How do you do digital marketing without third-party cookies? How will it impact your marketing and your marketing technology? Despite several years of warnings, half of marketing professionals don’t think their organizations are ready for deprecation, according to a study by Basis Technologies.
Sentry.io got rid of all advertising and tracking cookies last July. Elaine Szu, the company’s Vice President of Global Marketing, says that even with more than a year of planning, there were still things that surprised them.
“You just don’t even necessarily realize how deeply entrenched an advertising cookie can be in the ecosystem and in all the tooling that you use,” she said.
Sentry provides SaaS error and performance monitoring to developers. The deprecation of cookies wasn’t the only thing behind making this change. Szu and her team were reviewing how the company connects with its customers and were concerned about over-relying on Google for acquisition.
Dig deeper: Web analytics is badly broken
“Most sane rational marketers would say, ‘Well, we should delay [getting rid of cookies] as long as possible,’” she said. “We took a slightly different approach. We’ve been thinking about it since 2022. That was largely driven by conversations about our overall customer acquisition strategy, the different sources we look at to get in front of our audience, and the customer journey overall.”
While there was a lot of concern about the impact of this on customer acquisition, there was also a lot of concern about how customers would react if they didn’t get rid of cookies. Sentry’s customers are developers and, as Szu put it, “Developers by definition are very, very, allergic to being tracked.”
Also, developers are hyper-sensitive to anything that even smacks of standard marketing. This has always been at the core of Sentry’s efforts to connect with them.
The software Sentry sells is based on an open-source program and is itself open-source. Developers can modify and use it as long as they agree not to compete with the company for two years after purchase. As it says on the company website, “Sentry is an open-source company because the right to learn and to share what is learned with others is fundamental to product growth and relevance.”
The community of open-source developers will push back hard if their values aren’t respected. However, if they see a company as authentic and in synch with those values they are incredibly loyal and happily tell others.
“We’ve been very conscious and thoughtful about the fact that we don’t push monetization or marketing onto that open-source community,” said Szu. “We’ve really been investing in developer content, technical tutorials, building this more of an educational approach to marketing and letting the developer then choose when they’re ready to speak to us and talk to us.”
In short, their customer base is people who were likely to notice and react positively to Sentry removing cookies from its site.
The Sentry team wasn’t only worried about customer acquisition without cookies. There was also the issue of the impact on their technology and data collection.
Matt Henderson, the company’s global marketing lead, put it this way in a blog post: “As a growth marketer, I was excited at the prospect of it, but then I started thinking about all of the things that are going to break. The initial response I had was to raise concerns for why we shouldn’t do it or to delay it.”
Here are the breakpoints that initially worried him:
Attribution models (first and multi-touch in Sentry’s case).
Marketing reports in our business insights tool.
All SEO reporting (built on GA and GA4).
Google Ads smart bidding.
Procurement of onboarding new tools.
For Szu, this meant the team would have to go back to the marketing mindset people used when search marketing was still in its infancy.
“It’s almost like going back to a decade ago when it was less about pushing constantly and moving more towards a pull mechanism,” she said. “You have to create great brand investments that are consistent, that are resonant. More thoughtfulness in the way that you market and educate the world, and using creative ways to get in front of the audience versus trying to force them into your funnel.”
While the Sentry team had a pretty good idea of how the loss of cookies would affect their technology, they didn’t understand all the ways it would affect working with the rest of the digital world. For example, partner peer reviews were an important part of their marketing. Those reviews are driven by using a cookie, so they can’t be used anymore.
“It’s not something we anticipated because it’s not something you’re not thinking about,” Szu said. “You’re thinking more about your directly owned and operated property, tracking attribution, retargeting capabilities. We were aware that retargeting our core traffic would need to be revamped. We knew we would have to focus on UTM parameters and rebuilding all of our business intelligence to build UTM parameters. That was all anticipated. It’s these third parties and other areas that we didn’t really think through.”
Dig deeper: 5 GA4 issues and why they’re a good thing
Unanticipated problems with just hosting videos caused them a last-minute problem and delayed the relaunch of the Sentry’s website by two weeks. Sentry had been posting recordings of its workshops to YouTube and then embedding them on their website. However, doing that means YouTube embeds a cookie on the site. Once the team understood that, they had to upload all their videos to Vimeo and then redirected all of the website-hosted videos to Vimeo.
Here is Matt Henderson’s list of the other things they hadn’t realized would cause them problems once they removed all tracks on their website. “Some of these were more due to extreme measures we took by removing user tracking and the cookie banner as a whole, but most were due to the removal of cookies,” he writes.
Our GCLID from Google Ads passing onto our site.
Our bot-blocking tool for ads relying on both a pixel and GCLID.
Lookalike modeling not being an option b/c of data sharing with ad platforms.
Not being able to use Salesforce x Google Ads integration.
Completely losing signups from display marketing and YouTube.
Losing referrer data for content analysis.
4xing our website’s 404s and redirects by accident (whoops).
The last point helped the team improve the quality of the website by identifying old pages that hadn’t been updated and weren’t being used properly.
Getting rid of all the cookies requiring visitor consent also let Sentry get rid of a ubiquitous, often overlooked irritant: The consent banner.
“That in and of itself is sort of the first disruption that every user has to every website right now,” said Szu. “[Visitors] are like, ‘I don’t even want to have to think about this. I was just trying to look at this one product.’ So, we were able to reduce the friction in the user experience — that’s part of our core values.”
The Sentry team knew that the lack of signals with Google would have a severe impact on their traffic. They decided to approach it as an opportunity to learn other ways to find their audience.
“Display and other channels were severely hampered after cookie removal,” writes Henderson. “Our traditional retargeting motion died off pretty quickly as we couldn’t use GA4’s audiences or our Google Ads pixel.”
They decided to rely on ad engagement retargeting — rather than traditional retargeting — on most of their ad channels. That isn’t the same thing, of course, but it gave them a semblance of a funnel. So they tailored ads to focus on the middle and bottom of the funnel to their existing audience.
At the same time, they realized prospecting for customers with display ads wasn’t working out for them. Szu and her team repurposed that budget for sponsorships and publishers they knew appealed to their core audience. “This may produce less flashy conversion counts depending on your business, but it gave us a place to tell our story to our audience,” writes Henderson.
Once the cookies were gone, CPC in Google Search increased by about 30%. To mitigate the drop in audience numbers they recommend using hashed pass-backs or offline signals/APIs and smart strategies like Target CPA bidding and Maximize Conversions.
For Google Search, they went back to using Enhanced CPC and Max Clicks which optimize for less important factors than they could with data from cookies. They now monitor negative keywords and regional data more closely and use it as a proxy for a bidding model to sort through user intent.
For YouTube, they used Target CPV for any video viewer or engagement and Target CPM for awareness. They couldn’t use video action campaigns anymore so they focused more on reach. The results have been impressive: “We ended up 13xing our views on our videos period over period, and are getting more quality site visitors than we used to,” writes Henderson. The benchmarks for view rate are at industry average and above the benchmark from Google. so this ended up being a successful shift. They also had 100% growth in people mentioning they heard us from YouTube in their onboarding survey quarter over quarter.
For Display, they shifted to viewable CPM, this cut the cost nearly in half and produced a large increase in impressions (+20%) month over month. However, conversions tailed off and then dropped off of a cliff with the shift in bidding models.
The specifics of Sentry’s case won’t apply to every organization. How many places have a customer base that will appreciate or even notice the absence of cookies? Szu and Henderson know that. By sharing their journey, they hope people will understand that going cookieless is very doable and can result in better, more creative marketing.
Get MarTech! Daily. Free. In your inbox.
US singer-songwriter Taylor Swift performs during her Eras Tour in Inglewood, California on August 7, 2023 – Copyright GETTY IMAGES NORTH AMERICA/AFP Patrick Smith Alex PIGMAN Fans of Taylor Swift and politicians, including the White House, expressed outrage on Friday at AI-generated fake porn images of the megastar that went viral on X and were...
Google’s DeepMind published a research paper that proposes a way to train large language models so that they provide more reliable answers and are resistant against reward hacking, a step in the development of more adaptable and efficient AI systems. Hat tip to @EthanLazuk for tweeting about a new research paper from Google DeepMind. AI Has...