Paid Media
Updates by Casey Anderson, Paid Media Director
ChatGPT Ads Get Visual (and More Measurable)
For a lot of advertisers, ChatGPT ads have felt like something to watch rather than test. The audience is there, but the formats have been limited and the data hasn’t given us much to go on aside from top-line metrics. OpenAI’s latest update appears to be aimed at improving both of those problems.
OpenAI is expanding ad types in ChatGPT, and one of the first new formats shows visual ads while users are generating images. Until now, ChatGPT ads have only appeared below responses, so this is a new placement worth keeping an eye on. For now it’s a limited U.S. test, so don’t expect to see it in your account right away.
Other things to look out for:
Better conversion data. Hightouch, Tealium, and LiveRamp integrations can send conversion data from advertisers’ existing systems to ChatGPT Ads, so you can work with the tools you already use instead of building something new.
Incrementality testing is on the way. Haus, Measured, and WorkMagic are working with OpenAI on geo-based experiments to measure whether ChatGPT ads drive additional business results. This is the piece we’re most excited about, since it gets at whether those conversions would have happened anyway.
With ChatGPT reaching 1.2 billion people every week, it’s hard to ignore as a channel. If you’re thinking about testing it, we’d recommend starting with a modest budget and making sure conversion tracking is set up before launch. We’ll stay hopeful that these measurement updates give advertisers the clarity they need to decide whether ChatGPT ads earn a spot in the mix.
Google Ads: More Previews, More Testing, More Automation
It’s been a busy month in Google Ads. Most of the updates come down to two things: Google keeps automating more of the account, and it’s (slowly) giving advertisers more ways to check that automation before it goes live.
Preview AI Max copy before you commit. Giving up control on ad creative has always been one of the hesitations of going full-throttle with AI Max. Google Ads now lets advertisers preview up to 10 AI-generated headlines and descriptions before enabling AI Max text customization. The preview also reflects text guidelines, including excluded terms and themes, so you can check whether your brand rules are actually working. However, keep in mind that these are samples, so it’s still worth reviewing your live assets after you opt in.
Automated promotions are on by default. Starting Oct. 12, eligible advertisers could have website coupons and deals automatically turned into Google Ads assets unless they opt out. This applies to Search and Performance Max campaigns that have linked location assets and don’t already have promotion assets attached. Our advice: check your site for any old deals still hanging around. If you’d rather keep control, you can turn off Automated Promotions in your account-level automated asset settings.
A/B testing gets easier in Performance Max. Creative testing in PMax has always been tricky. An A/B testing toggle has appeared inside a live asset group, suggesting Google is testing a more direct route into its existing experiment system. Rather than setting up a separate experiment, you could test creative changes right where you edit your assets. It may not be in every account yet, but we’re excited to put it to use.
Automation is still moving fast, but the new previews and testing options give advertisers more room to stay in the driver’s seat. We’ll stay hopeful that Google keeps building in these kinds of checks as it rolls out more AI-powered features.
SEO & AI
Updates by Brett Woodward, Sr. SEO
Google Is Making Life Hard For 3rd Party SERP Trackers
In the past few months, multiple reports throughout the SEO community have been documenting new practices from Google that appear to be aimed at making it harder for 3rd party tools to get an accurate picture of search results.
On numerous occasions, SEO Brodie Clark has been helping the community keep tabs on these developments, showing us that Google has been:
- removing the “num=100” parameter functionality that allowed trackers to view 100 results in one look,
- replacing search links to encrypted “goto” URLs that do not provide scrapers with the final page destination,
- and according to SEO tracker tool Nozzel even potentially “poisoning” trackers by feeding them completely false search results that are different from what they show to users.
Why is Google doing this? One reason is because rank trackers can saturate search data with tons of queries that no human actually searched in order to pull rankings. But the biggest reason is likely because AI tools are scraping Google’s index to provide better responses to their users, and Google wants no part in helping competitors like OpenAI take a bigger slice of the search pie.
This is a constantly evolving situation, and my advice here is to continue to take 3rd party data of search engine results pages with a grain of salt. By all means keep using those tools, we sure will, but remember that Google Search Console is the only first-party source data we have despite its own shortcomings. Other tools are great at providing directional information but don’t take their data as gospel, since Google is actively trying to make that data less accurate in order to retain their dominance.
Publish Written & Video Content Together For Better AI Citations
A new study from OtterlyAI found that their content received more citations from LLMs when they published both blog and YouTube video format content simultaneously.
By tracking 8 stories across 7 major AI search platforms, the experiment revealed that written blog articles dominate early AI citations, earning nearly 80% of overall format citations within the first seven days compared to just 20% for videos.
The degree to which different AIs favored citing videos over written content fluctuated based on the model. However, the study noted that video citations ramped up over time, climbing from 15% to 26% of daily citations across the week. That’s a substantial increase in a short period of time, and it seems like the two content formats help boost the performance of the other.
OtterlyAI goes on to recommend treating video and written content as a paired unit rather than competing formats, and they even provide some helpful tactics you can use in your own content strategies near the bottom of the page.
This advice fits nicely into the same message we’re giving clients: the best content strategy is multimodal. Creating content in multiple formats will expand the surface area your content can cover on search engines and AI responses. It will make your marketing more durable since search tools will have multiple options for showcasing your content, and the study indicated, one format can help boost the visibility of another.
Analytics
Read our November Analytics Roundup for this month’s marketing analytics update. This month, a Google Tag/Google Tag Manager overhaul, how leading analysts are feeling about Claude’s new dashboarding functionality, Google’s new Data Agent Kit to connect Google Data Cloud to any agent, and more great news and articles from around the web.
