In this article written exclusively for ExchangeWire, Matt McLaughlin (pictured below), COO of DoubleVerify, responds to concerns over the firm's use of "clouds" in the aim of maintaining brand safety and negating advertising fraud.
At DoubleVerify, a key component of our service is blocking. The blocking function executes just before an ad is rendered on a device, after the ad has already been effectively purchased. This typically occurs when the brand has won a programmatic auction or, on direct buys, when the ad was selected by the publisher’s ad server for delivery. DV blocking code inspects all characteristics of an ad impression in real time, including evaluating the device, impression geography and surrounding content. These characteristics are compared with the standards established by the brand and trafficked in the DV system during setup. When the impression meets these standards, the ad is “Allowed” and the creative is rendered to the screen. Alternately, when the impression does not meet the brand’s standards, the ad is “Blocked” and the brand’s creative is not rendered; in this case, the ad slot is replaced with alternative content.
As brands adapt their suitability strategies during the coronavirus pandemic (see DV's recommendations for brands here), there have been questions raised regarding DV’s blocking service – why it works the way it does, and why DV Clouds appear in an ad slot. This piece briefly covers the basics of the blocking service and answers these recent questions, demonstrating why this solution is best for brands, publishers and consumers.
Are ads blocked because the content is not safe?
Yes, but that is not the only reason blocking occurs. DV’s media authentication service allows brands to activate blocking for three reasons:
- The content adjacent to the ad does not match the brand suitability settings established by the advertiser;
- The ad impression may be invalid or fraudulent; and/or
- The ad impression may be attempting to deliver outside of a campaign’s geographical restrictions.
Why is blocking needed?
No brand wants to pay for ads delivered outside of the standards they have established for their online media plan. Often, DV is asked why we can’t just render the ad, report on impressions that either complied or did not with the advertiser’s standards, and allow the seller and buyer to use this data to determine the impact on media budget. It’s a reasonable question, and our rationale is as follows – stopping an ad creative from rendering serves several key business objectives for a brand, beyond protection of their online media investment alone.
First, blocking helps to protect a brand’s equity. Brands spend far beyond their media budget to build their brand equity with customers. Carefully crafting the online and in-store experience; planning promotions across channels; developing products and their packaging; delivering customer service and advocacy; and protecting intellectual property, are just a few facets of the overall brand equity equation. An ad placement alongside content that is illegal or not aligned with the brand’s values can damage those investments – with long-term consequences to customer loyalty and commerce. Similarly, ads that are delivered in geographies outside the delivery capabilities of the brand can damage brand equity. Seeing an ad in a location where the brand can’t actually deliver the service can be a frustrating experience for the consumer. By preventing the ad creative from rendering in these environments, blocking stops the consumer from experiencing these negative interactions, and ensures the advertiser’s investment in their brand equity – above and beyond media budget alone, is protected.
Second, blocking for fraud and invalid traffic helps maintain the integrity of the online marketing metrics employed by the advertiser. Rendering an ad creative within fraudulent conditions will certainly impact impression delivery metrics, like impression count, video completion rate and more. While this could potentially be accounted for in ad serving systems without blocking, the human-like behaviour of certain fraud schemes can be far more damaging. Bots will click on the ad, visit the brand’s website and consume numerous content pages. Marketing teams invest significantly in web analytics, conversion path optimisation, copy testing, and many other metrics-driven systems on their digital properties. Allowing fraudulent bot traffic to click through on an ad and corrupt the metrics produced by these systems wastes money and effort far beyond the media plan. By preventing the ad creative from rendering on fraudulent impressions, blocking stops this activity and helps maintain the integrity of these metrics and marketing platforms.
Third, blocking for out of geography helps ensure regulatory compliance. Many industries have licensing or other regulations that determine where they are allowed to do business. Blocking ad creative from rendering to users in geographies outside of permitted jurisdictions helps ensure brands maintain compliance with those restrictions.
Is blocking needed on programmatic buys when pre-bid targeting can do the same thing – before the impression is purchased?
DV offers the most comprehensive and accurate pre-bid avoidance targeting available in the market. Pre-bid avoidance targeting helps brands drive efficiency from their programmatic media spend by preventing their bidding on auctions misaligned with their ad delivery standards. The solution also helps publishers by preventing the sale of impressions that ultimately result in a block and potential forfeit opportunity. DV’s Authentic Brand Safety targeting ensures that an auction is evaluated using the same suitability settings (including keyword blocklists) that we employ in blocking. Our millions of fraud signatures are updated nearly 100 times per day (every 15 minutes) in our DSP integrations to ensure near-immediate programmatic protection from invalid traffic. However, despite the extensiveness of DV’s pre-bid avoidance solutions and the benefits delivered to brands and publishers, pre-bid avoidance targeting is not a standalone technique that can fully replace blocking on programmatic buys.
The key difference between programmatic evaluation of auction quality and the blocking evaluation of impression quality is as follows – in programmatic evaluation, the information being utilised is the auction data passed to the DSP, while in the blocking evaluation, DV measurement code can directly observe the device and placement environment of the ad. In many cases the decisioning information available in an auction can be incomplete (eg, URL is truncated), incorrect (for legitimate purposes, like masking or due to misrepresentation), or absent entirely (the auction doesn’t contain the complete serving chain used to evaluate traffic validity).
For example, DV recently identified that among the Top-100 US news publishers, the URL for more than one third of programmatic auctions on article webpages was incomplete – containing only the site of the impression and not the complete path. During the auction process, this incomplete URL prevents DV from fully evaluating the content beside which the ad impression is going to appear. In these cases, blocking must be utilised to ensure the brand’s creative is not rendered alongside unsuitable content.
If we could be certain that programmatic auction information was complete, accurate and comprehensive – as compared with the data available when DV blocking code executes on the device, then pre-bid avoidance could fully replace blocking. Unfortunately, the data gaps between the auction and direct measurement are still too large for brands to rely exclusively on pre-bid avoidance as the mechanism to prevent their creative from rendering on programmatic buys.
Does the buyer pay the publisher when an ad is blocked?
As an independent media authentication service, DoubleVerify doesn’t handle media payments and is not directly aware of the payment terms between buyers and sellers. Whether or not the buyer pays the publisher for these impressions is dependent on the buying method (direct vs. programmatic), definition of the system of record, terms around quality standards (geo, IVT/fraud, suitability & viewability), and discrepancy and reconciliation processes. These terms are typically part of the negotiation and outlined in the various agreements between buyers, intermediaries and publishers.
Aren’t there alternatives to the DV Clouds?
At DV, we sometimes use a non-branded rendering of clouds (the “DV Clouds” as shown in the image below) to replace a brand’s creative when blocking the ad impression. The recent increased presence of DV Clouds has raised questions about potential alternatives that could be rendered and we would like to address them here:
Question: Why not execute a passback to the publisher?
We do when we can! If the ad impression meets the minimum requirements to be monetisable for another brand (eg., the reason for blocking is not fraud, piracy or other illegal categories), DV’s preference and first course is to return the ad opportunity to the publisher to show a replacement ad. When this happens, blocking is not visible to the end consumer because an ad creative renders in the slot. Unfortunately, not all publishers have enabled this functionality for direct buys. Also, in a programmatic environment involving DSPs and other intermediaries, executing a passback to the publisher is a challenge because the buyer may not even know who the publisher is (eg., blinded inventory). DV’s system preferentially executes a publisher passback when possible, but uses the DV Clouds as the default alternative when a passback can’t be executed.
Question: Why don’t you show a PSA instead?
That said, DV will continue to explore the viability of delivering a PSA when possible. Currently, the easiest mechanism to execute a PSA ad is for DV to deliver a passback of the blocked impression to the publisher or buying platform, who then can make the decision to show a PSA.
Question: Why show anything at all? Why not just leave it blank?
Regardless of what’s on the page, most ad networks require that a piece of creative always fill a request, regardless of whether the brand’s creative is rendered or blocked. In addition, filling the ad slot ensures the page layout and user experience isn’t harmed by empty real estate on the page. Even if the traffic is fraudulent, we must show something. If we could show nothing, we would – but since we can’t, we opt to show the DV clouds.
When an ad is blocked, how are the DV Clouds a better solution for the publisher and consumer?
People notice the DV Clouds because the creative is filling real estate on the page that looks like an ad position – while obviously not ad creative. The DV Clouds is an unbranded, non-promotional CSS overlay. If it were an ad (such as when we passback the impression to the publisher), users would never realise that an earlier ad had been blocked. There are several critical elements of the DV Clouds execution that are designed to optimise the publisher and user experience when ads are blocked and a passback to the publisher can’t be executed.
For example, certain competitors of ours often choose to display their own ad over a blocked impression (feel free to contact us for specific cases). Most people will see this and assume they purchased this inventory from the publisher as an advertiser. With this approach, competitors are basically co-opting an impression and a publisher’s inventory for their own brand. Their ad creative is competing for the consumer’s attention, when that attention should rightly be directed exclusively to the publisher’s content and any ads that were successfully rendered to the screen. DV doesn’t believe this is an appropriate use of a publisher’s inventory and we have made a deliberate choice not to do it. We consider it a conflict of interest to block an unsuitable ad with DV promotional content that competes for consumer attention.
Second, DV’s Cloud creative has been technically implemented to ensure the fastest possible privacy-friendly response that doesn’t impact the user experience. The DV cloud implementation doesn't require a call to another third-party server in order to execute the brand’s block. Instead, DV Clouds are rendered using a CSS overlay that immediately covers any sized ad slot, with no latency and no additional outbound calls from the device. This technically superior solution supports the end user experience and benefits the publisher.
In certain limited instances, ad platforms require DV to fill a blocked ad slot with a clickable creative. In those cases a CSS overlay can’t be used and we employ an alternate, non-branded creative – that clicks through to an informational landing page that briefly explains the use case, with little overt DV branding. This technique also supports our stance that publisher inventory, when blocked, should never show creative that either offers a service or competes directly with both publisher and paying advertisers for the attention of the consumer.
At DV, we are proud of the approach we’ve taken to media authentication. Blocking is a critical component of our service that protects brand equity and the integrity of marketing metrics, and ensures compliance. The DV Clouds support the digital ad ecosystem at large – advertisers, publishers and consumers – and we will continue to lead our category with the technology and services that support our singular mission – “Let’s build a better industry."