Affectiva Bids For The Emotion Attribute

By
  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

AffectivaAffectiva CEO David Berman recently spoke to AdExchanger after the announcement about his company’s most recent round of funding. Though it’s early for this notion at Affectiva - think emotions for audience targeting, potentially.

Berman dug into the details…

AdExchanger: First, can you review how strategy over the past year has evolved at Affectiva?

DAVID BERMAN: A year ago (2011 AdExchanger Q&A), we took funding that included investment from [agency holding company] WPP Group. They have a $4 billion division called Kantar and provide measurement capabilities for all the WPP companies.

WPP saw the strategic fit to objectively measure emotion, at scale, on any device. The way they currently do it - as it is in most market research companies - it's primarily survey‑based. We have a new way to get high value data passively. We took the investment to build out some applications to work with them and, specifically, we announced a partnership with Millward Brown at the beginning of the year.

Millward Brown is the leader in the states for ad measurement. We rolled out the product earlier this year and have been getting fantastic traction. We just announced the roll‑out in their emerging markets, too.

With this latest round, we weren't looking for funding, but we were approached and it gave us an opportunity to go deeper and wider in market research and also take this type of technology to consumers who want to receive content that's more engaging.

We're also going to be doing some more advanced development of our Q Sensor product, which we talked about last year as well.

So really, it's an expansion round. It's $12 million of new capital, our total raised to date is $17.7 million. We're thrilled to have such marquee investors like Horizon Ventures, which is Li Ka-shing's fund, and Kleiner, Perkins, Caufield and Byers. Frank Meehan is going to join the board from Horizons and Mary Meeker, a partner at KPCB is going to join as a board observer.

We were fortunate to be in her 2012 Internet Trends report which got us a huge spike in traffic as a disruptive technology. So we're using the money to accelerate development of the Affdex platform and Q Sensor.

Regarding strides you’ve made with Affdex, what's the reality of the platform ‑‑ how can marketers put that technology to use for themselves today and what do you see down the road for the marketer with the platform?

Affdex will allow them to test any content, literally, in seconds at scale. If you want to test a spot and see how it's going to perform around the world, we can do it in seconds and the measurements that we give are unbiased.

There's a high cultural bias in surveys. People generally tell you what they think you want to hear during a survey, but with our product you just read the faces. And the faces don't lie.  And if you don't believe it, we can show you all the faces.

There's a live demo available - the Super Bowl app - which is a subset of the technology. We're tracking many more emotions now. The database is now the world's largest database of human emotions, which is fantastic. Getting the information from the different countries is enabling us to build out the cultural nuances, which is powerful. We were testing an ad and it had great response in the US and Europe, but bombed in Latin America. Before they spent millions of dollars, they knew it, and were able to run a different version of the ad in Latin America.

Can you review a use case on how the product and the data capture works?

Sure. From a participant perspective, you need a webcam and a browser and then you click on a link, click on an ad, and then it's going to ask you to opt in. (Try it here.) All the ads we test are opt‑in, so people have to know that they're using it. The camera comes on, you watch the ad, and at the end it will aggregate all your results and be able to tell you the parts that you liked, the parts that you didn't like, or if you were confused or frustrated. And then it correlates all that with survey data, and you get an interactive dashboard.

What the marketers like is being able to see the trends and the specific “scene‑by‑scene” where the emotions change, and you can break it down by gender, age and location. We're working on an app that is a political debate app. As the different political commercials come on, it can gauge responses in battleground states that are pretty insightful.

Thinking about sort of the incentive part for the consumer, why do they participate? Do they receive a Starbucks card or something like that?

With Millward Brown, we use some of the panel companies and there is some sort of compensation for the people We do find that people like to share how they feel, especially if it's engaging content. We think there's a nice social piece to this and when we check in [with AdExchanger] again, hopefully less than a year from now, you'll be able to see some of those social apps that we have out there.

Picture being on Facebook - instead of having to "like" it, picture being able to show how much you really liked it and be able to show a full range of emotional responses. We joke here that the “like” button is going to be obsolete and be able to share how people really felt while they saw that photo of a friend from the past, for example.

It sounds like another way of panel‑based measurement around emotion similar to the way comScore and Nielsen do panel‑based tracking. Do you like that comparison?

I'm not an expert on how they do it but what we bring to the table is a passive way of doing it. You don't need to click anything, you don't know need to collect diary information, you just watch it. It's easier for the end user, and the marketers are able to get valuable results. There’s a science behind this, called the Facial Action Coding System, that we talked about last year.

How big is a panel for a particular survey?  What are the expectations in terms of extrapolating insights for a larger audience?

We find - because of the way we're built with an Internet capability - the panels can be much larger than they used to be. A lot of this is the qualitative business where they're doing a lot of onsite stuff. So we think the panels can go large. We can work with any size panel. It depends on how they've been doing it historically.

What would you say the usual panel size is?

It really varies per customer. The average is probably 150. You can go bigger and you can go smaller.

Back to the consumer product, why “go consumer”? Is it another way for you guys to get data to drive the B2B side, for example?

The primary driver is to get the technology out there. Get people comfortable using it, and letting them see the value. Knowing about yourself and being able to get content that's relevant and interesting is a real big win for consumers. That can transfer over to the B2B play, but we just want to get the technology out there, and get people building on it.

We think there's a huge opportunity to work with the social companies in sharing content, and making the content more engaging.

There's a big opportunity to include this stuff in gaming and changing your response to the game, based on your emotions: to know when you're intense and then be able to improve the content of the game, for example.

The way we're building this thing, and our vision from the beginning, was to build an effective data platform where we could use this data for multiple market segments and emotion‑enable all types of content.

It's starting to come to life. This investment allows us, as I said, to go deeper into the market research. We're going to be expanding into some consumer offerings to get consumers using the technology in an opt‑in way.

We think there's an opportunity to leverage both markets for the business. This allows us to have the runway to do that.

It feels like you’re turning the human into his or her own “remote control” in a way – removing the physical remote with the emotional.  Like that?

Yes, that's where it's going. It's personalization of the content that's going to be highly relevant and engaging. To go to your TV and know that your Affdex score on this one is a 10 out of 10, I'm going to watch that, versus the Netflix rating that I have right now. My kids messed up my [Netflix] rating system so now I get zombie movies.

It's going to know who I am, it's going to know what I like, and it's going to be able to help me sift through the amount of content and get the stuff that I'm going to enjoy the most.

Follow Affectiva (@affectiva) and AdExchanger.com (@adexchanger) on Twitter.

  • Facebook
  • Google Plus
  • Twitter
  • LinkedIn

Email This Post Email This Post

By on at

Leave a Reply