Go to the Globe and Mail homepage

Jump to main navigationJump to main content

dv495014?(RF) Decisions Digital Vision ROYALTY FREE (Getty Images)
dv495014?(RF) Decisions Digital Vision ROYALTY FREE (Getty Images)

These companies help guide your online purchasing decisions Add to ...

The abundance of options in today’s marketplace might seem like a good thing, but studies suggest that too many choices can actually increase consumers’ anxiety to the point where it prevents them from purchasing anything.

“As choice increases, we get overloaded,” says June Cotte, a professor at Ivey Business School. When faced with a large number of options, consumers spend more time deciding and, as a result, more time regretting their purchases. “Because we spend more time,” she says, “you’re going to remember the foregone choices.”

More Related to this Story

And with increasing awareness of ethical options, some customers are facing even more choices than they did even five years ago, she adds.

Some companies are seeing an opportunity to capitalize on all this confusion.

Sortable has built a business around helping consumers make decisions. The Waterloo-based company, which was acquired in 2012 by Rebellion Media, runs several websites that assist consumers decide between things like cars, cameras and computer hardware.

And they’re seeing results, according to Chris Reid, CEO and co-founder of the 25-person team. Customers are more satisfied with the purchases they make based on the company’s recommendations, with retail partners reporting a 75 per cent reduction in return rates.

“Online shopping is still not the best experience,” says Mr. Reid. “The information is on disparate websites, research takes lots and lots of leg-work … the user can’t read thousands of reviews.

The company’s software aggregates everything from reviews, to comments on social media websites and government safety ratings for cars. It then structures the data in a way that makes it easy for customers to compare features.

The company’s websites vary depending on the product being recommended. On some, like Carsort.com, users indicate their price range, the size of vehicle, whether they have a brand preference and can prioritize a variety of features, The website then provides a ranked list of vehicles that fit the criteria.

On other sites, like TabletRocket.com, users pick two competing products. The website then provides them with side-by-side comparison of features and aggregated reviews.

Building software that can not only gather all this information, but also understand it, is a big challenge, says Mr. Reid. “It’s ridiculously complicated to do it well and to do it accurately,” he says. Another hurdle is dealing with products that have multiple versions.

For instance, he says the Honda Civic has different versions available in different countries and a wide variety of optional features. Creating software that can understand the variations between the different versions while also understanding that they all fit within the same category is particularly difficult, especially if all the versions have some features in common. “If all the models have the same a.c. but different engines, doing that is very difficult,” says Mr. Reid.

Mr. Reid says that in order to do this, they’ve had to move one step at a time, building websites focused on a specific product type.

The next step is giving recommendations based on a product the user already likes. So far, the company has launched websites that allow users to input the name of a movie or TV show that they enjoy to get recommendations of similar programs or films that they also might like.

As the software continues to develop, Mr. Reid says it could eventually recommend anything in a similar manner.

“You could tell us a few things you like and it could probably take that data and answer [any] question,” he says.

Ms. Cotte says she expects to see the use of recommendations engines grow. She also believes that third-party recommendation engines might have an advantage over the recommendation systems on websites like Amazon.com because they aren’t pushing a sale as visibly.

“Consumers need to trust that it’s accurate,” she says. “Third-party sites are even more trustworthy.”

But even though Sortable might not be selling directly to their users, they’re still hoping to drive purchases. “There’s actually a lot of monetization options for us,” says Mr. Reid. “We generate very high revenue per visitor.”

He says the company makes money both through advertising and partnerships with online retailers.

And because of the nature of Sortable’s websites, the advertising on their website is highly targeted, and particularly effective. “Our software is really good at understanding what people are looking for in products,” he said. And because people are selecting the features they want, the company can micro-target advertisements to people who are interested in the specific products being advertised.

“We can quantify what people are interested in,” Mr. Reid says.

The company is also looking at licensing their knowledge graphs, essentially the brains of their system, to phone manufacturers, though no deals have been signed yet. That’s the route taken by Kitchener-based Maluuba, a company that natural language processing technology. Their product is an app that gives exact results to short queries, says Andrew McNamara, the company’s infrastructure architect.

He says users could ask the app, by voice, what time a movie starts at their local theatre and get a specific answer. While their product doesn’t give users recommendations exactly, it can give answers based on suggestions posted online. For instance, it could tell a user what the highest rated restaurant in their area is.

“The next step is tying in recommendations,” says Mr. McNamara.

The Maluuba app is a free download, with no ads. Instead they’re looking to license it to phone manufactures to power voice-based assistants.Their first partnership, with LG, sees their software used on the new G2 phone, released earlier this week.

Both Mr. McNamara and Mr. Reid say that working with natural data and structuring the unstructured data of the Internet presents tremendous challenges.

Even after several years of work, Mr. Reid says the problem of giving perfect recommendations still isn’t solved. “There’s still this race to build something that gives really good recommendations,” he says. “There’s no one who’s really good a doing this yet.”

And his ambitious go even further than that – he said his company hopes to redefine the whole search engine business.

“We want to change the way people think about getting answers on their devices.”

Follow us @GlobeSmallBiz and on Pinterest
Join our Small Business LinkedIn group
Add us to your circles
Sign up for our weekly newsletter

Follow us on Twitter: @GlobeSmallBiz

In the know

Most popular video »

Highlights

More from The Globe and Mail

Most Popular Stories