How to prepare your marketing and analytics for a world without cookies

The business interests of large advertising platforms (Google, Facebook) along with legal requirements have resulted in technical restrictions that will reduce the lifetime of browser cookies and the ability of advertising systems to collect user-level data. Already in 2023, marketers won't be able to collect a significant part of the data they currently collect about the effectiveness of advertising channels. Due to incorrect data, marketers today are wasting 21% of their budgets, and this percentage will become even more significant with the abolition of cookies.

The task for your company’s marketing team in 2022 is to transition to a new analytics format that will be better prepared for global post-cookie changes. In this article, we explain how these changes will impact the effectiveness of marketing and the work of analysts.

What exactly is changing:

• Browsers have started limiting the lifetime of cookies or completely refusing to use them. Firefox already blocks cookie tracking by default. Google plans to abandon third-party cookies in the Chrome browser by the end of 2023. In Safari, the lifetime for third-party cookies has been reduced to seven days, and in some cases to one day. This means that the retargeting possibilities for third-party websites will be limited to one day. If a user doesn’t interact with the website during the course of a day, the cookie will be deleted, and on the next visit, the website will identify a new user. Because of this, it will be difficult to identify users and correctly determine their source of transition.
• Advertising services will limit the ability to track campaigns at the user level. Worldwide, there is growing concern about the constant collection of personal data, so advertising services will show aggregated data in the context of campaigns to protect users. This will significantly complicate cohort analysis and data audit.
• Platforms and browsers will limit user-level tracking. For example, iOS already blocks IDFA, which is an analog of cookies for users of Apple devices. By default, the use of IDFA in mobile applications is disabled. Without this ID, actions that an application sends aren’t associated with a specific device. Restrictions regarding the User Agent in the Chrome browser that will make it impossible to use most fingerprint techniques are also actively being discussed.

These changes are happening for three main reasons:

1. Business interests of market players. As the market consolidates and competition increases, big players (Apple, Google, Facebook) are making these changes to defend their interests.
2. Compliance with user data protection requirements (GDPR, CCPA).
3. Technical limitations. Some platforms and tools are making changes precisely because the technical capabilities for applying data or working with advertising campaigns are changing.

In analytics:

1. The share of new website users will increase, though plenty of these new users will actually be old users who have been assigned a new cookie.
2. The share of direct traffic will increase because most analytics systems use the Last Non-Direct Click attribution model by default. If a user clicks an advertising link on Monday and on Tuesday returns to the website directly, the system will assign the source of the second session to an advertising campaign, not direct. But if this is a new user, then there will be no connection with the advertising campaign that attracted the user.
3. The length of the conversion chain will decrease. If earlier it was possible to observe that a user made several clicks before completing a conversion, now the number of touchpoints that can be associated with one user will be reduced.
4. Cohort reports will be limited. Because cohort reports are built on user properties, these properties (a sequence of actions, an order) will no longer be possible to combine.
5. Attribution quality will decrease. Previously, you could rely on conventional and simple methods, such as Last Click, and directly connect campaigns with sources. Now, the connection of advertising campaigns with conversions will be less accurate.
6. Campaigns will look more like the last click, as only the user history during one session or one day will be available to the analytics system. Accordingly, assessments by such systems will become closer to the Last Click attribution model.
7. The role of associated conversions in evaluating campaigns will grow. Advertisers have already started to evaluate many channels by considering associated conversions since it’s impossible to assess them by the proportion of users who clicked on advertising.

In advertising campaigns:

1. The cost of attracting a client will increase. Advertising services will have less information to target an advertisement to a suitable user. The less ability an advertising service has to determine that an offer is in the interests of a particular user, the less relevant its offers will be. This means that CTR will decrease and CPA will increase.
2. Coverage of lookalike and retargeting campaigns will decrease (assuming these types of campaigns remain). In fact, retargeting will work only for as long as a third-party cookie lives. After receiving third-party cookies, if the user goes to another website, such as Google, then the life of the third-party cookies will be extended. This means that retargeting, for example, in Google or any other walled garden can happen for longer than retargeting within a hypothetical Criteo. Because Criteo has less coverage through its websites, inaccessibility of information at the user level will not allow you to use lookalike audiences.
3. All this will lead to small advertisers leaving the market. It will be difficult for small players to prove their value with associated conversions.

The most important thing is to collect first-party and second-party data.

First-party data is what you can collect in your app or on your website. This is information the user provides you.

Second-party data is collected by the advertising service — for example, data from advertising accounts.

1. First of all, you need to implement a marketing data lake such as Google BigQuery, AWS Redshift, HP Vertica, or Hadoop. To do this, you need to collect data somewhere, and it also needs to be controlled by you, not by an advertising service. As long as your campaigns and user activity are stored in different systems, especially those that don’t belong to you (advertising account, Google Sheets, CRM), a comprehensive analysis of your advertising campaigns isn’t possible.
2. Collect raw user activity data from your website and mobile app. By raw data, we mean that each user activity is registered and saved. This isn’t aggregated data on the number of visits, for example. There should be no metric as a required parameter. Each interaction is a separate hit. The value of raw data is that it isn’t aggregated and, therefore, is accurate. In addition, ad blockers and browsers don’t limit the collection of raw data.
3. Import the most granular data from advertising accounts into your data lake. Many accounts are limited only to UTM tags for tracking transitions from advertising accounts. But this isn’t enough to build an analysis without connecting data to a specific user. For example, Facebook Ads allows you to upload up to 200 fields. Using raw data, you can build reports of the depth you need with any parameters. For example, with geodata, you can analyze the effectiveness of your advertising campaigns in different regions. In addition, you can use accumulated data for machine learning and forecasting the best marketing mix. Granular data from advertising services is required for performing calculations.
4. Motivate users to log in to the website and mobile app. This is a significantly underestimated activity that allows the advertiser to collect data in the context of a particular user without IDFA and cookies. We’ve collected more than ten ways to increase the share of identified users in a separate article. We recommend that you check whether you’re using all the methods in your business.

Considering the upcoming changes, we've developed a solution that will help companies smoothly switch to a new analytics format without losses.

With OWOX, you will receive:

1. A clear plan for preparing for the post-cookie world that is developed for your specific project.
2. The ability to use first-party cookies with a long lifespan through OWOX BI server-side tracking to identify users. This will significantly reduce the percentage of direct/none traffic.

The main reason for the inability to determine sources of user sessions (direct/none) is the absence of a stable user ID.

The primary identifier for an unauthorized user is currently Client ID. It’s set by Google Analytics JavaScript code in the user’s browser and stored in a cookie. This cookie is treated as a third-party cookie and is subject to restrictions in browsers including Safari.

The result of ITP in Safari is a reduction in the lifetime of a cookie with a Client ID to no more than seven days and in some cases only one day.

How to solve this problem? Make cookies perfect for browsers. Here’s what this means in technical terms:

• Set cookies not with a JavaScript code but with a server
• Assign cookies with httplOnly, Secure, and SameSite parameters
• Install cookies on behalf of the domain of the same site (via A-DNS records)

OWOX writes its unique user identifier to the owoxUserId cookie, which allows you to correctly determine transition sources. Our own experiments have shown that collecting data in the first-party context results in a:

• 2.1 times increase in the number of tracked returned users to Safari Mobile
• 9.6 times increase in the number of tracked returned users to Safari

1. Comply with personal data protection requirements (GDPR, ePrivacy).
2. Configure a GA/GBQ proxy to resolve data loss problems due to AdBlock.
3. Regularly collect and enrich raw hits in the usual session format. For analysts, it’s easier to work with data this way, and as a result, users get more opportunities to derive value from the collected data.
4. Model business-ready data for use in marketing and product reporting.

In addition, the OWOX team can help you with setting up Google Analytics 4 to collect complete, high-quality data both there and in Google BigQuery.

What the OWOX team will do for you as part of this solution:

1. Audit your Google Analytics and Google Tag Manager accounts and compile a personalized roadmap of changes to switch to post-cookie tracking.
2. Set up data collection in Google Analytics and Google BigQuery.
3. Transfer all web/app data you collect into Google Analytics 4.
4. Set up OWOX BI server-side tracking, which collects first-party data and combines it with marketing data in your repository.
5. Configure Consent Mode for correct and secure first-party data collection.
6. Set up cross-device profile generation and ready-to-use (business-ready) data.

Familiar ways of working with marketing data are gradually changing. Personalization of advertising, retargeting, and so on is becoming more difficult due to public outrage about privacy violations and the resulting laws and rules. Influential companies have already begun to change the way personal data is used, and this situation can no longer be ignored. Businesses should prepare their marketing and analytics for a post-cookie reality and new conditions for working with personal data.

It’s obvious that you should prepare for a post-cookie future before you lose access to some of your data. You can evaluate campaigns in a changing environment by calibrating your analytics and metrics systems.