Vitaly Pecherskiy and Ildar Shar met while working in the digital media industry and often talked about how online advertising could be improved. In June 2013, they decided to join forces and do something about it. They knew they needed stronger technology skills, so they recruited Yang Han, who had been working on real-time stock trading technology at Bloomberg, to join them as a third co-founder. Together they came up with many possible improvements and they weren't sure which was the most promising to run with. How could they figure this out?
Advertising companies run ad campaigns for clients by placing ads in media such as magazines. Digital advertising is no different, except that the ads are placed on the websites of diverse organizations – called publishers – who want to generate ad revenue from them. The difficulty in advertising has always been matching the ad to the people who see it, and digital technologies offer great potential to match ads with people.
It's a challenge: the online audience is heterogeneous, geographically distributed and accesses the web through many different devices. Although digitization has the potential to change technology, Mr. Pecherskiy says that it has been slow to do so: "Banner ads remain the dominant form of online advertising and have just celebrated their 20-year anniversary. However, they aren't all that effective. They don't translate well to mobile devices, which more and more people are using to surf the web. And people are coming to ignore banner ads, so advertisers pay less for them, which is leading to a massive decrease in monetization for publishers. To compensate, publishers are adding more and more ads, which decreases the value of any one ad and jeopardizes the user experience on the web. We knew there was a clear need for a resolution for this downward spiral." The next step was to figure out what it was.
The co-founders started with what they knew. Since banner ads are prominent in digital advertising, they started by exploring ways to engage users with them. For several months they worked on experimenting with technologies that created ads in real-time. For example, they learned how to dynamically pull – from platforms like Instagram – a user's pictures relevant to an advertiser in order to create ads in real-time from them, and to change the colour of a banner ad in real-time in order to match the colour of a publishers' website. Mr. Pecherskiy states that exploring this functionality led to bigger ideas: "Once we could handle colour separately, we realized that we could handle all the ad elements separately. This meant that we could provide 'native advertising,' which is advertising that is customized to each device – a desktop, tablet or smart phone – on which people might see an ad."
The partners then built on this idea and developed a technology that recognizes the stylistic features of the site an ad will appear on and dynamically alters individual ad elements to make the ad native to that site.
StackAdapt went live in March 2014, but had a chicken-and-egg problem: to get advertisers to sign on, they needed publishers of websites, but to get publishers to sign on, they needed advertisers. Although there are thousands – even millions – of potential advertisers and publishers, the founders decided to work closely with a few to get it right. They started working with several agencies that were interested in testing new platforms to deliver clients' message. Ads are traditionally put on contextually-relevant websites; websites that people go to when they're thinking about buying a car or a vacation.
"Through trial-and-error, we learned that behavioural targeting – based on a user's online cookie data – produced better results for the client. Our technology examines a user's cookie data in a fraction of a second when a web page is loading, looks across all the advertising campaigns StackAdapt is running for clients, decides which ad is the best match for that user, and presents the ad to match the website's style," says Mr. Pecherskiy.
Importantly, the partners realized quickly that behavioural targeting also provides better results for StackAdapt, because it is a much more scalable model to grow the firm with.
Mr. Pecherskiy explains that once they decided to emphasize behavioural targeting, the marketing challenge became a technological challenge in that it had to be able to process tens of thousands of requests per second. Again, they worked closely early on with a few companies – website publishers this time – to make sure that the technology worked and the system wouldn't crash.
The founders of StackAdapt waited to scale their business until they had developed a strong value proposition and technological platform by working closely with innovative customers. Mr. Pecherskiy points out that the value proposition scales up as more advertising clients and publishers sign on. "It may be really hard to find the perfect target for your advertising message when you have only a few websites on which to place that message, but when you have thousands of websites, you have a much larger pool to draw from."
There are now 22 people at StackAdapt, with offices in Toronto, New York City, and London England. Mr. Pecherskiy says that they already place several billion loads of online ads per month and, in September 2014, received external seed financing of $900,000 to invest in further product development.
Becky Reuber is a professor of strategic management in the Rotman School of Management of the University of Toronto.
This is the latest in a regular series of case studies by a rotating group of business professors from across the country. They appear every Friday on the Report on Small Business website.