The advertising industry has long debated the reliability of third-party data. Recently, we published a study on issues related to the accuracy of socio-demographic data, which can be found here. Keeping these insights in mind, we decided to run an experiment to further test the reliability of third-party data – directly comparing it to Adlook’s Smart Audiences, powered by PrimeAudience.

‍

Our goal was simple: verify which audience strategy drives more meaningful engagement under real campaign conditions.

‍

Why we did it

‍

When it comes to digital advertising, the most important thing is to reach real people who care about the product or brand we are promoting. But unfortunately, that’s not always easy.

‍

We wanted to put our Smart Audiences to the test and see how they perform compared to other data providers on the market. The key questions we set out to answer were:

‍

• Who delivers better results?
• Are the audiences truly interested in what we’re promoting?
• What does user engagement look like?

‍

How we tested

‍

We wanted to make sure that all the other factors stayed the same, so we used the same DSP, the same bidding strategy, and the same budget. The only difference was the audience source.

‍

We ran a skincare campaign in the U.S. over four weeks, split into two line items:

‍

• One line item using third-party skincare data segments
• One line item using Adlook’s Smart Audiences segment

‍

Each banner included a short interactive survey asking users, “How interested are you in skincare?” Respondents answered on a 5-point scale, where 1 meant “very interested” and 5 meant “uninterested”.

‍

To ensure reliability, we applied two additional safeguards:

‍

• Control questions about declared age and gender, to verify consistency.
• Bot-detection filters that removed suspicious responses (e.g. unusual IPs, extremely fast replies, or multiple contradictory answers).

‍

Here are the screenshots of the survey on different sites:

‍

‍

‍

‍

‍

‍

‍

The results

‍

After filtering out suspicious responses, the campaign generated 1,594 unique survey answers. On delivery side, Adlook’s line item reached 3.47 million impressions, compared to 1.32 million from third-party segments, and drove over 6x more clicks at nearly the same cost.

‍

From there, the differences became even clearer:

‍

• Stronger reach and scalability – Adlook’s Smart Audiences delivered 160% higher reach.
• Engaged audiences, not just numbers – Click-through rates were more than 2x higher.
• Cost efficiency – The cost per click was over 80% lower.
• True relevance – Adlook respondents were nearly 2x more likely to declare themselves “very interested” in skincare, while third-party data had more users saying they were uninterested.

‍

• On-target delivery – Adlook’s Smart Audiences reached 32% more declared female users (a group responsible for over 60% of global skincare spend) and over-delivered in the 25–34 age group, another prime segment for the category.

‍

Key takeaways

‍

• Better engagement – Adlook’s Smart Audiences consistently outperformed third-party data in both declared interest and interaction.
• More reliable targeting – Results held steady over time, while third-party data showed declining engagement and lower interest.
• Future-proof strategies – Custom audience modeling not only scales better but also avoids the pitfalls of outdated third-party approaches.

‍

What this means for marketers

‍

The verdict is clear: relying solely on third-party data makes campaigns less efficient, less scalable, and less relevant. Adlook’s Smart Audiences, on the other hand, combine privacy-preserving, AI-driven modeling with real campaign validation – unlocking more engaged users, consistent results, and a great chance to achieve lower costs.

‍

It may be time for brands to rethink their audience strategies. With the right frameworks in place, reaching the right people becomes not only possible – but predictable.

‍