From Clicks to Context: Mapping User Journeys with Chrome's Game-Changing Topics API!

The perfect scenario for marketers is knowing the online journey of their customer base. Even with all the effort from digital platforms - search, social, third-party apps etc. - it's still a distant dream. This got me thinking. What about browsers? Particularly Google Chrome? Here's what I found after days of research…

A chance discovery …

I was reading an interesting article online about photography. I closed the Chrome browser tab with the half-read article in a flurry of things. After a while, I started thinking about the interesting work featured in that article. I opened Chrome's History tab to find that link. That's when I noticed two small tabs which read, By time and By group. And I thought, By group? Clicking on that tab, I found that Chrome groups my searched links by keywords! That's interesting, isn't it? It helped me find the link I was looking for and finish reading.

Just a few days ago, I was thinking about that under-the-hood feature. Actually, it was the icon next to 'By group'. Icon showing journey from one point to another. Has Google started tracking user journeys in its browser? Yes, for nearly two years.

… led to a very interesting observation …

I've been mentioning this for almost a year - user behavior and preferences are changing fast, and algorithms are evolving faster, which means online user journeys are fragmented. Of course, this has caused problems for tech companies and marketers. It's also a reflection of the challenges people face finding reliable information online. When you mix that with strict, but valid, regulatory oversight, the situation gets complicated.

I made an interesting observation while thinking about this. On one hand, Google pulled its third-party cookie deprecation plan. On the other hand, it's implementing user journey tracking in Chrome, which dominates the global browser market. That sounds about right. But, does it? And, is it?

… which revealed Google’s answer to mapping user journey with Chrome Journeys.

Finding the sweet spot between delivering trustworthy, reliable information, having efficient mechanisms for delivering contextually relevant ads, enabling the best possible user experiences, and keeping users' privacy is hard in this attention-based digital economy.

Launched in 2022, as part of the Privacy Sandbox initiative, mapping user journey with Chrome Journeys is designed to offer a more interactive way to browse through your online search history. I might have noticed it a couple of times but never really paid attention. Located in Chrome's History section, Journeys makes browsing history more interactive and informative and shows details like how you found a particular website, your browsing history, and what you found in between. Let's say you searched Google for a black hoodie with a zipper. In your search for the best hoodie that fits your budget, you browsed a few websites. For every website you visit, Chrome makes an entry in History. This is the most common way to find recently visited websites and pages. But it's not that easy. For instance, if you browsed through something else in between, the history will record those sites as well, in chronological order. History is great for finding recently visited sites, but it doesn't solve the big problem of keeping similar things together. This is where Chrome Journeys comes in.

Google intends to ‘stay on topic’ in a secured way …

From a user's perspective, Chrome Journeys gives them more control over their search history. What about Google, marketers, and advertisers? Google's deep tech behind its user-friendly UI could prove to be a better alternative to the previous Federated Learning of Cohorts or FLoC and third-party cookie deprecation initiatives. It’s called the Topics API. Topics API is designed to run interest-based advertising, but in a more privacy-friendly way than third-party cookies.

Simply put, Topics API analyzes a bunch of data samples that have been labelled with categories they belong to using a classifier model. Analyzing these samples helps the model recognize patterns and relationships. Then, when it gets new, unseen data, it can use its learned knowledge to predict which category it belongs to. By using this approach, it will map website hostnames to topics.

Topics API will assign a maximum of 5 topics to a user each week based on what they visit, along with a sixth random topic to provide reasonable privacy protections. Topics will be taken from a user's device (like a laptop or smartphone) and won't involve any external servers, not even Google's.

Whenever the Topics API is called by, say, an AdTech platform, anywhere from 0 to 3 topics are returned. Topics related to the user's search history will only be returned to the platform. If three topics are returned, then these three topics will be the top three topics of that user for the past 3 weeks - one topic from each week. Topics associated with a user will only be kept for 3 weeks, then new ones will be created. In the Topics API, any publisher or AdTech company can get topic IDs for each of the three topics, which can be looked up in Google's taxonomy. The topics will be empty if the user has cleared their web-browser history, browsed in incognito mode, or opted out.

Google’s new approach to gathering user interest data and sharing it with advertisers has some major benefits, like:

• The topics are chosen carefully to avoid sensitive categories like race and gender.

• It's easier to see and control how your data is shared when Chrome is used to gather interest data.

• With users' topics of interest, businesses can keep serving relevant ads without using covert tracking techniques like browser fingerprinting.

• Topics will be selected on people's devices, not on external servers. It's very similar to Apple's on-device approach.

• In order to limit the amount of data associated with one person at a time, topics will only be stored for three weeks.

• Topics of interest won't be captured in Incognito mode.

• With Google Search being the most popular search engine, Chrome being the most popular browser, and Android being the most popular mobile OS, data collection across platforms would provide better insights into user journeys.

… with a few visible gaps that could be filled with AI.

The Topics API approach has some gaps, too. For example,

• Topics are chosen based on popular sites visited by users, with each publisher representing only three categories. As a result, inventory and reader interests may be inaccurately represented.

• It's possible to collect data cross-platform, but there's likely to be gaps when the user searches on an iPhone or other browsers like Safari or Mozilla. It could also make it hard to build a complete picture of what users like.

• Topics are limited to domain-level interests, so there could be gaps in determining user intent, impacting user experience and context-based targeting.

Google introduced three new AI-powered features in its Chrome browser earlier this year. One was about organizing your browser tabs smartly. Chrome will automatically suggest and create tab groups based on your open tabs. If you're doing a bunch of things in Chrome at once, like researching a topic, planning a trip, and shopping, this can be helpful. That could be the front-end feature that links back to the Topics API, which makes sense. There's a pretty good chance that this and other future AI-powered features could fill the gaps I mentioned.

Topics API lets advertisers target users based on their browsing history, maps websites to topics of interest, and identifies the most important topics for users based on their browsing history. Through Chrome Journeys and smart tabs, it looks like a good middle ground between personalization and privacy. This could shape the post-cookie future of user journey mapping and contextual advertising.