
Shocking Schmidt, the machine was able to guess which webpage would perform best at a rate of 79.7 percent. Humans, meanwhile, were only right half of the time.
“The surprising thing is just how impactful words ended up being,” he says. “We were able to predict a lot of things based just on the words. That told us that words played a potentially outsized role in the performance of these pages. We’ve said this a lot, that the other elements are clearly important, but perhaps—and we still have to do more work here—there’s a strong signal here that copy and writing is dominant.”
Using its machine-learning algorithm, the company was able to compile a number of insights that marketers could use to target consumers. Real-estate pages, for example, perform best when written at a seventh-grade reading level, whereas college admissions websites are more likely to encourage user engagement with university-level copy. Language that builds a feeling of anticipation or pressure in the home-improvement industry turns visitors off, while an excess of positive words on a website for legal services makes it seem less credible to readers.
After gathering a glut of research data, the company felt sufficiently confident about its results to release its AI into the wilds of the Internet. Over 30 Unbounce clients offered to be guinea pigs for new technology, and connect their campaigns to the company’s engine. After a few weeks, those businesses saw a lift of around 20 percent in conversions, with one tallying a 117-percent improvement. Schmidt attributes that success to the AI’s ability to deliver the right page to the right person.
“There’s no one message which is going to work best for your entire audience,” he says. “We know that words affect us differently based on our own biases and perceptions and experiences. There’s no way that one page can be the best for all of your visitors. Right now this is patented tech, so we’re going to hold it a little close until we figure out how we’re going to productize it. Then we’ll be talking a bit more about how it does what it does.”
If the company is able to replicate those results on a wider scale, Unbounce’s AI technology could spell big changes for consumers across the Internet. Although the company has not divulged exactly how its creation will operate, machine-learning technology in the future could be able to tailor each page for an individual, creating a more personalized experience, and encouraging more sales for companies. In Schimdt’s view, AI is a win for both browsers and businesses.
“The reason AI is important is because the world is becoming pretty overwhelming,” he says. “We’ve got flooded inboxes; we can’t keep up on our social feeds. It’s hard to just get the information that is really relevant. I think that AI can help us as consumers ultimately deal with the information firehose better, and be a filter for us. In the end, that will give us back more time.
“AI is going to be one of those differentiator technologies,” he continues. “Same as when the Internet came along and changed a whole bunch of business models, and challenged print, challenged mail-order—all of these things. It’s that class of technology.”
Follow Kate Wilson on Twitter @KateWilsonSays