The Post-ATT World: What DTC Brands Need to Know

Apple's App Tracking Transparency (ATT) has undoubtedly impacted the DTC and eCommerce marketing landscape. But in the midst of these changes, forward-thinking startups are harnessing AI and generative technologies to reshape performance marketing. As marketers, it's crucial to adapt to these innovative tools, optimize campaigns, and stay competitive in this privacy-centric environment.

To explore the effects of ATT and the revolution brought on by AI-driven solutions in marketing., Treat’s CEO and Co-Founder, Matt Osman, recently joined Eric Seufert (Analyst and GP, Hercales) and Rishabh Jain (CEO, Fermat) on Greylock’s Greymatter Podcast, hosted by Mike Duboe

In this post, we’ll highlight key moments from the conversation focused on how generative AI tools are helping DTC brands transition their performance marketing in 2023.

You can listen to the full podcast here.

The Dawn of the Probabilistic Era:

We can view the Post-ATT world as the result of a transition from a deterministic state, where marketers had come to rely on repeatable and efficient results from ad networks, to a more probabilistic one, where attribution has become increasingly unreliable and ad networks have shifted to machine learning models to target audiences.

There’s been this kind of game of whack-a-mole leading up to this, we’ve moved to a probabilistic world where honestly, I think the ad networks have made the call that the only people who can really measure correctly or solve attribution or even attempt to are those with very, very large machine learning staffs who understand this space exceptionally well.

Matt Osman, CEO @ Treat

The Deterministic Era

In the deterministic era, platforms like Facebook (now Meta) allowed marketers to achieve near-perfect mapping of consumer data by utilizing post-back links. This approach facilitated a convergence of data on the brand side and performance on the ad network side, creating a highly efficient advertising engine.

The Probabilistic Shift

With the implementation of ATT and other privacy-focused measures, the deterministic advertising approach has been disrupted. Today, we find ourselves in a probabilistic world where ad networks must rely on large machine learning teams to tackle attribution challenges.

The big thing that all of adtech recognized at that moment – and it was sort of a strange gentleman’s agreement that if we admit that there’s no solution to this for the next six months, our stock price is going to tank. That’s just true. And I wish there was another sort of collegial way of saying this, but this is actually what was true.

Rishabh Jain, CEO @ Fermat Commerce

Enter: Machine Learning

In response to this shift, major players like Meta and Google are moving marketing activities in-house, leveraging tools such as Advantage Plus (Meta's rebranded automation suite) and PMAX (Google's programmatic offering). These moves are an attempt to regain control over the advertising landscape, driven by the need to address economic pressures and maintain high stock prices.

What This Means for CMOs and Growth Marketers:

Marketers must stay informed about the changing landscape and adapt their strategies accordingly. The shift to probabilistic advertising calls for a deeper understanding of data attribution and a greater reliance on machine learning capabilities.

Fading Attribution

So when Apple said, “Hey, you have to get a consumer to opt in,” by definition, a certain set of those consumers are opting out. And so you no longer have that signal. So you now need to like Matt has been saying, you need to move to probabilistic models.

Rishabh Jain, CEO @ Fermat Commerce

Small and medium-sized merchants, in particular, are grappling with two significant problems:

• Attribution becoming more difficult due to signal loss
• Scaling issues on individual channels like Facebook.

The way to fully prevent this behavior – which Apple calls tracking, which is kind of mixing the third-party and the first-party context, it’s commingling that data – is to just starve these (ad tech) companies of the data. There’s no way to sort of ask them to abide by this policy unless you remove the data from that sort of transactional frame where it will be used.

Eric Seufert, Analyst and GP @ Heracles Capital

With the shift to probabilistic models and a growing need for multi-channel marketing, merchants are turning to multiple tools for attribution. Brands are now triangulating between Google Analytics, third-party attribution tools (such as Rockerbox, Northbeam, or Triple Whale), and in-platform data to navigate these challenges.

I think there is a begrudging acceptance now (at least amongst the brands that we speak to) that there’s no one source of truth anymore, and now the kind of big topic is triangulation and using like a multitude of different techniques to try and get to the answer.

Matt Osman, CEO @ Treat

The complexity of attribution has increased, and there is a noticeable gap in the market for tools and education tailored to small and medium-sized businesses. Merchants are looking for more effective ways to measure the success of their marketing campaigns and better understand the best practices for attribution across multiple channels.

The need for improved tools remains, leaving room for innovation and growth in the performance marketing space.

Owned Data is Gold

Shopify, has ventured into traffic generation by developing their Audiences product. As the platform evolves into a marketplace-like model, it presents new opportunities and challenges for both Shopify and its merchants.

Shopify Audiences has shown early promising results, striking a balance between compliance with ATT and addressing merchants' needs. By leveraging first-party data, Shopify offers a high-margin solution without violating the privacy of brands on their platform.

The broader discussion here is the importance of brands better leveraging their first-party data and exploring novel approaches to improve their marketing strategies. As Shopify continues to grow and adapt, businesses should become more strategic in how they use the available data and tools to drive success in the ever-changing e-commerce landscape.

Data powered ad-networks:

As the advertising landscape evolves, retail media networks are becoming increasingly popular for businesses in search of targeted and direct response ads. This trend is driven by companies like Amazon and Walmart, who are leveraging their vast audience data to launch their own ad networks.

The technology required to create these networks is becoming more accessible, allowing even smaller businesses with sizable audience profiles to monetize their data and explore new advertising opportunities.

But I think that’s behind the dynamic of what I call the “Everything is an ad network phenomenon.” Everything is becoming an ad network because well, Facebook’s not the everything store now. And hey, that’s an opportunity for me to spin up pretty meaningful super high-margin revenue.

Eric Seufert, Analyst and GP @ Heracles Capital

The Shift Towards Creative and Generative AI

As third-party cookies and other tracking technologies face increasing scrutiny, the focus has shifted toward creative as the new frontier for reaching audiences.

The recent rise of generative AI is revolutionizing the marketing landscape, offering new ways to streamline the creative process and enhance ideation.

I think most can agree that image models are going to give designers superpowers. There’s a bunch of new apps being marketed to marketers, given it’s such an obvious use case with a quantifiable ROI of all this stuff.

Mike Duboe, Partner @ Greylock

Adapting to New Creative Tools:

As a variety of powerful creative tools emerge, the focus should shift towards developing better filters and upstream tools to determine what to create. The rapid advancements in AI technology offer promising opportunities for marketers to harness its potential in the creative process.

Generative AI in Incumbent Design Tools:

Major design tool providers like Adobe and Figma are incorporating generative AI into their offerings, making the production aspect of creative work easier and more efficient. Adobe's recently launched Firefly suite and Figma's upcoming Stable Diffusion integration are examples of this trend.

The Power of Generative AI in Ideation:

The true potential of generative AI lies in its ability to enhance the ideation phase, where the bottleneck in creating high-performing creative content often occurs.

And as the cost of content creation (both images and copy) kind of tends to zero, which is basically what we’re riding right now, the value of knowing what to create and why increases proportionally in the opposite direction. And it gets super, super important.

Matt Osman, CEO @ Treat

Generative AI can facilitate rapid creative variant testing and help marketers pinpoint the most effective creative concepts and visual elements.

Content Abundance and the Future of Marketing Tools:

As we enter a world of content abundance, marketers should focus on tools that help them understand which creative concepts to feed into generative AI systems. The production process will likely be integrated into existing tools like Photoshop and Figma, so the real value lies in finding tools that aid in the ideation and decision-making stages.

Generative AI is a powerful tool for ad creative, but it is crucial to ensure that third-party tools used for generating creative are not anchored into the actual ad network, as their incentives may not align with yours. The rise of generative AI highlights the competitive advantage of skilled creative directors and heads of ad creative, showcasing their value in the marketing process.

Discovering the Potential of AI in Marketing: Treat's Lookalike Creative

Treat's Lookalike Creative is an AI-based tool that analyzes a brand's past campaigns, including images, videos, and copy, to uncover visual patterns and correlations with specific goals or target demographics. This capability can help brands optimize their creative assets to better connect with their audience.

Uncovering Hidden Patterns:

In one case study, Treat's Lookalike Creative identified a specific type of marble bathroom that correlated highly with click-through rates for a beauty brand. This pattern was challenging to perceive across 15,000 creatives, demonstrating the tool's effectiveness in handling large data sets and pattern matching, tasks where machines excel.

It’s just very hard for a human eye to perceive, which was that we were working at a beauty brand. And there was one specific kind of marble bathroom. And it was a specific kind of marble, and it just happened to be really, really eye-catching for the machine that was highly correlated with CTR, which was the thing that they were trying to optimize for. Armed with that, the creative team can kind of do its magic and focus on the stuff that they’re good at. But the pattern matching across, I think this was across like 15,000 creatives. That’s just not something a human really should be either manually tagging or reviewing themselves. That should be outsourced.

Matt Osman, CEO @ Treat

Leveraging AI for Specialization and Division of Tasks:

By incorporating AI-powered tools like Treat's Lookalike Creative, creative teams can focus on their core strengths, while machines handle areas where they outshine humans. This symbiotic relationship between human ingenuity and machine precision is the future of marketing.

The Future: Shared Intelligence Database:

As we look ahead, there is potential for creating a shared intelligence database, where brands can opt-in to share effective visual elements with others. This collaborative approach fosters collective growth and innovation, further leveraging the power of AI in marketing.

Looking Ahead

The integration of AI in marketing represents a crossroads where growth marketers and design teams can specialize in what they excel at while allowing machines to handle tasks where they clearly outperform humans. In the future, it is possible to create a shared intelligence database where brands can opt-in to share effective visual elements with others, fostering collaboration and collective growth.

The value of AI in marketing lies in the specialization and division of tasks between humans and machines, allowing each to contribute their unique strengths to the marketing process.