Updated Jan 25th 2022: As of this date Google has announced that FLoC is no more. In its place is the Google Topics API. You can read the full API readme here, but in short...

"The intent of the Topics API is to provide callers (including third-party ad-tech or advertising providers on the page that run scripts) with coarse-grained advertising topics that the page visitor might currently be interested in. These topics will supplement the contextual signals from the current page and can be combined to help find an appropriate advertisement for the visitor."


Since this blog post was published, Google has announced that it is pushing back the final deadline for its phasing out of third-party cookies to late 2024.


With any luck, we will have left third-party cookies behind us by this time next year.

As Google drip feeds us more news and updates about “removing support for third-party cookies” in Chrome, the death of these trackers is increasingly on the lips of digital marketers everywhere.

Understandably, each time Google makes an announcement people sit up and listen. But remember, Firefox, Edge and iOS users have had significantly restricted cookie environments for a while, and we are still running effective ad campaigns.

So, there is no need to worry - but things do need to change. And you do need to prepare.

How do you prepare for a future without third-party cookies? In short, work out what you use third-party cookies for now and plan to replace them with an alternative solution.

We have summarised below what we believe the impact of this will be based on current information (this is an ongoing evolution). We will update, share our view, and help with critical actions as and when we learn more.

What Won’t Be Impacted Once Third-Party Cookies Are Disabled in Chrome?

The "Walled Gardens" of Ad Tech

The big ad tech platforms will likely benefit from this, particularly Google.

In fact, authorities are becoming aware that a consequence of ditching third-party cookies looks set to cement the monopoly of the likes of Google and Amazon.

Those users logged into Google, Facebook, Microsoft, Amazon, etc... will still experience tailored ads, because the tech platforms have an information-rich, first-party relationship with them. That means they can use additional signals in real-time, based on context and interactions.

Within these “walled gardens” of first-party data, there’s no reliance on third-party cookies - they can pretty much carry on as they were, as publishers on the “open web” are impacted.


Advertisers will also be able to continue to use first-party data mechanisms (Google’s customer match, Facebook’s custom audiences).

They will also be able to buy clusters of audiences based on the platforms’ newly developing tracking solutions and typical contextual and behavioural targeting. This is based on users’ interaction with the big ad tech platform websites not using supplementary data from other 1-2-1 website interactions.

First-Party Cookies

First-party cookies will be around for a little while yet. Remember, Google is developing "a path towards making third-party cookies obsolete", not all cookies. That may come, but it is not imminent.

Using cookies on your own websites (such as Analytics cookies) will continue.

What Will Be Impacted Once Third-Party Cookies Are Disabled in Chrome?

Conversion Tracking

Conversion tracking is already changing. Currently, ad platforms rely on third-party cookies (and “pixels”) for a large amount of their conversion tracking. They are investing lots of time and effort into uncovering new solutions that address this.


Advertisers who rely significantly on remarketing. The days of visiting a shoe retailer site, almost buying those brown shoes, then being chased for days with images of those same brown shoes look set to be over - thankfully!

Those advertisers who rely on this as a significant chunk of their ad campaigns will need to start using the smarter, newer, alternative ways of reaching prospective customers.

Profiling and Targeting

Audience profiling and targeting will have to change. In the past, the targeting of ads on the “open web” - outside of the walled gardens - has been driven significantly by cookie-based, behavioural data shared with the main ad tech platforms.

The ad tech platforms are focusing on solutions to protect their own inventory.

Effectively, this should help eliminate the murkier corners of publisher data trading, meaning a much more transparent buy market. Ad inventory prices may rise as authorised publishers claim their inventory has “higher value”, but we’d expect increased effectiveness of this inventory to help balance any price rises.

Frequency Capping

Frequency capping will be difficult. An unfortunate consequence of increased privacy is that we will have less visibility of how many times we serve the same ad to an individual.

Not a problem for some advertisers who have never seemed to care anyway, but for the rest of us we’ll need to be smart with how we limit our display ad spam levels.

Programmatic Measurement

Precisely measuring the impact of programmatic display advertising. The days of justifying display activity by using metrics like “view-through” conversions may be over.

Joining the dots between the specific “impression” on an “open” website and the downstream conversion will not be possible. Other planning, measurement and modelling approaches will need to be taken.

What Are the Proposed Solutions?

Privacy-Conscious Cookie Solutions

As we have already said, not all cookies are going to fully disappear overnight.

Third-party cookies will certainly be phased out, but there is still a significant amount of investment in developing more secure first-party cookie solutions (cookies that are set and used by your website).

If cookies were to be disabled overnight, it would have a prohibition-like effect (we’d be swamped with the cookie equivalent of moonshine). To help avoid this, Google and others are simultaneously developing technology to prevent these deceptive and intrusive solutions, such as browser fingerprinting.

One example of continued development of cookie solutions in this area is with Google’s “consent mode”. This is in trials as a mechanism within Tag Manager that will easily allow management of tags based on user consent levels. It is a “privacy-preserving” mechanism that still uses cookies and shows the cookie isn’t dead quite yet.

Cookie Alternatives Are Evolving in Many Forms

There are a variety of alternative tracking solutions being examined and tested.

Server-To-Server Tracking

A significant part of Facebook’s ongoing battle with Apple over tracking restrictions is for advertisers to use their “server-to-server” conversion tracking solution, known as Conversions API (“CAPI”).

Essentially, by connecting Facebook users and the advertisers’ customers, there is no need for a “browser- or device-based” tracking connector (e.g. a cookie). Instead, the source and destination servers can talk to each other directly.

Google now also enables this by using server-side tagging within Google Tag Manager. The principle is the same - no need for cookies within a browser, we can join a source and destination together in the cloud by matching up the “logged-in” users.

On-Device Targeting

Instead of keeping user data secure by storing it on your first-party server, an alternative is to do everything on the device. Questions about privacy largely arise when the user data is shared across the web, so this is equally as valid a solution as server-to-server.

Two examples being developed by Google are:

FLoC. (Federated Learning of Cohorts). This approach “hides” individual users in a crowd (or cohort) of people with similar interests and uses on-device processing to keep individual browsing habits private.

FLEDGE. Interest groups (possibly determined by FLoC) are stored in the browser and assigned to arbitrary, benign metadata. The metadata will then act as a signal to the ad platforms for bid management or targeting in real-time.

In both cases functionality such as targeting and bid management should be possible by using on-device, clustered signals to deliver real-time experiences for users. These solutions are all still being tested.

Other User ID Solutions

The advertising industry has been exploring alternative ways to track and connect users with their behavioural attributes for several years.

Exploring these solutions will not stop, even though Google has said they will not be investing in them. Google has said that it will allow publishers to use their own solutions for targeting users where the publisher has a direct relationship, plus they have said they do not intend to block other solutions.

However, they have mentioned “authorised publishers”, so we can expect that publishers will need to conform to the ad platform standards in some way (possibly by obtaining certifications).

Growing Your Own First-Party Data Pools Is Critical

First-party data is already essential to the ad platforms themselves. Think of all the logged-in users that Google, Facebook, Apple, and Amazon have.

Advertisers could piggyback off this data for a long time. For example, we all know how effective the targeting can be when you use Google’s “customer match” and Facebook’s “custom audiences”. If you are not using these existing customer match type solutions, you should be.

Furthermore, as cookies phase out, the more first-party data you have, the more you can take advantage of not just these existing mechanisms, but the inevitable new first-party mechanisms that will follow.

Make sure you are being creative with ways of collecting more first-party data. For example, you could develop compelling content and assets that have enough of a “value exchange” that users will hand over their data to you.

Do not underestimate how important this will be - by building these first-party data sets now, you will be heading into 2022 in good shape.

Sampling and Modelling Helps Fill the Gaps

Machine learning, Artificial Intelligence (AI), and algorithms. These are already the key, unseen foundations of effective digital advertising platforms.

As cookies phase out and users become even more protective of their data, the depth of rich personal data on some individuals may reduce, but we now have tools that enable us to use the remaining data to “fill the gaps”.

The latest version of Google Analytics (GA4) is a great example of this. Using already existing data sets (e.g., Google signals) Google is building ever more powerful modelling to ensure the reporting built on a smaller set of data is robust and accurate.

Many of these models are built out from the endpoint of a journey (where the customer spends money) so are often referred to as “conversion modelling”. By building statistically robust models, the platforms can use the observable, permission-based, data to make accurate assumptions about the non-observable data (the users who have not given permission).

More Creative Solutions

Dynamic creative will need to make better use of the other data signals available to them (location, context, weather, time of day).

In the past, we have relied too heavily on things tied specifically to each individual, but data that’s readily available to cohorts will still enable excellent dynamic creative execution.

Furthermore, chatbots and conversational tools could become incredibly useful. With these user-initiated tools advertisers can take people on a journey by encouraging them to interact with the ad creative itself, while doing so can tailor that experience.

Look at adlingo.com as an example of “conversational creative”. An advertiser may not learn anything about that individual beyond that session, but they will be able to tailor an experience at the time. They could also try to encourage the user to give permission to take the tailored experience beyond that session, which will require a worthy value exchange.

11 Ways to Prepare for the End of Third-Party Cookies

1. Make sure you are using existing first-party data solutions

The likes of Google’s customer match and Facebook’s custom audiences etc. You are missing out already if not.

2. Create a first-party data roadmap

Make someone internally responsible for considering questions such as: what first-party data do you currently have? With what permissions? How are you currently reliant on third-party cookie solutions? What first-party data could you gather to fill these gaps?

3. Find opportunities to grow your first-party data

Think beyond the customer data you are hopefully already using for “customer match”. How can you encourage more individuals to share their data with you?

    To supplement your customer data, could you build a pool of data of people who are product advocates? How can you encourage customers to tell you more about themselves? Club members? Newsletter subscribers? Effective conversion optimisation will be critical for persuading prospective customers to exchange their personal data with you.

    4. Continue to seek permission

    As you have already been doing since you heard about GDPR. Audit your tracking and your data to ensure you know what you have got and why. Update your privacy policy to be clear on how you use data (including if you are planning to use first-party data for the customer audience features we recommend).

      Make sure website visitors can manage their consent - consider the options becoming available with “consent mode” and other permission management tools. Further information and support are provided by Facebook here and Google here.

      5. Take an active part in upcoming ad platform tests

      The tech giants are still innovating and trying different things, so take part in the trials/betas when you can.

        Google is currently encouraging the use of “value-based bidding” for optimisation and a new first-party cookie solution to help with conversion tracking. Do not be suspicious of Google pushing these new approaches, their current priority is to move to privacy-focused solutions.

        6. Implement Google Analytics 4 (or other server-side analytics solutions)

        GA4 is designed to work well with server-side tracking (as opposed to cookies) and is being built with more powerful cloud-based, machine learning modelling, which will be better positioned to “fill the gaps”.

          It won’t replace the current version of GA any time soon, but it’s worth getting them up and running alongside one another.

          7. Implement site-wide, first-party tagging (Google Tag Manager)

          This will ensure you are collecting as much first-party conversion data as possible. These conversions are ultimately what the ad platform tools are designed to optimise towards.

            By gathering the largest data sets possible currently, your ability to learn and model from that data in the future will be increased.

            8. Explore ways to use your “conversion data”

            This means Facebook Conversions API (currently available) and Google’s enhanced conversions (in Alpha testing).

              These server-side connections allow a reliable end-to-end relationship to be formed, which can ensure optimisation of campaigns, without getting drawn into mapping all the touchpoints along the journey.

              9. Revisit “test and learn”

              There is a reason “control, test and learn” methodologies have been around in marketing for decades; they work. They also do not generally trigger a race to the bottom.

                Campaign performance can be continually improved by having a robust approach to intelligently segmenting audiences, carefully planning media campaigns, running incrementality tests, measuring the uplift, etc...

                You do not need to track every online interaction of every prospect. Deprioritising things like attribution modelling to focus on the bigger picture could be a blessing.

                10. Invest in your data team

                There have always been gaps in the data we can track online and there always will be. Rather than obsessively trying to fill the gaps, be confident in the data you do have by investing in people who can rigorously analyse that data and accurately build models, which support your test and learn approach.

                11. Start testing dynamic creative driven by other types of data

                The days of product-oriented 1-2-1 retargeting may be over, but there are loads of other ways you can serve up interesting, engaging, smart creative.

                  Whatever you do though, start with hypotheses about what creative executions will work for your audiences. Do not start with “what data can we find that we might be able to use”.

                  Final Thoughts and Further Reading

                  Just because advertisers can no longer track every user interaction online does not mean you can’t be smart and undertake brilliantly effective advertising campaigns.

                  Just be sure to plan ahead, know what the critical drivers of your advertising effectiveness are and be sure to have a way of evolving them as we head into 2022.