Kellogg’s Assesses Video Viewability and Sales Numbers


Kellogg’s has joined forces with Nielsen Catalina Solutions (NCS), Moat, and Yahoo’s BrightRoll in order to evaluate the relationship between the quality of video advertising and purchase rates. They measured factors such as ‘time in view’ and ‘audibility’ in order to determine how they influenced overall sales impact.

Kellogg’s purchased sets of exposures from BrightRoll and Moat provided a variety of metrics to provide insight into the video viewings. Some of these metrics include data on video player size, time spent viewing, and if the audio was muted during viewing. This data was leveraged against offline sales data provided by Nielsen Catalina Solutions in order to assess the sales outcomes of the video inventory.

Does Video Inherently Boost Sales?

The results were surprising to many in the industry and challenged the notion that video automatically improves sales numbers.

“More stringent in-view minimums, customized to the format of delivery, are required,” said the senior director of KNA Media and Experience Planning of Kellogg Co., Chris Osner-Hackett.

Evaluation from the test campaign revealed that videos that were viewed for seven to 16 seconds were the most significant contribution to sales. This points towards a flaw within the industry standard of two seconds.

It was also discovered that video with the sound active boosted sales by twice as much compared to videos with no sound or the sound off.

Nielsen Catalina Solutions was also able to pick up on some nuances of reach and sales frequency.

“Reach is particularly important, but not just reach among all people – it’s reach among opportunities to purchase,” said the chief revenue officer for NCS, Andrew Feigenson. “I may buy bottled water once a week, you may buy it once a year. So my ability to influence a purchase is largely dependent on how many times you’re going to a store and when I capture you in the purchase cycle.”


There Is Still No Black And White Definition

Despite many of the insights, there is still no clear definition of what works the best, particularly when it comes to time spent.

For example, if a consumer purchases a year’s supply of Corn Flakes, and Kellogg’s exposes them to a video ad afterwards, the consumer is less likely to be impacted by that impression.
Determining the connection between video views and sales numbers is starting to define more accurate measurement methods. In the past, studies that evaluated the connection between digital activity and offline sales were slow and expensive. However, as more offline data becomes digital, the ability to measure ad effectiveness has improved greatly.

This will enable advertisers to develop and run campaigns that are more directly related to performance and measurement data. Although there are still many challenges to this type of targeted advertising, it is becoming much easier as measurement methods develop. As programmatic transactions continue to rise and access to data grows, it is predicted that these types of integrations will become industry standard within the near future.

Google Eliminates Its ‘Last-Look’ Advantage in Ad Auctions


Finally, Google is leveling the playing field by getting rid of their ‘last-look’ advantage in ad auctions. This has been a practice that has always bothered publishers and exchanges because their bids could always be discarded in favor of Google’s preference.

Many people saw this as an unfair advantage, but as of the beginning of April, Google has relinquished their priority position. This was released in a document that detailed the transition to Google’s Exchange Bidding in Dynamic Allocation feature, which is still being tested to some degree.

Jonathon Bellack, the director of product management at Google, confirmed this recently.


“We are collecting the price each exchange would pay, including AdX, and then putting it in a unified auction where the highest price wins,” Bellack commented.

Here is a quick breakdown of how the programmatic auction will operate now: Any EBDA exchange participant, such as Index Exchange, Rubicon Project, Pubmatic, Sovrn, Smaato, or Gamut, simply submit their bids for an ad placement. The DoubleClick AdExchange will also submit their bid, but now, without any preferential treatment, the highest bid wins.

In the past, the DoubleClick AdExchange would allow every other exchange to submit their bid, and then it would take a final moment to possibly outbid the other highest bidder. However, now Google’s exchange enters the auction at the same place as other exchanges and will need to submit a bid that is competitive if they want to secure the placement.

Although there may be an initial impression that this new change will hurt publishers since there is a lost opportunity for a higher bid from Google, the outcomes will essentially be the same due to the rules of second-price auctions. This will simply open up the highest bid to more exchanges.

Eliminating the ‘last-look’ advantage is a sign that Google is changing how they work their partners.

“This is a big change in how exchange bidding works that shows they are open to having it and DFP be a lot more neutral,” said Drew Bradstock, the SVP of product at Index Exchange. “We like that they are willing to listen and be a lot more transparent.”

At a Google forum last year, one of the primary concerns raised by publishers and exchanges was based on the ‘last-look’ advantage that Google held. According to Bellack, the criticisms raised in this forum were a major influence on the new changes.

“The exchanges and publishers we’ve been working with like this [change], because they think this is a fair way to make competition,” Bellack said.

However, Google neglected to say how this change will affect its own win rate and revenue. Regardless, the DoubleClick AdExchange performs well without the ‘last-look’ advantage. Their performance includes unique demand from Adsense and they are still holding onto one advantage over their competitors. Google knows more about users than they share with other exchanges.

Despite this, the change is a definite improvement for other exchanges, and as Bellack commented, “We only succeed when our partners succeed.”

Ad Tech Needs a Human Touch to Solve Brand-Safety Problems


Malcolm Cox, CMO of Grapeshot, spoke recently about brand-safety issues and programmatic ads at a meeting sponsored by ad-tech company, PulsePoint.

“In the era of big data, there are more answers than ever, but people aren’t asking the right questions,” said Cox as he probed into some of the recent problems arising with brand-safety.

In light of some of the recent events with Google serving ads for some big brands in contextually inappropriate locations, Cox suggested there might be some better ways to protect brand-safety. He emphasized that the industry needs to be more transparent about their policies, methods of decision-making, and how ads are served.

“There’s no reason why we, as an industry, can’t help advertisers. Google is the No. 1 brand in this country and the most trusted brand, yet we don’t feel we can trust Google to place ads next to site content,” Cox said to the audience of mostly ad tech professionals.

Part of the issue is that ad tech alone is not sufficient enough to guarantee brand-safety. Rich Sutton, CRO of Trusted Media Brands, addressed this directly.


“If it doesn’t have human intervention, you can’t guarantee it will run with brand-safe content. I don’t think you can stop [it] with tech alone. I don’t think that advertisers realize that — and when they do, they’ll either pull back in a big way from programmatic and digital overall, or do business with premium publishers directly,” said Sutton.

Sutton also cited research that pointed towards more than 50% of advertising agencies believing that programmatic would not be able to do a better job of ensuring brand-safety. Tech on its own seems to be unable to determine some of the nuances that make a big difference to brand image and satisfactory ad placements.

This was proven beyond a doubt when some marketers’ ads were displayed beside offensive content on YouTube.

“There’s enough blame to go all around, but in order to move forward we have to start collaborating. How do you get through this finger-pointing moment and get to collaboration? It’s really easy to blame Google. The problem is that the technology was built to follow audience interest,” said Rob Rasko, CEO of The 614 Group.

Rasko argued that about 90% of the tech works, but there is still 10%, which can cause some major issues for big brands. This leaves the question of the best method for fixing that last tenth of ad serving.

“I argue if you’ve got data and real-time decisioning in a brand-safe environment, dollar for dollar, you will outperform other channels because of the scale, data, and real-time information,” said Chris Neuner, SVP and GM of digital health solutions at PulsePoint.

What is clear is that ad tech on its own will not be able to guarantee brand-safety until new processes are put in place. This means the only true course of action to reduce brand-safety risks requires a marriage between ad tech and human intervention.

Nielsen Catalina and Facebook Team up to Deliver Cross-media Marketing Solution


In the past, TV and Digital have been considered competitive platforms for advertising, but this may all change soon with a new plan from Facebook and Nielsen Catalina Solutions. The two major companies are pairing up to work on a new cross-media measurement product that will evaluate how the combination of digital and TV can produce better sales for CPG advertisers.

“The biggest ask from all advertisers has been to help them with comparability,” says the director of marketing science at Facebook, Fred Leach. “They know Facebook works, but they want to understand how it works in comparison with all of the other activities they’re doing.”

Nielsen Catalina Solutions are experts at evaluating cross-media sales effects. They are able to draw upon their impressive 100,000-person consumer shopping panel in combination with in-store purchase data from approximately 90 million homes in the United States in order to gain a glimpse into the benefits of using multiple media platforms for advertising.


From their initial study, Facebook and Nielsen Catalina Solutions has found that in the majority of cases, utilizing both TV and Facebook has resulted in the highest return on ad spending. That being said, there were some campaigns that used a combination of TV and Facebook, but it only resulted in a flat ROI.

However, this may not be a failure of the combination of platforms, but rather a lack of quality in the creative or the messaging, suggested Leslie Wood, the Chief Research Officer at Nielsen Catalina.

It goes without saying that the frequency of views has some effect on the incremental sales lift, however it is not the major determining factor for return on ad spending. ROAS was at the highest when both TV and digital ads worked in conjunction, regardless of the actual amount of times a consumer viewed the ad.

“We see strong synergy when you bring two media together,” says Wood.

At this time, the studies only covered the news feed ads, but there is hope to expand the research to include the cross-media effect of Instagram and the Audience Network.

However, this may not be as simple as some would like. It is already quite complicated to gauge exposed and unexposed viewers within the context of just one publisher.

“There is still research and development that needs to be done, particularly on the NCS side, to do more sophisticated cross-publisher comparison,” says Leach.
Nielsen Catalina is testing its new product only with certain advertisers in an invite-only beta version. However, the plan is to eventually allow other publishers to use the cross-media solution to demonstrate their value within cross-media marketing.

Facebook has been partners with NCS since the beginning of September, when they also released that they would begin a series of third-party partnerships with other organizations like Oracles Data Cloud, Visual IQ, and Neustar MarketShare. This is all part of Facebook’s efforts to increase transparency within the industry in response to rising pressure from outside sources.

Facebook Announces Launch of Header Bidding with the Help of Tech Partners


Facebook is finally opening up its platform to better header bidding by expanding the capacity of their Audience Network. However, they will be relying entirely on partners to provide the tech needed to support the transition.

Facebook’s foray into better header bidding was announced this month after the success of the beta testing period of their Audience Network. Their header-bidding format acted as a demand source for publishers who were integrated with a variety of pre-selected technology partners such as Amazon Publisher Services, Index Exchange, Sonobi, Sortable and AppNexus.

In simpler terms, this means that Facebook does not have a header-bidding product that is readily available to all. Instead, publishers who work with one of Facebook’s pre-approved partners can opt-in for access to the Audience Network by employing their wrapper integration.


Access to the Audience Network is determined by code that is loaded on the page, which drives demand directly towards a publisher’s property. The exact process varies depending on the technology partner, but the result is essentially the same. All types of formats will be available to publishers, including native, video and standard formats. There is also potential for Facebook to start using server-side integration.

This approach is a little bit different from the one that Amazon has chosen to take. Amazon’s server-side header tag lets publishers manage multiple header partners within one wrapper, but Facebook saw no reason to make radical changes to a system that works.

“We’re not coming out and trying to compete in the ad tech landscape or releasing third-party technology,” says David Jakubowski, the director of publisher solutions at Facebook. In fact, he mentioned that Facebook was content waiting on the sidelines until the industry reached its maturity.

Facebook took its first steps into header bidding this past August with the initiation of its beta testing in conjunction with a short list of publishers like Forbes, The Washington Post, and Daily Mail.

However, now is the first time that Facebook will allow its Audience Network to appear anywhere, rather than only through direct implementation. According to Facebook, this is the primetime to start making these sorts of changes.

“All the demand sources need to get the same information at the same time,” says Jakubowski. “And the ad space should go to the person who needs it most without any weird black-box arbitrage or bias to any owned and operated demand source.”
Facebook’s break in to header bidding is heavily concentrated on building better relationships with their publisher partners instead of just creating open ecosystem play.

“This is a departure from ad tech of the past,” says Jakubowski. “It’s a simplification and cleaning up of the technology process so that everyone knows where their dollars are going.”

Although Facebook will not divulge their exact match rate for publishers accessing Audience Network demand, they did claim that publishers who use their header bidding integration see an increase of revenue somewhere between 10% and 30%.

Google Plans Stronger Safeguards After Ads Appear Beside Extremist Content


For obvious reasons, big brands don’t want their advertisements appearing close to controversial content. However, this recently happened to a few large companies like McDonald’s and Mercedes-Benz when their ads were featured during extremist content on channels such as YouTube. Now, Google is feeling the lash back and has released plans to create stronger safeguards for ads in response to the criticisms.

Many big brands, agencies, media outlets and government bodies in the UK are considering restricting ad spending until better guidelines can be put in place to protect brand images.

This has Google scrambling to fix the issue and they have already expanded safeguard policies for advertisers, which will provide them with better control over the placement and timing of ads.

Philip Schindler, the Chief Business Officer at Google, apologized recently in a blog post. “We had a number of cases where brands’ ads appeared on content that was not aligned with their values. For this, we deeply apologize.”


In addition to improvements to their ad safeguards, Google is going to take a tougher position on content that could be considered offensive or contain hateful messages. It will also ensure that advertisements only appear on channels from legitimate creators that are part of its YouTube Partner Program.

Brand safety has always been a major concern for all advertisers, long before programmatic and other audience-based purchasing platforms were prevalent in the industry. That being said, in the digital world, it is becoming even more difficult to keep a handle on all of the moving components for programmatic advertising. It is anticipated that Google’s changes to their ad safeguards will be a step in the right direction for eliminating harmful ad placements that could potentially hurt a brand’s image.

However, in the wild west of the Internet, it is difficult for anything to ever be fully guaranteed.

With brand safety and image protection becoming an increasingly global problem, some agencies are taking further steps to protect their client’s image. As large companies desire to advertise through many different channels in many different countries, it is becoming more and more difficult to know exactly where those programmatic ads will end up. This has inspired some agencies, such as Wave, to integrate more safeguards into their programmatic advertising framework in order to prevent these types of situations. It is not always possible to simply rely on Google to protect any company’s image other than their own appearance.

This is even more important when it comes to advertisements that are placed within user-generated content. It can be difficult to determine the context of some of this content with 100% accuracy, which puts some programmatic ads at risk of poor placement, especially if they are not working with an agency that knows how to protect their ideals.

These recent events are a further reminder that programmatic buying is not a replacement for human beings. Constant monitoring and reporting should allow trading teams to catch potential risks early and correct them before brand partners become aware and too much spend is allocated. We are big proponents that programmatic is not a strategy, it is a tool. And with all tools in the wrong hands the outcome can be catastrophic.