Can the online customer experience remain intact in a world bereft of third-party cookies?

Marketers and journalists have been lamenting the demise of third-party (3P) cookies since 2020, when Google announced that it would phase them out.

 

Or at least, it’s thinking about phasing them out. Since the announcement, everyone has been waiting with bated breath for an answer, given that the tech giant’s Chrome browser controls more than 60 per cent of the global market.

 

Meanwhile, however, relatively smaller players – Mozilla and Apple – have been putting their cookies where their mouths are. The former blocked third-party cookies in Firefox in June 2019, while Apple’s Safari browser followed suit in March 2020.

 

Google’s prevarication over whether to ditch cookies was mostly thanks to the shortcomings of the new Privacy Sandbox ecosystem it designed to replace 3P cookies.

 

While the first iteration of the Privacy Sandbox raised regulatory concerns, with similar criticism voiced by the adtech industry, the API toolkit Google developed in response performed poorly against 3P cookies.

 

A 2022 report by Criteo, a company that specialises in ad retargeting using third-party cookies, for example, found that Google’s Topics API was five times less relevant for interest-based advertising than third-party cookies.

 

Fast-forward to 2025 and – following further performance enhancements to the Topic API and its user controls – 3P cookies are still an integral part of the Google advertising universe which, Google now suggests, should be used in combination with the tweaked API tools.

 

Have third-party cookies been overrated?

 

Although Google has been dominating the adtech market for years, and has the power to give cookies the thumbs-down or grant them a reprieve, third-party cookies seem to have a lifecycle of their own.

 

As they are falling out of favour, it’s not hard to come across internet opinion that reveal their weaknesses. They have been referred to as a privacy disaster, the tool of the lazy marketer and publisher and as “lousy proxies for a consumer’s identity.”

 

The data that 3P cookies – small text files saved on a user’s computer at the behest of a website they visited – add to a consumer profile are indeed far from perfect.

 

The factors that can skew their reliability include users’ tendency to share devices and clear caches – or, in other words, delete temporary files their browser has stored.

 

Moreover, clickbots, which are used extensively to fraudulently create traffic for ads, have also made the collected data noisy.

 

As cookies are placed on devices, targeting the same individual across their gadgets is a challenge, which often results in the frustrating buyer experience of being swamped by online adverts on a laptop for products the user has already bought on their mobile.

 

Changing habits, and the rising popularity of applications, have also weakened the appeal of third-party cookies, given that apps operate outside the browser environment and use alternative storage mechanisms.

 

The deficiencies of 3P cookie tracking and targeting are also demonstrated by metrics of what’s called cookie syncing. As cookies are domain specific and only the administrators of websites that placed them can read them, adtech companies only have access to their own cookies. To attain more information for better targeting, their own data must be matched with the data other platforms have collected about the same individual.

 

In other words, the cookie-synching process matches the user information that comes from another source with the data a vendor already has on the same person.

 

The fact that the average success rate of cookie-synching is between 40 to 60 per cent is eye-opening proof of the limitations of 3P-cookie-based targeting. It means that at least 40 per cent of users are not matched successfully in the process and thus can’t be shown more relevant ads.

 

You’ll miss them when they’re gone

 

It would be unfair, however, not to highlight the fact that, despite their constraints, 3P cookies have been providing the lifeblood of a highly successful ecosystem.

 

In fact, martech providers will need to stitch together a patchwork of technologies – old and new – to cover all the functionalities of the versatile 3P cookie.

 

The old regime is already being replaced by a browsing-session-based contextual approach, where users are shown ads that match the context of the site they are visiting.

 

For example, on medical websites they may see pharmaceutical ads, while websites focusing on interior design would be more likely to show adverts for soft furnishings.

 

At the core of contextual advertising, there lies the simple mechanism of matching URLs with a brand’s campaign brief.

 

The process is based on the analysis of websites’ contextual relevance and matching them with brands that feel good about displaying their ads in that environment.

 

Similarly to how web searching today goes well beyond simple keyword matching, identifying the most effective content that the product should be embedded in also leverages the functionalities that machine learning (ML) excels at.

 

ML can understand the totality of the contextual relationships that exist within a website – whether between words, pictures, videos or podcasts. This way, people can be targeted effectively using this analysis.

 

Some studies show that contextual advertising can even bring a better customer experience than ads powered by 3P cookies – at least in certain respects. When achieving a good fit between the advert and the website content, they can perform well on metrics such as ad-recall and message association – to put it simply, they stick better.

 

In contrast with 3P cookies, first-party cookies collected directly by the website that the user visits are here to stay. Used in combination with contextual advertising, they can hone targeting by providing data about user behaviour and preferences displayed while on the website.

 

Meanwhile, tracking across websites – the 3P cookie superpower – can be delivered by a consent-based version of identity trackers. These so-called universal IDs, which eliminate the need for cookie synching and enable retargeting, typically use encrypted, hashed email addresses as identifiers to protect privacy.

 

As this emerging ecosystem is built on trust and transparency, another major source of data it relies on is a customer indicating a potential intent to buy, which can range from signing up for a product demo, a webinar or a free trial to leaving an abandoned cart behind.

 

Is privacy-first tracking only a passing phase?

 

The technologies described above may grow organically into a seamless ecosystem which will rival the efficiency of the 3P cookie universe while also protecting user privacy.

 

Alternatively, the era of the post-cookie patchwork of technologies may turn out to be a fleeting one, just one element of a greater inflection point similar to the shift from horse-drawn transport to automobiles. In this scenario, it’s not only 3P cookies that are fading into insignificance but also web browsers as we know them today.

 

If or when search engines finally give way to generative question answering engines such as ChatGPT, Claude or Perplexity, marketers will probably shift their focus from humans to AI agents and their IDs.

 

However futuristic gen AI agents organising entire holidays and interior design projects for us may now sound, it could happen earlier than we think. Microsoft, for example, has predicted that conversational agents will make search engines and web browsers obsolete in three to five years.

 

No matter how fast things pan out, though, privacy must remain central to a new ecosystem as our AI agents will eventually reflect our own habits, budgets and preferences.