We sat down with Genesis CEO Mark Yackanich and CTO Rob Schwartzberg, as well as Smart AdServer US General Manager Romain Job, for an expansive talk on leveraging data for smarter outstream video placements, the potential for transacting against time (my pet topic), and overall how advanced infrastructure with server-side integrations and delivery could greatly improve the ad experience for buyers, sellers and users.
GAVIN DUNAWAY: So can you explain to me a bit what Genesis brings to the table? You suggested you can do real-time decisioning on the units that appear on the page?
ROB SCHWARTZBERG: We work with publishers to decide what formats they want to use. As those formats are served to pages, we record data about how they perform. There’s a constant adjustment.
MARK YACKANICH: Publishers in general are managing such an incredible amount of complexity, but the structure of monetization doesn’t change. You have a template, you nail an ad unit up here, and then there’s a scrum for those ad units. Why is that? It’s an artifact of the way newspapers and magazines were structured, and of the limitations in technology. Outstream gives us an opportunity to do more dynamic placements.
GD: Do you work more with the buy or sell side?
MY: We’re very supply-focused, but we’re working with everyone. Brand advertisers are actively looking for solutions for balancing ad performance with viewability at scale. If we can provide the publishers a systematic, mathematical way to provide those solutions to the agencies, then the agencies will buy. And tied to that is an understanding of the content.
Any site has a variety of content that creates different types of engagement—really high-quality articles where people spend a lot of time, and other pages where there just isn’t much time spent. Understanding that content more deeply provides a better foundation for the health of the publisher.
We can target to essentially any behavior you’d want. The issue sometimes with highly targeted audiences is, you don’t get the scale. So by concentrating on the format, as opposed to the audience, the scale stays the same, but you can still optimize toward your KPIs.
GD: How do you and Smart work together?
MY: We use Smart’s ad server and their SSP, and we’ve been actively engaged in helping build out that SSP. Smart has given us the flexibility to do a lot of customization. We can transact with any buyer out there, and we’re working actively with Smart to set up server integrations with all the major DSPs.
ROMAIN JOB: Genesis has great technology, but what they seek from Smart AdServer is a base of operations. Thanks to our legacy, we’re very good at providing a base of support. We have an independent platform where you can bring some supply and bring some demand, and we will facilitate the transactions. But we don’t have a media sales team or package ad inventory for sale—we’re demand neutral.
Genesis was facing a few major challenges—other major platforms didn’t have the flexibility to enable them to access the data points they required, or to customize the integration for optimal performance. There were technical reasons, but many purely business-oriented ones: vendors did not want to share data they thought they owned. We don’t hold back data, which is a major aspect of our relationship with Genesis.
MY: Our relationship with Smart is the hand-off—we’re ad decisioning. We’re focused on how to find the right spot to place ad formats—how do we do so dynamically and thoughtfully in such a way that everybody wins? And once we’ve created a placement, that is actually flowing into Smart, and Smart is then the arbiter out to our partners to ensure a transaction manifests itself in value to a publisher. Bottom line, an ad is delivered and the right ad is delivered.
GD: Genesis’ focus has been outstream, which has a reputation as a disruptive format. It sounds like you guys are trying to fix that through data.
MY: Outstream is good because there needs to be an evolution of preroll. There needs to be a way to figure out how to bring more value to the publisher for the creation of great content and engagement.
Frankly, outstream also gets a bit of a bad rap. Done well, it accomplishes something for all of the parties. It allows the reader to not have to pay for content. It can drive a much greater value back to the publisher. And from an advertising perspective, right now advertisers are desperately seeking big-player placements with sound on. There’s this kind of logjam that’s happening right now around all kinds of policies and thoughts around ad experiences.
I don’t think outstream is perfect, but I think it’s an important part of the solution. We like to think of what we do as outstream with intelligence.
GD: If you can have so much flexibility with the page, what challenges you? Does this affect forecasting?
MY: The forecasting is relatively straightforward. In-article is allocated a certain amount of impressions, using any variety of formats that we decide. There’s an allocation in there for no serving as well, because 10% or 15% of the opportunities we see just don’t make sense to touch.
GD: What are your views on major programmatic video trends?
RJ: The programmatic video space is migrating to a server-side implementation as we talk. It’s still not the majority of transactions, but we’ve been working extensively on this with Genesis as well as other clients.
The largest part of the programmatic video market still rely on old-school tag waterfall logic. This is stalling the emergence of a healthy ecosystem as it prevents both supply and demand sides from having the impression-level control they need. A true programmatic video market with server-side auction logic is just emerging and will shake up the status quo with a lot of benefit for both sides.
GD: With programmatic video, one of the main points was people didn’t think the placements were high quality, and they were getting a ton of fraud for a while. Has that gotten better in your experience?
RJ: These kind of things can be solved with more server-to-server connections between the supply and demand sides. With these setups, both sides get impression-level data and the value chain becomes much more transparent, preventing rogue arbitrage as well as the easiest fraud implementations.
MY: Fraud is effectively a really smart programmer figuring out a way to hack through the defenses we as an industry built up. It is reasonable to expect we bake the realization of fraud into the investments we make in digital media.
That said, we’ve taken a massive amount of it out of the ecosystem. A lot of bad players have been taken out, and a lot of measurement has been put in. Measurement and verification have now been embedded in open exchange transactions. The buy and sell sides now can have a more informed conversation about what’s actually happening at the campaign level, relative to delivery.
Also publishers are embracing programmatic. It’s no longer a remnant strategy—there’s a realization that it is the primary strategy. The industry has really turned over in the last year, year and a half. There’s a difference in the sophistication of the buying community, depending upon region and agency in some cases, but by and large we’re seeing much of our business come out of programmatic.
GD: Is there any appeal to you around time-based guarantees or attention-based guarantees?
MY: Nine months ago, it was at best a glimmer in the eye. We’re seeing some more structured attempts now. How quickly it’s coming is the big question. Conversation hasn’t yet yielded a marketplace. We’re in the early innings of experimentation, but promising experimentation.
RS: If you look at the history of what people want, it gets more stringent as time goes on, so measuring something like seconds in view makes sense, and we’d welcome it.
RJ: At the very least, these metrics are now being watched more thoroughly. SAS provides our own framework and integrated third-party tools to help pubs provide more accurate insights. It is not yet ”the currency,” but time has become major KPI.
GD: Have you figured out a way around latency, or is it still an issue?
MY: Latency for everyone is an issue. We all strive to make the lightest-weight formats possible. The thing is, we can only do what we can do. The actors upstream–the agencies and the creative shops and in some cases the brands–have, for a variety of reasons, decided to make creative asset that we have to wrestle with downstream. It’s a shared responsibility to clean it up. On our side, we’re fast.
RJ: Again, server-to-server is a key part of the equation. Latencies mostly come from the video waterfall, and if the auction happens within 100ms server side and the ad player uses the proper caching logic, 90% of the latency problems are solved.
GD: I wanted to finish off by asking about the bigger picture. It seems like you guys are pushing toward a system where you could completely customize the ad experience on a user-by-user basis. Is that the larger goal?
MY: The larger goal is to create a framework that will allow publishers to make better decisions, page by page, person by person, about how they should treat people. Should they be recommending content? Should they be monetizing aggressively? Should they be not touching the customer at all and growing the value of that user rather than harvesting it immediately?
This is something that Facebook does very well. They think about each individual user and each individual interaction and say, “Okay, in this particular instance, I’m going to grow my relationship by providing more content. Okay, I’m growing you enough. Now I’m going to harvest. Now let’s put an ad in there.” I think the publisher should have a similar philosophy about how they treat users.