"Discovery" has become one of the hot new alternatives to the more common forms of behavioral tracking. One company in the space, MatchMine, lets users create MatchKey profiles of their
content "likes" at MatchMine.com or one of its partners like music site Fuzz or film site FilmCrave. The technology creates this profile from the ways a user interacts with different
content. Opted-in members can use the MatchKey only at a specific site or bring it with them to other sites in the growing network to inform content recommendations. We caught up with CEO Michael
Troiano as MatchMine evolves from its test phase with early partners to a more expansive rollout with new partners Odeo, blogged, and others. Behavioral Insider: Distinguish
discovery from some of the other behavioral targeting approaches out there.
It is even broader than behavioral targeting. BT is really a subset of online targeting
behavior. It seems almost ironic that we have seen the reality of online targeting catch up with the original promise [of online advertising] in being able to target people by who they are but in a
way that consumers are almost a little creeped out by.
[For example, there's] the whole fiasco around the original iteration of Beacon. In round one people felt it was intrusive and
they felt anxiety about being targeted by someone who had an intimate understanding of you. I think there is an opportunity here for something different to step in. Discovery enables a third
generation [of advertising], to target people by what they like. It is in many ways more durable than targeting by what they want [i.e., search]. What I want now may not be what I want five minutes
from now. But what I like is reasonably constant compared to what I want. But it is not as intrusive or intimate as targeting by who I am, so I am more willing to share what I like to third parties to
act on. BI: Determining what people like can be just as creepy, though. And At MatchMine I have the hurdle of registering and opting into bringing the MatchKey to other
No question that involving the user in the process of discovery, to the extent that MatchMine does, creates this execution risk. We have struggled with this a
little bit and there has been some pressure to go in the cookie direction. I really don't think that is in the interests of users or of MatchMine. I think there is a much simpler way to split the
difference there. The vast majority of users will encounter MatchMine's technology in the context of a partner site, where they will see it as an integrated feature. So in that way getting the
user experience as an integral part in a partner site is absolutely fundamental to an opt-in model that keeps the burden of participation as low as possible.
Now, you don't get the full
benefit of that key if you only use it in a single partner. We are not siloed discovery. If you want to use that key in more places, then you need to save it to the Web. And that is the focus of
MatchKey.com. It gives you a place to save your key so you can use it all over. But getting someone to do that after they have had a positive experience at a partner site is a very different value
proposition than trying to get them from zero to 60 by going to matchkey.com. I agree, that is a model that will not scale, but that is not our strategy.BI: How does the user see
the benefits of that opt-in so the value exchange is apparent?
If you go to the music section of Fuzz, you will see ‘Get Recommendations,' not
‘here's MatchMine.' It will analyze all the stuff you have rated or interacted with in a positive or negative fashion and generate your "key" with a single click. It creates
what we call a local key. You will start to get recommendations within the frame of that partner site using that local key right away. BI: On the publisher site, is it up to them
how that set of user ‘likes' is used?
The display layer has a lot of flexibility for the publisher. We found in the beta period that people are very interested
in where these results come from. A significant percentage of people will click down to see why something was recommended to them and learn that there is an enabling technology under this.
The user is going through a series of stages when they encounter MatchMine. The first is that they see this as another feature inside a partner site. Then some percentage want to know what is going
on. Then they will learn that it is a platform. That is when we introduce the idea of portability and when we offer the call to action to save the key to the Web and use it wherever you like. We think
the highest order of participation is a state we call avatar. Because MatchMine results are a visual representation of your interests and taste, we see people being able to connect to that on an
emotional level and a desire to share their key with other. That is where we want to get over time.BI: What benefits do publishers realize?
obvious is in incremental page views. We have begun creating ways to deliver some kind of a CPM premium. It may be leveraging the preferences in the key in some sort of affiliate model, recommending
content of interest to people in a transparent way. You would be on user-generated content at a CPM that averages 50 cents because it is untargeted. For a MatchKey holder who is participating in that
space their preferences are well understood by us and we can serve up an affiliate ad, like an Amazon offer for season 1 DVDs of ‘Battlestar Galactica.' The effective CPM we can deliver back
to the partner is much higher.BI: On the user side, he is accruing this profile he actually can see. What are the possibilities for consumers leveraging their own data in different
Absolutely. We are active in the Data Portability Workgroup that advocates the rights of users to move their data around the Internet as they see fit. It advocates
open standards for the representation of things like identity. We are to the point where we can open up in both directions.
We started this week the ability to feed your key with your
last.fm profile. We are trying to empower people to feed their key not only with partners in the network but with any source that is available. We will launch a more generalized RSS solution that will
make it easier to import your YouTube favorites or Amazon Wish List to your key.
The other side of that coin is to let people create an output of their key in an open format. We are
helping shape the development of what is called APML standard (Attention Profile Markup Language). The goal is a standardized RSS format that enables people to express their interests and tastes in a
way that they can use all over the Web. You can't be committed to one side of that without being committed to the other.