Spring naar content

DDMA’s Data Driven Marketing Survey 2026 shows how far this has progressed. 70% of marketing professionals say their organisation uses AI for marketing activities, and 72% of organisations expect to expand that use in the coming year. According to the same survey, the main challenges are shifting from experimenting with AI to data quality, governance and organisational readiness.

This is the question at the center of DDMA NEXT on 26 November: how can organisations keep a competitive advantage when AI is available to everyone? When competitors use the same tools, the difference is more likely to come from what those tools work with. For most AI in marketing, that is tracking data: page views, conversions and the signals ad platforms use to optimize campaigns.

That data is often less complete than reports suggest.

Where marketing data gets lost

Data loss in tracking usually goes unnoticed. Reports keep updating, campaigns are running and the numbers look plausible. A lower conversion count is easy to attribute to a slower week.

A large share of events are removed before reaching analytics and ad platforms due to:

  • Ad blockers: According to GWI (2025), about 30% of internet users worldwide use some form of ad blocker. The most common reasons include too many ads and ads interfering with browsing. Privacy concerns come third. Many ad blockers also stop analytics and conversion requests, regardless of what the user intended.
  • Browser restrictions: Safari’s Intelligent Tracking Prevention limits the lifetime of cookies set by scripts, in many cases to seven days. When a visitor clicks an ad and completes a purchase after that period, the purchase can lose its connection to the original click.
  • Consent: Visitors who decline tracking have made a choice that organisations need to respect. A separate issue arises when an ad blocker stops the consent banner itself. Most consent management platforms (CMPs) load from well-known third-party domains that appear on filter lists. When the banner is blocked, the visitor sees the page without a consent prompt and the site has no record of a choice. This creates a measurement and a compliance gap at the same time.

Server-side Tracking reduces part of this data loss, but the first request still starts in the browser. When that request contains recognizable patterns, such as paths that include /collect, ad blockers can stop it before it reaches the server.

How incomplete data affects AI-driven decisions

Missing data used to be mainly a reporting issue. An analyst could spot an unusual dip and correct for it. Today the same data feeds automated systems that adjust bids and budgets continuously.

Automated bidding optimizes on the conversions the platform receives. When part of those conversions is blocked, the system works from an incomplete picture and treats it as complete. Audiences, keywords or creatives with more blocked conversions appear to perform worse, and budget shifts accordingly.

The loss is also uneven across audiences. TAGGRS tests show that the share of blocked requests is higher for technical audiences such as developers, marketers and analytics professionals. For organisations that target these groups, the impact on automated bidding is larger. A system that receives fewer conversions from a group will tend to bid less for that group, which can reduce the measured conversions from it further.

The same applies to AI tools that analyze or summarize performance. An AI assistant reporting on last month’s results works from the same data as the dashboard, and presents its findings with the same confidence whether the underlying data is complete or partial.

At RTB Day in Amsterdam this October, TAGGRS co-founder Niek Schlepers showed a standard Google Tag Manager script being blocked by an ad blocker in real time, next to a script that kept sending data. What we learned by talking to people during the event is that  most teams know that some data loss exists. Seeing it happen is often what prompts them to look at their own setup.

Beyond the AI-driven organisation: Ownership of the measurement chain

Measurement often sits between several teams. Marketing uses the data, developers manage the website, legal is responsible for the consent banner, an agency may manage the tags, and the ad platforms process what arrives. Each party manages its own part, and the connections between those parts often lack a clear owner.

The DDMO 2026 results point in the same direction. 66% of organisations are investing in AI policies, up from 54% in 2025, and 56% have established data governance practices. Responsibility for AI is moving to senior management (38%, up from 28%), and collaboration between marketing, IT and other disciplines is becoming more important.

In an AI-driven organisation, the tracking setup supplies the data that automated bidding and AI analysis learn from. Treating it like other business-critical systems means assigning an owner, monitoring it and agreeing on how changes are made.

The DDMA NEXT theme, Beyond the AI-driven organisation, focuses on how organisations keep an advantage when AI is common. Part of the answer lies in knowing which data AI systems use, and how complete that data is.

What a resilient measurement setup includes

A resilient setup consists of several layers. Each layer depends on the one before it, so data lost at an early step stays lost.

  1. Consent that reaches visitors. Serving the consent banner from a first-party subdomain keeps filter lists from blocking it. Setting explicit Consent Mode V2 defaults, with “denied” as the starting state, gives every tag defined behavior before the visitor makes a choice.
  2. Scripts and requests that reach the server. Many organisations already load their Google Tag Manager script from a first-party domain. The event requests sent after the script loads often still contain recognizable analytics patterns, and ad blockers can match them.

To measure the effect of this layer, at TAGGRS we developed and tested the Enhanced Tracking Script, which encrypts the full browser-to-server request instead of only masking the script URL. In tests with 30 partner organisations and on the TAGGRS website, masking only the script and using a first-party domain increased measured page views by 0.8% compared with standard client-side tracking. Encrypting the full request increased measured page views by 9.1%. For audiences of developers and marketers, the increase came close to 30%.

These results show that the point in the chain where data is protected affects how much of it reaches the server.

  1. A server-side control layer. Once data reaches a server environment the organisation controls, the organisation can anonymize IP addresses, remove personal data before it reaches ad platforms, add context to events and decide which platform receives which data.
  2. Monitoring. Comparing client-side and server-side request volumes shows how much data the server-side setup recovers. Comparing the number of recorded consent choices with the number of sessions shows whether the consent banner reaches all visitors. A consistent difference between these numbers is worth investigating.

These layers apply to data from visitors who have given consent. The aim is for consented data to reach the tools it is meant for and for every visitor to be able to make a choice. Visitors who use an ad blocker browse the site exactly as before, without ads. The only difference is that, if they have given consent, their visit is now included in your measurement.

Five questions to review with your team

These five questions help assess where your organisation stands before DDMA NEXT.

  1. What share of sessions shows a consent banner? Compare the number of consent records in your consent platform with the number of sessions in your analytics for the same period.
  2. Where do the consent banner and tag scripts load from? Check whether they load from your own domain or from a third-party domain shared by many other websites.
  3. Which audiences are likely to be underrepresented in your data? Consider whether your bidding strategies and AI reports account for this.
  4. Who owns the measurement chain from browser to ad platform? Name one person or team responsible for the full chain.
  5. Can you explain the data behind your AI’s decisions? Check whether you know which data an algorithm used when it shifted budget.

Conclusion

AI tools will continue to improve and become more accessible for every organisation in the market. The organisations that gain the most from AI are likely to be those whose systems learn from complete, consented data about their customers.

This work takes place in consent settings, script configurations and collaboration between marketing, IT and legal teams. It forms the foundation that determines how much value AI can deliver.

We look forward to discussing this at DDMA NEXT on 26 November in Amsterdam. To review your own setup beforehand, the guide How resilient is your tracking setup? explains the most common weak points and what to address first.

TAGGRS is partner of DDMA NEXT. Visit their session at DDMA NEXT | The New Marketing Reality 26 November in Capital C in Amsterdam.

Author:
Gosia Poddębniak is an experienced content marketer and writer with over five years in the SaaS, analytics, and SEO industries. At TAGGRS, she develops marketing materials across multiple channels and manages website content, covering topics such as data collection, marketing optimization, privacy, tracking, and analytics.

Malgorzata Poddebniak

Content Marketer | TAGGRS

Ook interessant

Lees meer
Artificial Intelligence |

Why AI-driven marketing depends on complete, consented tracking data

AI has become part of everyday marketing work. Automated bidding in Google Ads and Meta, AI features in analytics platforms and AI assistants that summarize campaign performance are available to…
Lees meer
Artificial Intelligence |

Hoe ethisch denkt de (data)marketeer? Help mee aan wetenschappelijk onderzoek

Over ethiek in marketing wordt veel gesproken. Zeker nu AI, data en personalisatie een steeds grotere rol spelen. Maar hoe kijken marketingprofessionals zelf naar ethische vraagstukken in hun dagelijkse werk?…
Lees meer
Artificial Intelligence |

Eerst de fundering, dan de AI-magie: zo verdubbel je het marketingresultaat

Wat nou als je conversie van 5% naar 10% gaat? Zonder dat het mediabudget omhoog hoeft, simpelweg omdat je als organisatie in staat bent om de juiste boodschap, op het…