A Robot Named Albert Wants to Revolutionize Digital AdvertisingBy Tessa Wegert October 27th, 2016
You’ve got to hand it to robots—these days, they can do just about anything. Powered by artificial intelligence, they’re writing news stories, working in hotels, and even running a McDonald’s in Phoenix.
But would you trust them to manage your entire advertising campaign?
Adgorithms hopes so. This August, the digital marketing company launched Albert 2.0, an artificial intelligence platform designed to make digital media buying decisions on behalf of brands.
It was a long time coming. Seven years ago, founder and CEO Or Shani created Adgorithms in the hopes of making marketers’ lives a little easier. If he could automate more of the advertising process, he thought, the energy media buyers spent on minutiae like analyzing click-through rates could instead be devoted to the tasks that require a human touch.
“Albert does all the execution work for you,” Shani said, “so you really have time to think about strategy and focus on creative.”
On the platform, users enter campaign parameters like target audience, KPIs, location, desired devices and channels, and ad creative. Using predictive algorithms and learnings from past campaigns, the AI builds profiles of target users and dictates which ads should be shown where and when, processing the data and adjusting decisions along the way.
AI has been a major point of interest for numerous martech companies in recent years. IBM’s Watson leverages machine learning to help people do everything from improve cross-channel shopping experiences to create branded content. Salesforce’s AI solution, called Einstein, aims to deliver a more personalized customer experience.
Albert, however, is entirely about ads.
In essence, Adgorithms attempts to take programmatic advertising to another level. Instead of just relying on automation for individual media buys, brands can turn all of their campaigns over to Albert. The AI is designed to consolidate and micromanage everything from email marketing to native advertising, and it integrates with platforms like Google Display Network, Outbrain, MailChimp, and Snapchat.
“People try to work in silos,” Shani said of the current digital marketing space. “You have your branding, direct response, retention, your Facebook team, your Google team. But everything needs to work together. What we’ve created is something much more flexible.”
Shani claimed that, with a little training, anyone can operate Albert, whether they’re on the brand side or with an agency.
This is hardly the first time a robot has tried to elbow its way into the digital marketing space. The transition from manual media buying to an automated system started close to a decade ago with the arrival of ad exchanges and technology that promised real-time bidding. Earlier this year, eMarketer predicted more than two-thirds of all digital display advertising in the U.S. will be purchased programmatically in 2016.
But Albert may be ushering in a new era—now, instead of just running paid media campaigns, Adgorithms wants artificial intelligence to become a full-fledged member of the marketing team.
Bridging the business and marketing gap
For a business like Harley-Davidson of NYC, Albert has a lot of appeal. Asaf Jacobi, president of Harley-Davidson NYC, will tell you that New York City is an “incredibly difficult place to break through using traditional advertising.” His prospective buyers represent just 2 percent of the population, which caused a lot of inefficient spending in the past.
Jacobi knew that digital advertising was the best way to reach his potential customers, but he wasn’t sure how to use it effectively with the limited resources at his disposal. It’s a problem shared by many business managers, particularly as the digital marketing technology landscape—now comprised of more than 3,800 different solutions—has grown increasingly complex.
“I’m not a marketer,” Jacobi admitted. “I oversee all business operations. But suddenly I found myself in this new position where I was trying to solve what were clearly marketing problems.”
What happened next is straight out of a movie plot.1 Jacobi was walking through his Manhattan neighborhood one day when he stopped to chat with a stranger. That stranger was Shani, who told him about Adgorithms.
“The rest,” Jacobi said, “is history.”
So is Harley-Davidson of NYC’s tradition of manually buying media. Now, Jacobi tells Albert his desired outcome for a campaign, his target audience, and his geographic priorities. Albert does the rest.
By sussing out online behavior patterns and creating custom combinations of creative and copy, Albert identified and targeted audiences Jacobi didn’t even realize existed. It then prioritized top-performing ad concepts, predicted the optimal pricing for each ad buy, and even redistributed the budget to favor those channels that performed best.
“We don’t add rules about when to send things out or what audience segments should get what messages at what time of day, or even what headlines to pair with what creative. That’s the self-learning part,” Jacobi said. “[Albert] begins going to work to meet our goals, processing thousands of events in seconds. He knows better what to do than we do.”
What Jacobi does do is monitor the leads generated by his campaigns, along with the conversions. Early on, to put Albert to the test, Jacobi devised a plan. On a good weekend, his dealership sells four or five bikes; Albert was expected to sell 48 in 48 hours.
“This was at the end of spring and, as any retailer knows, you’re always trying to get rid of a bunch of overstock. Every year I deal with it,” Jacobi said, “and it’s a real hassle if you’re in retail.”
Albert couldn’t deliver on the ambitious goal, but after 48 hours, Harley-Davidson of NYC had sold 15 bikes. That was triple what the dealership likely would have sold on its own, and, to Jacobi, that was a win.
“Tech companies aren’t here to create another acronym and add complexity to the business. That’s not what people want.”
It’s been about three months since Harley-Davidson of NYC started using Albert for both for direct response and brand awareness campaigns, and it now credits 40 percent of its motorcycle sales to the platform. Daily visits to the company’s website have increased by 566 percent. By the second month, leads more than tripled.
“What people in marketing tech and ad tech tend to forget is that tech companies aren’t here to create another acronym and add complexity to the business,” Shani said of the AI’s value. “That’s not what people want. They just need it right now because there’s no other option.”
Can brands trust a bot?
It’s tempting to see Albert as a panacea for the ad industry’s woes. But can media buyers really entrust a robot with so much autonomy?
In the past, we’ve seen automated marketing go horribly wrong when brands put their faith in automation but fail to keep a close watch on the results. For example, an ad for an SUV appeared alongside a story about a horrific car crash because marketing technology mistook it as contextually relevant. Misguided ad placements can make companies appear heartless and damage a brand’s image. It’s a digital marketer’s worst nightmare.
Placement isn’t the only concern. While spending on programmatic advertising continues to rise, studies show that advertisers still worry about ad fraud, wherein algorithms mistakenly deliver ad impressions to fake websites without real human traffic. According to the Association of National Advertisers, bot traffic will amount to more than $7 billion in wasted investments globally this year. That has brands thinking twice about going all in with automated advertising.
Jacobi, for one, isn’t worried about Albert betraying him.
“Because all of the creative and copy comes from us and our agency, we have complete control over the experience from a brand perspective,” he said. “We are infinitely more confident that the right people are seeing the right messages, and the wrong people aren’t seeing any.”
“Our system is completely transparent, like a glass box, so you can really see what happens,” Shani said. “We do have roadblocks in place to ensure bad things don’t happen, but probably some will. I’m coming from a profession where I believe everything can be worked out with data. But you do need humans.”
To that end, the platform allows for humans to watch campaigns as they happen—and interrupt them if they see something going wrong.
In that sense, like all good AI robots, Albert has a kill switch. Now, the question is whether Adgorithms or its competitors can make automated advertising campaigns so seamless that marketers never have to use it.Image by imagewerks / Getty