Google introducing machine learning to manage ad frequency when cookies are missing
Last week, Google announced its new privacy-safe approach to managing ad frequency. After a number of user consultations, it was deemed that one way that Google could take a more privacy-centric approach to publishing and advertising was in the management of ad frequency – or the number of times someone is served an ad across devices.
At present, if a user has third-party cookies blocked or restricted, there is no way for an advertiser to know or control how many times it has served an ad to a particular individual, which can lead to spammy targeting, and ultimately lots of unnecessary wasted media spend.
In response to this, Google will soon be rolling out a new feature in DV360 that uses machine learning to help advertisers manage ad frequency. This will be done by creating models that predict traffic patterns when cookies are not present across Google Ad Manager. This can be used to estimate how likely it is that users will visit different publishers serving ads through Google Ad Manager.
Given all user data is aggregated before Google applies its machine learning models, no user-level information is shared. The feature relies purely on a publisher’s first-party data to inform the ad frequency on the site.
We are really pleased to see this new model being implemented, as user privacy is an ever-important topic, especially in a post-GDPR landscape. Capping ad frequency is crucial to a successful paid media campaign, to avoid wastage and to optimize saliency without reaching the point of being spammy for the customer and draining unnecessary budget. While all of this is great, we are still keen to ensure that ad frequency capping is done manually wherever it can be to deliver the best possible results for our clients.
YouTube “Video Reach” campaigns introduced for multiple creative rotated within one ad placement
YouTube has now introduced its new product Video Reach Campaigns, which allows advertisers to upload multiple video creatives in the same campaign, and from there Google’s machine learning will automatically serve the most efficient format to reach the target audience at scale. This new campaign type will be available across multiple formats, including 6 second bumper ads, skippable in-stream ads and non-skippable in-stream ads.
Car brand Ford has been an early adopter of the new format and has seen a 20% reduction in campaign costs compared to previous YouTube benchmarks.
This will be an interesting development for YouTube advertisers with multiple creatives, looking to maximise audience reach without massively increasing frequency.
YouTube testing ad buying tool that lets you reserve ad placements 120 days in advance
Google is currently testing a new Instant Review Tool for YouTube, which would enable advertisers to reserve inventory up to 120 days prior to a campaign going live. This new feature could be particularly helpful for advertisers looking to run promotions around certain events, or in line with seasonal holidays.
While there are a lot of advantages to this, should it roll out permanently, we could also see that it establishes an early play for inventory and higher prices to pay to reach key audiences at peak times when booking closer to a campaign going live.
Conversational e-commerce (c-commerce) - the latest insights from Facebook
The latest stats from Facebook Insights has thrown up some interesting insights around conversation in e-commerce:
- Last Christmas, 42% of those surveyed in Europe messaged a business over the holiday season – for a variety of reasons:
- Seeking product information (45%)
- Personalized advice (31%)
- Negotiating prices (30%)
- Instant responses at any time of day (35%)
While advertising via messenger still has lower adoption rates than other more traditional ad formats, it’s something we have seen gain traction over the last year, and it is certainly a product that Facebook has been pushing to increase adoption.
It’s still early days in terms of measurement and understanding the full impact that messenger bots and other conversational tools can have for e-commerce advertisers, but it will be a trend that the team will be following with interest in the coming month for all of our lovely e-commerce clients.