Mike “JB” John-Baptiste is CEO of Peerset, an advertising technology company.
AdExchanger.com: What is Peerset?
MJB: Peerset is an ad targeting service focused on leveraging social psychographics to connect a marketer with a user. Social psychographics is a category of consumer data we are helping the market understand and embrace. Think about it most simplistically as user generated content that is most reflective of people’s interests (better than polls and better than tracking web visits). Core to our IP is that our software understands the relationship between human interests(i.e. based on 5 things someone says, we can predict their other likely interests). Our first product, Peerset Advertising, can be used by marketers, agencies, social networks and ad networks to predict the likelihood that a specific user’s interests are matched well enough to a specific advertisement to recommend that ad to them. Peerset can either deliver the ad or deliver the user and have other services deliver the ad.
Discuss your competitive set.
We believe that the competition is less about a specific company and more about “business as usual” in how marketers and agencies plan and execute display and search advertising. We believe that consumer’s interests should play a bigger role in discovering what options are available to advertise online. We believe that beyond the endemic websites that make sense to an advertiser it is important to market to an audience wherever they are spending their time and therefore the WHO(the person) should be considered as important as the WHERE(the website). We support initiatives around re-targeting, real-time bidding and ad exchanges as efficient vehicles to access relevant users.
In that we leverage, social data in advertising, we do get discussed in the same breath as companies like 33Across, Media6 and Lotame. Those companies are all making bets to one degree or another that if you can identify Influencers in the social web then you can be more efficient with your advertising. Some use the social graph and others don’t but it’s still about the Influencer. Peerset doesn’t have anything to do with the social graph and with identifying Influencers. Peerset is focused on Interest data and predicting relevant audience segments and individuals based on Interests.
Facebook and Myspace we hope will be customers of Peerset and two years ago we filed a patent for our science and service in this area. They may try to create their own Interest-based predictive systems and I would consider that to be competitive and inevitably a violation of our patent. We are speaking with Myspace.
How do you see display advertising evolving over the next 12-24 months?
I’m not the type of guy that likes to predict things but I can tell you what I hope will be the key drivers to increase more dollars into online advertising.
Creative – I hope that advertisers don’t think of display advertising as throwaway and reach and rather continue to invest effort in learning about consumers and their web behavior to improve the messaging and design of video and non-video ads to keep users engaged.
Measurement – I hope that the click-through as a standalone metric continues to lose its popularity and that we as an industry find other metrics and data that can be used to constantly improve the marketing experience online
Targeting – I hope that marketers continue to diversity their media allocation mix away from picking websites and towards audience sets. To do this they really need to understand the consumer at the individual level and
that means analyzing relevant data.
Can you give us some detail on how you create your psychographic datasets? Please feel free to get in the weeds. And, is their a shelf life for these datasets?
Since 2005, Peerset has been collecting and processing human interest data. Initially this data came solely from crawling the social web for anonymous non-PII UGC. Data record #1 may include (Age=37, Gender=Male, Location=San Francisco, Interests=Tennis, Soccer, 30 Rock, Toronto Maple Leafs) whereas data record #2 may include (Age=24, Gender=Female, Location=San Diego, Interests=Surfing, anthropology, Twlight, chocolate, Job=researcher. We built a proprietary system that studies millions of these records.
There are two types of datasets.
Dataset Type A – Interest Correlations – every single keyword(interests) in our system is associated with every single other word in our system and so there is literally a rank order from 1 to x,xxx,xxx that demonstrates the correlation of interests. You can imagine then how this info would be used when considering whether or not an xBox360 advertisement should be shown to a user that lists things like birds,John Mayer, Starbucks, etc. in their profile.
No shelf life for most Interests. Our data shows that interests don’t change a lot. If you like Scarface today, you’ll like it Scarface five years from now likely.
Dataset Type B – User profile – where we have partnerships with social media companies or other entities with user level data, we create anonymous unique profiles in our system that include explicit information and predicted information on each one of them. So if Heinz comes out with a new mustard brand and they want to start an awareness campaign for it, we can dynamically create a user segment of people interested in Mustard or Heinz or Hot Dogs or….
–Key to understanding a big part of our value is how Peerset brings scale to targeting. A huge limitation to the targeting sector is that most services don’t scale well. Why? Because they rely only on actual user action (i.e. visiting a travel site and typing in Las Vegas). Peerset only relies on explicit action or data as a starting point. We expand from the explicit data to predict a much broader group of interests and therefore targets. Try going on Facebook to their self-service ad platform and type in a keyword to see how many users they can find with that keyword. The limited number of users that query will give you is a good example of how current targeting approaches sabotage themselves because of their inability to scale.
What is the sweet spot when prospecting for clients? Are particular verticals strong, for example?
For agencies – our solution works in most marketing segments. Some verticals might be larger than others (i.e. Cosmetics vs. Equestrian Sports) but that has more to do with the market size vs. our technology limitations.
What’s making brand (awareness) marketers hesitate in bringing their brand budgets online?
Current measurements are not a substitute for experience and I believe that budgets haven’t shifted quickly enough to the internet because the promise of the internet was about measurement and perhaps that expectation was always unrealistic. You can accept that someone watching live tv or listening to live radio will experience the ads. You can’t prove it necessarily but marketers accept this anyway. The same cannot be said about a typical banner ad(marketers will not assume that the banner will be noticed) so the click and the action has been forced into the spotlight. For me it comes back to my point about (Question #3) what I hope for. If marketers get the ad creative right and get the ad in front of the right person we can take pressure off of the old metrics and accept that viewership are noticing those ads alongside their content experience.
How does Peerset help the publisher?
- Money – placing targeted ads we secure from advertisers
- Targeting technology – giving publishers the opportunity to optimize their own inventory by making better recommendations of ads to users based on user psychographics(interests)
- Insights – we can help them understand their audience more completely by:
- Packaging their users into marketing segments
- Profiling users for general business purposes
Is it just social publishers you’re helping?
Our value prop for social publishers is multi-layered. For non-social publishers we are helping them indirectly via ad networks and exchanges who can re-target our social partner’s users. Consider however that over time every publisher is a social publisher or should be as traditional publishers incorporate social tools (bookmarking, comments, chat, blogs, video, FB connect, etc.).
Are you buying through exchanges and yield optimizers to retarget users? How is that working? Any success or challenges you can report?
Actively in partnership discussions. Nothing to announce yet. We see these relationships as the Fulfillment component of Peerset Advertising. Many of these companies also look to us as a source of data that they can use to help bring more intelligence to the cookied users that they are posting on exchanges for advertisers to acquire.
Peerset re-branded about a year ago from “Ontogenix”. What was the reason?
The name was dorky. The name change was before I joined so I don’t have all the details. “Onto” was for Ontology. “Genix” I suppose was to suggest the science behind our IP. Peerset takes “Peer” for more the techie, peer-to-peer notion and “Set” more for data.
How has momentum been since that time?
I wouldn’t tie momentum to name change. For me momentum has been mostly around the point in time earlier this year when we recognized that marketers are still using primitive approaches to plan and reach their target audience. We can make a big impact here so long as people get to know of us. 🙂
Are their non-internet advertising applications for Peerset technology?
Absolutely but we are not focused on any of them for the time being in the spirit of focus. If you consider that at the end of the day our core IP is about our understanding of the correlation of human interests and our
ability to predict things knowing that there are many applications to get involved with. Think about big research sectors such as Pharmaceuticals or the intelligence sectors such as Homeland Security. Also think eCommerce and consumer applications for things like dating or job search.