The data paradox: 4 changes that brands must prepare for

As a result, privacy is becoming a card to play in our marketing arsenals. Privacy features are being routinely pulled out as key selling points in marketing communications.

There’s no bigger example than Apple, which has the perfect scenario of being able to address legitimate consumer concerns and hurt direct rival revenue streams of Google, Meta and Microsoft by restricting their access to Apple’s customers’ data. Despite consumer privacy options spreading over the past few years with GDPR in Europe and CCPA in California, 96% of consumers still choose to opt out of sharing data.

Consumer nervousness to share data creates a paradox. Consumers are nervous to share data for personalized ads, yet it’s personalized ads that power much of the technology consumers use daily.

Consumer education and understanding will take time. Consumers want to understand what data is being shared, what it’ll be used for and whether there’s a clear value exchange offered. For marketers, this is where the real challenge lies. First, leveraging privacy as a potential selling point and second, offering value in exchange for consumer data.

Although this is how marketers can get ahead of the competition, privacy will affect marketing efforts whether you act or not. There are four key changes that businesses can prepare for.

1. Conversion modeling

We've seen a huge rise in conversions being 'modeled' rather than accurately tracked. Conversion modeling leverages machine learning to estimate the number of conversions being driven by marketing when tracking limitations exist, giving a guesstimate number rather than an accurate one. Most notably this is across iOS devices but expands to wherever cookies are rejected.

This is leaving more marketers in the dark as platforms (namely Google and Meta) begin to further mark their own homework.

2. First-party data sharing

Google and Meta have been pushing first-party integrations. This encourages businesses to share an email address or phone number of converting customers to better allocate marketing credit to their platforms.

It provides substantial increases in performance, especially for Facebook campaigns, which without it are blind on Apple devices. The integration of this data allows platforms to better self-optimize and is becoming an essential requirement for successful online campaigns.

3. Web analytics

GA4 caused a stir when it was announced back in October 2020. It’s taking some getting used to for those who have taken the leap from Universal Analytics.

Aside from the monumental shift in tracking models and all the new features of GA4, user privacy has been pushed closer to the forefront.

New privacy features allow businesses to be more in control of how Google Analytics collects data and how long it’s retained; they also increase the ease of data deletion from the platform.

However, a 14-month data retention period (in which marketers are unable to look back on data from this period) will be a big hindrance to marketers. Luckily, Google has made it free for GA4 data to be extracted to BigQuery should GA4 users need to store this data for longer.

With BigQuery comes endless opportunities to drive more valuable insight using GA4 data. Supplementing with wider business data or bringing in other datasets can uncover factors to help businesses better understand their customers, channel performance and effectiveness of marketing against wider data and business strategies.

4. Federated learning & contextual marketing

Federated learning is a workaround that many browsers are using to continue allowing interest-based advertising to power ads. Rather than using the behavior of an individual, the browser amalgamates the browsing behavior of individuals and creates cohorts of users. The data is grouped together and processed using machine learning to understand behavior patterns.

However, whether it’s cohorts of users or individuals, the browser needs data, and lots of it. The push for privacy removes specific data assigned to one user but introduces more powerful methods of processing the data.

Federated learning is a workaround for audience-based targeting, but arguably the most accurate and futureproof method of targeting is contextual targeting. It does what it says on the tin: it scans pages in which ads can be shown, understands the content and uses keywords that advertisers provide to understand if it’s suitable to serve an ad.

This sounds straightforward but machine learning and AI still play a part in enhancing this method of targeting. The words ‘killer’ and ‘die’ in most instances would be flagged as inappropriate but increasingly sophisticated ad tech is able to appreciate the wider context of the page. These words could be used legitimately in a brand-safe environment for ‘killer recipes’ and ‘brownies to die for’. The world of contextual targeting has a long way to go but it’s a sure-fire way of knowing you’re hitting the right audience.

These are just a few of the changes coming our way, with some like conversion modeling being slipped in without much fanfare. Our advice to marketers? Own them, understand them, and test them. Businesses that attack these things head-on typically outperform their market.

For more on how the world of data-driven advertising and marketing is evolving, check out our latest Deep Dive.