This edition of Real-Time Banter features a discussion about mobile advertising data with Gil Elbaz, founder and CEO of Factual. Gil was interviewed by Jennifer Lum, Chief Strategy Officer for Adelphic.
Jennfer: What is Factual?
Gil: Factual is a neutral data company. Our job is to make data accessible to other companies, and to help them innovate and be more productive. We have this real focus on neutrality. For example, we will never build a consumer app because we want to be absolutely aligned with our customers, providing the highest quality data to make their jobs easier. We specialize in mobile and location. We found that this is a massive and interesting opportunity to provide data about the entire world, about every spot on the world. We think there are a tremendous number of opportunities to provide value by focusing on location data.
You are the founder and CEO of Factual, can you share Factual’s founding stories with us?
I have been addicted to data since I was a little kid. I was in love with the reference section of the library – from dictionaries to the Guinness World Records – and anything that was structured. I turned this into a career and my first company, Applied Semantics, started out with a focus on building ontology. Ontology is the organization of the world’s information: every word and phrase that you can find anywhere in language. We realized that this is the basis for natural language understanding technology, which then became the basis for AdSense. Applied Semantics was acquired by Google in 2003, and so I spent a few years with Google. Factual represents taking knowledge and structuring it to the next level. The goal of Factual is to make this data accessible; to make structured data, high-quality data accessible because there is such an opportunity to be neutral, to power all other innovations. We are not going to innovate across every industry; we can never hope to do that. But by being a provider, we get to work with companies across many different industries and create value across the landscape.
Are there any important lessons you learned in building Applied Semantics that you are now applying to how you build your business and your company at Factual?
When you have a big vison, there are certain elements where there are long secular trends and things don’t change overnight. So at Applied Semantics, we pioneered contextual ads. For a couple years, our AdSense revenues were exactly zero because the supply partnerships just weren’t interested in this form of ads. The standard ways of doing things had been targeting on demographics. And then it started to work…we were patient. As it started to work, it felt like the large companies at the time – Google and Yahoo – had their eye on us, and it seemed like we had to get acquired and they were going to win. While it was exciting to get acquired by Google, I learned a tremendous amount. It was interesting to learn that it wasn’t the last opportunity to create an independent company in the space. Actually, several other companies launched after we were acquired and were very, very successful in contextual ads. So, actually when a trend emerges, there are going to be plenty of opportunities for many companies within that space for years to come. So I think that there are going to be tremendous opportunities to build great data companies for many, many years. And I will be committed to this new industry.
“I think that there are going to be tremendous opportunities to build great data companies for many, many years.
And I will be committed to this new industry.” – Gil Elbaz, Factual
Factual has three core costumer segments: enterprise, developers and mobile advertising. Can you describe Factual’s products that focus on mobile advertising?
We built this core data, and we found that there are use cases across several different categories, including mobile apps and enterprise. There are tremendous use cases for this type of data in the mobile advertising industry as well. What we did is we packaged the data in a way that makes it much easier to integrate and much more actionable for the mobile adtech ecosystem. So the key thing that we’ve done with location data is turn it into audience data. That’s much easier to act upon within this industry. We’ve also built software that a demand-side platform can run on premise. It allows us to continuously push the freshest and most relevant targeting or audience data. It also enables a DSP to acquire the data at blazingly fast speed, so it doesn’t negatively impact its own logic and internal processes. In our packaging of this data, we got very good feedback from across the industry for this type of model.
What are the sources of Factual’s data?
We’ve been building location data for seven years now. We’ve put a tremendous amount of engineering resource on it, and we have a number of partnerships that allow us to build the high quality data that we have. The partnerships come in many different varieties. So we have companies that come to us and push us highly accurate information about their businesses globally. So that is an example where the data is pushed to us. And there is one that is perhaps that key distributor of such location data. We benefit from people coming in and bringing data to us. We also have a write API that allows a developer to push data to us in real-time. We’re also very thankful for a great number of partnerships where companies want to push data, and ensure that any business that wants to be found will be found, because it gets pushed to us. We also have a tremendous number of technologies that are able to validate and verify information using a range of other sources and it helps us keep tabs on what is fresh and accurate.
There has been quite a bit of discussion regarding the quality of location data recently. What does Factual do to ensure data quality?
We put a huge amount of time into this. There is nothing more important to us than data quality. It’s in our name; it’s our goal to continuously raise the bar on quality, especially the data that we distribute. We’ve built a location validation stack that gives us quite a big edge when it comes to ensuring that the device location is representative of where people are, so that’s on the mobile device side. In terms of the core place data, I just discussed with you the tremendous work we do to bring in a variety of sources. We also have a lot of algorithms and machine learning technology that let us sort through millions of sources and automatically pick the best information and rank the best sources. If you don’t do it automatically, it turns out to be a very expensive manual process. We have made a lot of building automatic. If you have a hundred conflicting answers about, let’s say, what kind of food a restaurant serves, how do you automatically figure out the correct information? So this is where more than half of our engineers work.
“There seems to have been an inflection in the last 12 months, where advertisers are moving from geo-contextual or location-orientated – something that is experimental – to something that has been proven.” – Gil Elbaz, Factual
Are there any new emerging use cases for location data in mobile advertising?
There seems to have been an inflection in the last 12 months, where advertisers are moving from geo-contextual or location-orientated – something that is experimental – to something that has been proven. So there is definitely a transition from insignificant budget relying on this type of targeting to bigger budgets. We are learning all the time, all the different use cases and types of campaigns that can benefit from geo-contextual. There are many interesting ones, so being able to understand patterns of activity – such as people that hike or people that are looking to buy a car – and the types of learning that we can put on that, types of machine learning we can place above that data in order to build very high quality, very personalized, very accurate audience segments. It seems to be gaining a lot of traction.
The way that we would like to think about that is like this: location data generated from a consumer’s mobile device is becoming the most powerful set of contextual data, and marketers and brands can now paint a picture that is the actual consumer journey through out. You can start learning about behaviors, and you have the technology to really find out what it is this person truly does, where they go, and why they go there.
Yeah, I love how you phrase the geo-contextual consumer journey as a physical consumer journey.
Yeah, the consumer journey, the path to purchase. Consumers may start their journey with a retailer’s website, and then they check the price on Google. Maybe they go in-store, or maybe all of their shopping will be done online. But mobile data creates a physical map of what a consumer does on a day-to-day basis and that is powerful.
And that data has been underutilized. There are a lot of apps that collect location signal. But that signal was just a latitude and longitude. It’s a piece of information that doesn’t necessarily – without a layer above it providing context – give you the additional information you need to make real decisions.
Ok, now two personal questions. What was your first mobile phone?
It was so sleek…a very cool phone. And then second question, what’s your favorite mobile app at the moment?
I am in love with this vision that apps are going to be predictive, and they are going to know me even better than I know myself. And so we are going to see a ton of innovation there. But the app that stands out right now is Google Now. It leap frogs you in terms of guessing what you are going to or should be searching in the moment, and provides all sorts of useful information. For example, letting you know that you are going to be late to your flight because it figured out for itself where you are going, what airline, what flight, the path from where you are right now and then the traffic along the way.
Yes, I completely agree with you. Product people strive to create products that delight the costumer. Google Now is the app that most recently has created moments of surprise and delight for me and it proactively pushes me data that is useful and relevant.