Jeff Hirsch, CEO of Audience Science, formerly Revenue Science, name-changing and leading behavioural targeting technology, you’re obviously flattered to be interviewed by Goallover! But joking aside, let’s get to some questions: Jeff, can you tell us a little about your career before you became CEO of Audience Science?
Immediately prior to Revenue Science (now Audience Science) I was the SVP of Business Development for ValueClick, Inc. Before joining ValueClick, I was a founder and Chief Revenue Officer of Fastclick. I’ve been in media and marketing for over 30 years.
When did you join Audience Science – and how did the appointment come about?
I joined Revenue Science in September, 2006, originally as Chief Revenue Officer. My original charter was to modify our business model to develop an audience marketplace to provide the company the ability to sell to marketers directly in the US, to compliment our targeting technology business.
What position was the company in when you joined, and how has it advanced?
Our primary business was supplying targeting technology to premium web properties around the world. The technology allows these publishers to sell behavioural targeting to their advertisers. We are paid when the technology is utilized. Since then, we have developed a business that is complimentary to our technology business that allows us to more directly address the marketer side of the equation.
Just to be clear, you mean that you now aggregate the inventory of your technology clients, and sell across their properties to marketers and advertisers directly?
Our technology clients are now able to go beyond the confines of their own site: they can reach their users across a network of sites; they can export behavioural data to other publishers; they can import behavioural data to enhance their own segments. Basically, our technology clients are able to create virtual networks of audiences easily and quickly. In the US, we also participate as an audience network, working directly with advertisers to stimulate the entire BT market.
When we have emailed each other in the past, you mentioned that there was a value for the inventory, but also a value for the information. Are you saying that you could effectively target the audience of one of your clients, but while that audience was on another website, but that you remunerate the client website for providing you with that permission and information?
Publishers spend a considerable amount of money creating content to attract consumers. Doing so provides them with two key assets that they can sell to marketers: 1) ad space or inventory; and 2) audiences. Most of the work that publishers do is around maximizing the yield of their ad space. We feel that doing the same around the audience they have created is equally important. If a publishers audience is going to be used for targeting by anyone other than that publisher, that publisher should be compensated accordingly. We will see more sophisticated strategies around audience yield going forward and Audience Science is uniquely positioned to participate in that ecosystem and provide a publisher with the tools to maximize this opportunity.
What do you think stands you apart from the other behavioural targeting technologies?
We were originally an analytics company. That genesis means that our data storage and management capabilities are significant. We were a technology company first, before entering the business of applying that technology to media. This means a more robust backend, and subsequently scalability, than other companies. We have over a dozen technology patents and dozens more pending.
When you say “more robust and thus more scalable than other companies”, can you be more specific? I’m asking because I am surprised a competitor would share their architecture with you.
A recent report by Attributor (www.attributor.com) showed us as the #4 adserver in the world reaching over 385 million unique users each month. No other company in the behavioural targeting space can claim numbers anywhere close to these.
Do you think that some of your competitors are responsible for bringing negative press to the field of Behavioural Targeting?
There is a lot of confusion around Behavioral Targeting. We still need to iron out standards in terms of nomenclature and certainly in the area of data rights. The use of data encompasses both who has the right to utilize it AND what consumer rights MUST be protected. There are some companies who have crossed the self-regulation lines and have brought negative press to the entire industry.
With reference to the Thelma story here – how many datasets do you believe are required before a behavioural targeting company can accurately “profile” a user?
That depends on the data. A search is a behavioural targeting event with an immediate display of relevant content. A search is very specific. When you start looking at web surfing behaviour to develop a profile and ascertain interest or intent, we believe that you must be able to combine data elements to more fully understand the consumers’ mindset.
I guess what we are asking is – do you believe that there is a tipping point at which behavioural targeting can become user profiling, without the user being aware they are being profiled?
This certainly can occur. Part of the need for standards is around properly informing consumers as to the data that might be collected.
And does Audience Science have a public policy on where they suggest the line should be drawn?
We do not collect personally identifiable information. By far, this is the most important building block for creating a consumer friendly environment. Beyond that we require notice by publishers as to any data collection activity and provide the ability for consumers to opt-out. The NAI in the US has established guidelines that we believe will drive this conversation.
With respect to the European Union’s “discussions” with Google, MSN and Yahoo, what is Audience Science’s policy on cookie life, and anonymous information retention?
Since we are in the business of understanding interest and intent and combining that with recency, we believe that shorter periods of time are required to drive relevant advertising than what we have seen by other players.
Again, does Audience Science quantify that “shorter period” and have any suggestion for what is long enough to be commercially viable whilst short enough to satisfy government officers?
This is often a function of the type of audience and campaigns that are running. Our default storage period is 90 days.
Jeff, can I ask you some candid questions? You have a reputation for asserting that Audience Science is more ethical and transparent in its operations than your competitors. Can you give me specific grounds for these assertions?
Each of our data sources are contracted for directly. Each of our data sources MUST comply with NAI (www.networkadvertising.org) standards around privacy and disclosure. Although this is a US organization, we believe in the concepts they promote. In no cases will we utilize data that has been generated without the consumer having been informed and having the ability to opt-out. On the business side, we believe that publishers that create content are creating two assets that can be monetized to advertisers: ad space, and audience. With specific contracts around data collection, publishers are compensated for the value of the audience they create. We believe this to be essential in protecting the ecosystem that exists around free content in exchange for advertising.
How exactly can you value each publisher’s contribution if you are consolidating data across so many publisher data points? Does each publisher get rewarded? If 100 publishers contributed to the dataset, are you saying 100 publishers would be remunerated for the subsequent revenue you could generate?
That would be highly unlikely to occur, however if the audience segment was made up of multiple data sources, then each data source would participate in the revenue that they generated.
Do you believe there are activities being undertaken by competitors that are or may constitute a breach of UK or European Union data laws?
No comment (This question should be asked to our competitors directly)
Do you retain a legal counsel who is expert on UK and European data laws?
We do.
In your opinion, what is the biggest threat that the entire privacy debate poses to the online advertising sector, and specifically to Behavioural Targeting companies?
The collection of personally identifiable information and/or sensitive information such as health issues, and the inability for consumers to make a choice by opting out of the collection of information for behavioural targeting ads.
Do you think companies and technologies such as Nebuad and Phorm have any greater ability to abuse personal privacy rights than the large search companies?
We believe that the “disconnect” between content and technology creates a potential gap in a consumers ability to know what they are opting in or out of in terms of data collection.
And in closing, please can you explain to us how you ensure that your partners are adhering to UK and EU data laws when they implement your technology? Is it your responsibility or is it the responsibility of the publisher?
It is contractually agreed to between us and our publisher partners, the implementation of which is managed by the publishers themselves.
It seems like you saying that partner publishers are responsible for ensuring that the end consumers are made aware of their interaction with Audience Science technology? If a publisher provided links to a privacy policy that was deemed inadequate by (for example) the British Data Commissioner, are you saying that only the publisher and not Audience Science would be liable for any breach?
Audience Science requires in its contracts with publishers to maintain an adequate privacy policy. If Audience Science becomes aware of any violation of this contractual obligation, the NAI Principles require Audience Science to make reasonable efforts to enforce the contract (either by requiring the Publisher to correct the privacy policy or terminating the agreement with the publisher).
I am trying to ensure I am 100% clear on this point. Given the assertions of “collective responsibility” currently being levelled at Nebuad and the ISP’s, do you think that the balance of responsibility can actually be determined by Audience Science and her partners, or do you think the precedent actions suggest a collective responsibility exists?
We believe that there have to be firmly established guidelines that appropriately protect consumers while providing the means to support the value proposition that advertising supports – free content in exchange for advertising. The ability to accurately target ads is key to this agreement. Industry players should self regulate against these standards.
Jeff, thank you so much for taking the time to talk to Goallover and explain the Audience Science position. Certainly, commentators are suggesting that behavioural targeting is a “smoking gun”. We appreciate your clarifications and answers, and hope we can return to ask you further questions. (If you stop changing your company name, we’ll know where to find you). As you quite rightly say there is a lot of confusion around behavioural targeting – so we appreciate your help in clearing the smoke. We wish you all the best with your business.
Jeff should be a Press Secretary for a major political figure or office. His answers are neither here nor there and are really just dodging the issues. He essentially takes no responsibility or assigns no real blame for the practices employed by his company. Revenue Science (audience science) do create user profiles and those profiles while not specific to a name can and do get very specific and can be used to identify a person.