The following is the second piece of a four-part series by Matt McLaughlin, DoubleVerify COO, on the issue of brand suitability to explain how suitability technology best practices during the coronavirus crisis can help advertisers, technology partners and publishers.
Amid the coronavirus crisis, confusion over keyword-based blocking has intensified. This confusion often results from not differentiating how various providers implement their keyword blocklist services. Advertisers want to avoid unnecessarily taking good supply off the table, while publishers rightly have concerns that imprecise keyword blocking could demonetize suitable news content. We have offered very specific guidance regarding advertising on reputable news publishers amid the coronavirus outbreak. With the right partner in place, however, keyword blocking is a key feature of the brand suitability toolkit and over-blocking through use of keywords is generally rare.
According to our own data and based on campaigns from earlier this year across many of the world’s top advertisers, across billions of ad requests globally, the incremental rate of brand suitability blocking due to keywords sits at well under one percent. Because of the diversity and types of content produced, news publishers experience a higher rate, but still under two percent. It is important to note that this is the incremental rate of avoidance -- avoidance due exclusively to keyword blocking and not for any other brand suitability settings. If a brand is avoiding content classified in the category “politics” and is also avoiding the keyword “Trump”, the incremental rate of avoidance due to this keyword is limited only to web pages where both the word “Trump” is in the URL and the content is unrelated to politics. This limited impact is because DV’s keyword technology is smarter and more precise than ever before.
How are Keyword Blocklists Used?
Keyword blocklists prevent ads from appearing next to content that may be unsuitable, based on the appearance of specific words in the URL of an article. Historically, the URL has been an important SEO tool and, as a result, is an effective proxy for an article’s title.
There are three primary use cases for brands to implement keyword blocklists to maintain suitability:
- Avoiding emerging negative content. Incidents like a terrorist attack, mass shooting or other emerging stories can create high volumes of web pages in a short period of time. DoubleVerify’s semantic science solution is fast - retrieving and classifying this content and distributing the categorization to our global serving infrastructure in minutes. In these critical minutes brands wish to be protected from appearing alongside this unsuitable content.
- Increasing avoidance precision on specific topics. Sometimes brand suitability concerns may be more narrow than a particular topic. Suppose a brand has a negative association with a political topic like minimum wage. In this case it is more precise if the brand blocks on minimum wage-related keywords as opposed to blocking all content in the politics category.
- Blocking narrow, brand-specific concerns. Some brand suitability concerns are truly specific to an individual brand. These concerns could include product recalls, legal judgements, or workplace accidents. Utilizing keyword blocking is particularly effective in these cases where there is a narrow, brand-specific concern.
It is important to note that the last two use cases actually increase publisher monetization by enabling brands to implement more precise and granular controls. By allowing narrow suitability concerns to be addressed with keywords, it enables brands to continue to buy content in broader categories like politics and news.
I want to add a quick note here about keyword list hygiene. DV encourages our customers to regularly review their keyword list to ensure it meets one of the three criteria below. This is especially important for keywords added for the first reason listed (avoiding emerging negative content). Once an incident is no longer emerging, a best practice is to remove these keywords from the blocklist.
At DV, we recognize that the coronavirus challenge is a unique situation that will not simply “go away.” We also recognize that trusted news is a value add that advertisers should embrace, unless there is a direct connection between a news incident and their brand. For that reason, we have encouraged brands to exempt trusted news sites from keyword blocklists.
DV Keyword Blocklist Confusion
With that said, we understand that there might still be some confusion over keyword blocklists and how they work - especially because there are assumptions being made in social media about “keyword blocking” that don’t distinguish DV’s far more precise and advanced solution from services offered by our competitors. For that reason, we would like to address the most common misunderstandings about DV’s keyword blocking and ask again that you note that not all media authentication providers execute keyword blocking the same way:
Misunderstanding: A single keyword anywhere within the content of the page can cause a block.
In fact, DV’s keyword service scans the URL, which serves as a proxy for the content title, and is not comparing keyword blocklists with the content on the page. This ensures DV applies more sophisticated semantic science categorization techniques to the page content and limits false positives that may occur if the entire page content is scanned for individual keywords.
Misunderstanding: Keyword blocks happen on news homepages or news section homepages.
In fact, because DV’s keyword service only scans the URL, keyword blocks never happen on these pages because they have no URL path to scan (unless a brand mistakenly entered a very generic term like “sports,” which DV advises against).
Misunderstanding: Words like “Sussex” can trigger a block when a brand is trying to avoid the word “sex.”
In fact, DV’s keyword service allows advertisers to set up exact match keywords where the word must stand alone in order to be blocked. If a brand enables “ sex ” as an exact match keyword in DV’s service, then it will not block when the word “Sussex” appears in the URL.
Misunderstanding: Keywords cannot be avoided using pre-bid tools, resulting in blocks on a publisher’s site after the ad is purchased.
In fact, DV’s Authentic Brand Safety targeting solution allows the exact same keyword settings to apply to programmatic auction evaluation as post-bid blocking. This solution is available in nearly all leading DSPs including Adelphic, Amazon, Amobee, MediaMath, The Trade Desk, Verizon Media DSP, and Xandr. However, this solution is only as effective as the information available in the auction. If a publisher truncates the URL data prior to initiating a programmatic request, which DV observes on over 35% of US top-100 news auctions, then pre-bid keyword avoidance doesn’t have the auction data necessary to avoid this content during the programmatic bidding process.
Misunderstanding: Large keyword lists are “blunt” instruments that harm news publishers.
In fact, when DV applies keyword lists, less than 2% of impressions on the top-100 US news sites are blocked solely due to this reason. Use of keyword blocking, even in large lists, enables brands to monetize over 98% of news content. The real “blunt” tool is news avoidance, where brands take the perceived “safest” path and avoid an entire category causing significant harm to news publishers. DV encourages brands to advertise across trusted news sites as broadly as possible while implementing tools like keyword blocklists that can be used to precisely manage their legitimate suitability concerns.
Misunderstanding: Keywords with different meanings in different languages (think “die” in English versus German) can have unintended consequences for global brands.
In fact, DV’s keyword technology allows brands to create keyword lists per language to eliminate the ambiguity of different meanings in different languages. DV first determines the language of the page and then applies the appropriate language-specific keyword list to the blocking or pre-bid avoidance evaluation.
Proper understanding and use of keywords is one of several tools, alongside semantic science content classification, inclusion/exclusion lists, and other settings that enables brands to have the clarity and confidence they need in their digital ad campaigns. DV has built a keyword blocking service that is differentiated from other service providers, and useful for brands to manage key suitability concerns, while ensuring maximum monetization for trusted news publishers. Like any setting in a media plan, keyword blocklists need to be executed properly, with the right partner, and reviewed periodically to ensure they are effectively serving their intended purpose.
Read Part 3 in the series here.