Navigating marketing with a focus on data privacy and compliance

After driving online tracking and personalization for years, third-party cookies are on their way out. The privacy concerns associated with tracking software have led internet giants to discontinue them. Apple and Mozilla have already got the ball rolling with their respective browsers, while Google is on track to follow suit sometime in 2024.

While this is great news for privacy enthusiasts, marketers are not thrilled with the development. They are also exploring alternate ways of driving their marketing and personalization efforts, such as bringing artificial intelligence (AI) into the loop.

“When marketers think about AI, they broadly think about three separate types of capabilities. First is generative-AI stuff, which is the ability of AI to create text or images (like GPT-3 or Dall-E) for content production,” Andrew Frank, distinguished VP analyst for Gartner, told VentureBeat. “The other area is the analytics side, which covers things such as emotion sensing or eye tracking to infer attention and response to an ad. Finally, the third capability is AI-based decisioning, where an AI decides how to personalize a site or what offer to send you next.”

Bringing AI and data into the mix

With these AI-driven capabilities, marketers can quickly scale to reach more customers with relevant content and campaigns. However, when the talk is about AI, data becomes an inherent part of the conversation. It is the “fuel” that powers the models, but also one that should be used responsibly to ensure regulatory compliance and the trust of customers.

“The increase in data privacy regulations is forcing brands to take greater control of how they understand their customers,” said Raj De Datta, CEO of ML-driven personalization company, Bloomreach. “…They must truly understand who their customers are (instead of using cookies) to personalize and scale marketing efforts. The challenge this presents for marketers is that it forces them to take stock of how they capture and organize first and zero-party data.”

First-party data is information gathered from customers’ web activity, while zero-party is the data consolidated through surveys and interviews. Both are equally important for marketing, which is why companies have already started looking at their whole data infrastructure to figure out how it has to change to store information in accordance with laws and technical restrictions. The matter, as Frank said, has escalated up to the highest level across organizations.

The ideal approach: Minimize use of personal data

To use first-party information, which is often a point of contention, with a focus on privacy, organizations need to focus on a few key elements, starting with minimizing the collection and use of personally identifiable information (PII).

“A lot of algorithms today don’t necessarily rely on personal data. In fact, there’s maybe a misconception that in order to do personalization well, you really have to have a lot of personal data and you really have to have intimate knowledge about people’s personal behaviors and habits from history,” Frank said. “This isn’t the case. It’s possible to infer things about people’s intent and priorities, just based on their behavior in a session, without necessarily capturing any data about who they are or any data that would be considered personal data.”

Marketers, to begin with, should try to make the most of technical information at their disposal – like device, language or operating systems being used — as well as contextual data that’s available through multiple tech touchpoints. The latter could include anything from general location or time of customer interaction to content-based contexts, like what the person was looking at or what path they took to arrive at a certain location or experience. Then, marketers can build on it by bringing synthetic or artificially generated data into use.

Getting to know customers without relying on cookies

“In the cookieless world, privacy-respecting, AI-based marketing will be able to learn from broad, unlabeled signals and generate engagement and sales from a much larger set of people than today,” Nadia Gonzalez, CMO at Scibids, “The best practice here is to only utilize signals that do not require user tracking and profiling for targeting.”

If all these elements are not enough and the company still sees clear consumer value from gathering some level of personal data, then it should clarify the value proposition and seek consent for the data. This way, customers could weigh the benefits against the level of intrusion, and make a rational decision on whether they want to opt into sharing or not.

“In case you’re a specialist retailer or a mono-brand retailer, you might have a good case for wanting someone to personalize their experience because they’re intimately involved with the products and you have a transactional relationship that you can use to recommend new products or remind them when it’s time to get service,” Frank said. “However, if you’re a CPG manufacturer that sells mostly through the retail channel, then maybe you don’t have such a good case for why someone would want to share data with you.”

“For the most part, you’re going to have to think about how you can acquire aggregate data through collaborative strategies, like data clean rooms, to understand patterns and trends and help inform your marketing strategy, without getting involved in the acquisition of personal data,” he added.

Transparency is critical

However, whenever one resorts to collecting personal data, just making the value proposition clear will not be enough. The organization should also ensure transparency by fully disclosing and explaining how the data sought will be put to use — and that it will be deleted or modified if the user chooses so. This way, marketers can gradually build trust with their customers, ultimately giving them enough confidence to be open to sharing their information. 

“You can’t assume you’re too small to get caught or that certain regulations don’t apply to you. Data privacy regulations will only continue to grow, and businesses that are building their practices (marketing or otherwise) with their customers’ privacy at the forefront today will be ready for whatever may be down the road,” Datta said. “When it comes to your marketing, be transparent about how you are using peoples’ data. Simple things such as ‘These items are here because you marked them as favorites’ help to build trust with your customers.”

Yes, at times, going into the details of data use can prove tricky, owing to the complex nature of AI algorithms. However, in these cases, companies should focus on creating a pathway that could provide the required information while not becoming burdensome or disruptive at the same time.

Prepare for conflicting scenarios

In addition to this, enterprise marketers should prepare to deal with certain conflicting imperatives owing to regulatory restrictions. 

Many privacy laws require the deletion of data when requested by users or when it has outlived its purposes. And at the same time, there are also regulations that require certain data to be retained for different periods depending on the category of data and the region in which it was collected. Alternatively, there can be conflicts between wanting more data for reasons such as eliminating bias in algorithms or keeping less of it for privacy. 

While this can be complicated and become worse if states continue to pursue the enforcement of local laws outside their geographical boundaries, Frank suggests planning for different scenarios instead of settling on a single strategy.

“You can’t settle on one choice and have to create enough flexibility and adaptability in your solution. This way, if things take a turn one way or the other, you’re able to adapt quickly and effectively to the way that the situation unfolds because it’s very unpredictable and unstable right now,” he said.

The big opportunity

More than 80% of industry experts already integrate some form of AI in their marketing activities, and the adoption is expected to grow thanks to the obvious benefits around targeting, personalization and analysis. According to Statista, the market for AI in marketing was estimated at $15.8 billion a year ago and will likely surpass $107 billion by 2028. 

Clearly, there is much innovation to come by all players in the industry that will make marketing more effective. But, as Gonzalez pointed out, driving ROI will become increasingly dependent on AI solutions that do not rely on personal data or cross-site behavioral analysis.

“We all have a great opportunity today to grow the category and significance of digital marketing while reducing friction between consumers, brands and the economic engine of the Web.

“The future of ad tech and post-cookie digital marketing and the business growth it enables is bright. We have the technology to leave burdensome, intrusive, legacy technology from another age behind,” she said.

Originally appeared on: TheSpuzz