SCOTT SPENCER: Basically, we’re going to be rolling out a few more tools to help DoubleClick Ad Exchange buyers buy quality inventory, and to check their campaigns.
Taking a quick step back; when we launched the exchange about a year ago, we engineered it with best-in-market buyer and publishers controls, as well as extensive crawl-and-verify inventory screening. Together with the real time bidder, these were the biggest upgrades we made.
As part of a long line of improvements in this area over the past year, we’re taking the wraps off a couple of additional features to give buyers even more control, quality and transparency.
The first is “Site Packs” – these are manually crafted collections of like sites based on DoubleClick Ad Planner and internal classifications, vetted for quality. These allow buyers to get a set of high quality sites for their particular campaigns, covering anonymous and branded inventory.
Second, we’re making some changes to our Real-time Bidder (in beta). The biggest change here is for Ad Exchange clients who work with DSPs. Historically, Ad Exchange buyers were hidden from publishers behind their DSP. By introducing a way to segment out each individual client’s ad calls, inventory can be sent exclusively to an Ad Exchange buyer even when that buyer uses a DSP. It increases transparency for publishers and potentially give buyers more access to the highest quality inventory, like “exclusive ad slots” – high quality inventory offered to only a few, select buyers as determined by the publisher.
Thirdly, we’re soon going to be rolling out a beta of what we call “Data Transfer” – this is a report of every transaction bought or sold by a client on the Ad Exchange. Effectively, it’s a daily log file of everything that happened. Clients can then review every branded URL that they purchased to ensure everything was what they expected.
A recent report suggested that exchange inventory is unsafe for marketers. How does Google respond?
SCOTT SPENCER: You won’t be surprised to hear this, but when it comes to the DoubleClick Ad Exchange, I disagree. We have high quality DoubleClick and AdSense publishers.
Of course, it’s always possible to find a single page among millions that is objectionable in one way. But we’ve built extensive checks to ensure the quality and safety of our inventory, including strict participation policies, continuous automated scanning of all publisher sites, and automated inventory review to identify improper traffic patterns. A flag can trigger either a human review or an automatic blocking. These validation tools apply to our combined pool of AdSense and Ad Exchange publisher inventory. They operate at a page-level granularity, pre-screening individual ad units, then report to the client in great detail afterwards.
We’ve also developed ways for marketers to choose high quality inventory, like through the Ad Planner 1000 filter, and the ability to buy or exclude specific publishers, URLs or categories of content.
What are you investing in as it relates to providing a brand safe experience for advertisers across the Google Display Network?
JASON MILLER: On the Google Display Network, we genuinely believe that marketers and agencies get access to the most brand-friendly network out there.
It’s important to emphasize that Google's technological expertise is in doing two things very well: first, crawling and analyzing content at great scale (as a search engine that’s Google’s core competency) - and second, targeting ads precisely at massive scale. We’ve applied this to the Display Network for quite some time.
We broadly invest in 4 areas to ensure brand safety:
First, strict policies. We don’t permit sites with content that violates our program policies.
Second, technology. Several layers of review - before admission to the network, the precise time of ad serving, and continuous automatic scanning - uphold our program policies. We do real-time blocking of ads if the page violates policy. We do this at massive scale and at a page/URL level - not just a site level.
Third, transparency. We offer domain - and, importantly, URL - level reporting. This provides deep insight into the sites and pages where our clients’ ads were served, on-demand. Major advertisers regularly get a report that shows what their ads looked like in the context of a given page.
Finally, advertiser controls. We give advertisers a robust set of controls to closely manage their ad delivery across the GDN, including hand-picked bundles of sites (such as the Ad Planner 1000) or individual sites, an above the fold filter, specific site and category exclusions.
Recently, we signed up to the standards and practices under the IAB’s Networks and Exchanges Quality Assurance Guidelines.
Do you allow the use of ad verification technology from companies such as AdSafe Media or DoubleVerify across all inventory of the DoubleClick Ad Exchange including the Google Display Network? What's the plan here?
SCOTT SPENCER: Yes, we allow Ad Exchange buyers to use these companies if they want to and if certified. Of course, sellers can decide whether they allow the tags - it replicates the situation elsewhere on the web. Some do, some don’t. We have no plans to change this. The Google Display Network is treated like any other seller on AdX.
JASON MILLER: We’re working through these issues methodically for our network, with a view to coming out in the best place for users, marketers and publishers. There’s very significant privacy, methodology and accuracy issues to consider, but importantly, there’s also the issue of significantly slowed ad and page load times caused by third parties’ site-level (as opposed to page level) blocking. We know a lot of other publishers are going through these same considerations. If an advertiser is getting highly inaccurate reports from a third party, as we have seen happen, and all these tags slow network performance, it’s not value-adding.
So, in short, we have provisionally certified a third party verification provider and are running some tests with them. But mostly, we’re continuing the investments I’ve described earlier and providing those services to advertisers at no extra charge. We believe that Google has a unique strength in crawling, content analysis and ad targeting across the entirety of our network and that this puts our marketers in an extremely safe place.
How do you justify for publishers the use of buyers' ad verification technology which may compromise the anonymity of their inventory in the DoubleClick Ad Exchange and present potential channel conflict issues with their direct sales team?
SCOTT SPENCER: Ad Exchange publishers make their own decisions on transparency and the allowance of 3rd party validation tags, over their branded and anonymous inventory.
The first point to make is that the anonymous inventory in the DoubleClick Ad Exchange typically belongs to the most brand sensitive publishers and is among the highest quality inventory - by definition, publishers make this space anonymous because they’re selling the same space directly. For these clients the prospect of any exposure of their brand of course makes them wary of ad verification technologies.
We unapologetically prohibit the identification of anonymous inventory. Verification companies that have adequate protection of publishers’ anonymity are more likely to be accepted by publishers than those that attempt to expose the anonymous inventory’s underlying site.
I understand that you only allow one pixel in a DoubleClick Ad Exchange served creative. Does that mean that a client needs to decide between an intender data pixel, agency pixel, ad verification beacon, etc.? If so, why limit the buyer?
SCOTT SPENCER: Again, we’re balancing the needs of marketers, publishers and users. We want to protect everyone from latency issues and from having a daisy chain of 10 or 20 pixels show on their site when an ad is delivered. That’s a lose-lose-lose scenario for user, advertiser and publisher.
Our policy is that there can be one network and one pixel delivered per call. This is to enable, when allowed by the seller, post-impression reporting or data collection for subsequent remarketing. It’s up to the marketer if they want to use a verification pixel on all, none, or a sample of their impressions.
Can you share the types of fraudulent activity in the DoubleClick Ad Exchange today? Any trends concern you?
SCOTT SPENCER: There’s always a tiny number of bad actors in any network, and our job is to pre-empt and detect them, using technology and human review. This sort of activity is bad for the whole ecosystem. Because of not wanting to tip my hand about the specific security and anti-spam measures we have, I’d rather not go into the attempts at impression and ad spam that we’ve detected - you can see some of what we monitor for here. I can say that we pick almost all of it up proactively and punish offenders, and in terms of trends, I can say that on our exchange, the proportion and volume of this stuff is falling.