Commentary

Anodot Working To Predict Future Data Anomalies

Data paints a picture of aggregated human behavior that recognizes and identifies certain traits, but artificial intelligence can identify anomalies built on a baseline to determine why consumers may or may not complete a purchase.

Predicting anomalies in data with help from AI in real-time is great, but the real Holy Grail is the ability to predict two weeks in advance the real price of an advertisement or how much inventory a brand will need in a specific area based on sales the ads support, said Rebecca Herson, vice president of marketing at Anodot.

"When I talk about the future, it's really something we're working on and not something we've rolled out," she said.

Anodot, founded in 2014, focuses on real-time analytics and automated anomaly detection in data. In September 2016 it raised an $8 million series B round of funding, and then earlier this year the company was named one of the 30 machine-learning startups to watch in Israel.

Andot's technology looks for anomalies in data that would suggest a change in human behavior. One example might be a change in ad bidding or a drop in conversions for a specific demand-side platform on a specific location for consumers using smartphones running Apple's operating system.

AI in data can identify "the unknown in the unknown," Herson said. "A human can't possibly ask all the questions that need to be asked, and that's where AI comes in."

The technology learns all the behavior of the data metrics based on specific behavior, identifies the anomalies, alerts the users, and provides the changes to companies it works with such as Rubicon, Kenshoo, iView, and many others.

Google and Microsoft are both Anodot customers, but not specifically for advertising, per Herson.

For ecommerce brands, Anodot helps to identify a spike in orders in a specific region to track potential product shortages in stores in real-time.

The technology also identifies a correlation between bidding prices, weather and other data sources.

The data can tell advertisers when media buys don't line up or are mismatched by monitoring the quality and relevance of the data and identifying why revenue might fall.

Anodot also can support search marketers using platforms like Google Analytics or Adobe Analytics, The technology can track the traffic coming from various search engines to see if there's a drop in searches per keyword or product and correlate it with other variables.

In one case a drop in traffic from search engines enabled a brand to identify when a competitor raised its bid prices on keywords. This was possible because the customer also streamed competitive data to Anodot it had purchased from a third party. 

Next story loading loading..