Sramana Mitra: Where does the data come from? Where are you getting the data on which you are running your personalization?
Diaz Nesamoney: We work directly with brands. Let’s take, for example, a large hotel chain. A lot of the people would go to the hotel chain website. There’s a tremendous amount of data there. What category of hotel do they look at? That tells us a lot.
There’re also third-party data platforms like Oracle’s BlueKai and other data management platforms. There are a number of third-party data providers who are providing aggregated data across the web. There’s also basic stuff like weather. You’re likely to buy certain things when it’s warm. Where are you physically?
Sramana Mitra: So you do geo-targeting as well.
Diaz Nesamoney: Yes. We basically build, in real-time, a personal profile of the individual. Where is this person at demographically, geographically, and contextually? That tells us exactly what kind of message they’re likely to respond to.
Sramana Mitra: Then you use the retargeting kind of model to follow the customer?
Diaz Nesamoney: Retargeting is one of the triggers. Retargeting simply says that this person expressed interest.
Sramana Mitra: But you have to somehow cookie the person to see who that person is. My question is where do you do that.
Diaz Nesamoney: Typically it’s when the user visits the brand’s website, which is why it works better when a user visits the website. However, that’s not always the case because a cookie is a cookie and works across all brands. The identifier is not unique to that particular brand. In other words, if we know that you went to a particular hotel site and then went to a car rental site, it’s still you. We’ve just learned one more thing about you. That identity is what ties together all these pieces of data together.
Sramana Mitra: The reason I’m probing is what is the starting point. You’re deriving the context from the site of your client.
Diaz Nesamoney: Yes, that is where it usually starts.
Sramana Mitra: If you are mapping me, you are mapping me in the context of your client’s website.
Diaz Nesamoney: Yes.
Sramana Mitra: You want to understand what my hotel-related interests are.
Diaz Nesamoney: Yes, and then we can build on that.
Sramana Mitra: You may model me on a car’s site differently later on. Each of those are different models.
Diaz Nesamoney: Exactly. We are also modeling you on how you engage with the brand. For example, when you click an ad, that’s also an indicator of interest. We may actually serve you multiple versions of the ad to see which one resonates best. One might have a special offer. Another one might have just really nice visuals.
The system’s then optimizing to say, “This is a person who typically responds better to an offer.” Somebody else may engage for entirely different reasons. The more data points we have, the more likely we are to present the consumer with something that is of interest to them.
This segment is part 2 in the series : Thought Leaders in Artificial Intelligence: Diaz Nesamoney, Founder and CEO of Jivox
1 2 3 4 5 6