Display frequency: When is it really effective and when does it just look that way?

3. July 2023

Which ad impression frequency per user in brand campaigns is the most effective? The optimal frequency is generally considered to be between four and seven impressions. Let’s take a look at how we know when we’ve actually achieved an effective frequency, and when it just looks like it.

Firstly, let us remember what the term “average frequency” means. This metric shows how many times an ad was shown to one user on average during a campaign (or over a period).

The calculation is as follows:
Average frequency = number of impressions

Let’s take an example from a model RTB campaign:
In total, we displayed 4,690,177 impressions during the campaign, while registering 1,035,462 unique cookies. Add these numbers to the formula.

  • Average frequency = 4,690,177 /1,035,462
  • Average frequency = 4.53

The value of the average frequency came out to be 4.53. This falls within the effective range of 4–7 above and seems fantastic at first sight. However, looking at the frequency distribution graph below, we see that we’ve actually achieved the average or similar values to it for only a few of the exposed cookies.

What is the ideal frequency in brand campaigns? Find out in our article.

In the case of this campaign, the effective frequency range (4–7 impressions) represents less than 15% of all unique cookies exposed. Over 71% of unique cookies had a display frequency of less than 4 and almost 14% had a higher frequency (8+).

Achieving the desired frequency and capping – bringing real effectiveness

To achieve effective frequency for the maximum number of unique cookies, let’s tweak the campaign settings:

    1. We will add a strategy focusing on cookies that have already received impressions but have not yet reached the effective frequency (at least 4 impressions) or the optimal frequency that we select. This optimization technique is called achieving the desired frequency.
    2. We set the maximum required frequency (called capping) for unique cookies (7 impressions at most).

As the next graph shows, the results fundamentally change thanks to optimization. In the case of this model RTB campaign, we managed to achieve an effective frequency range (4–7) for more than 61% of all exposed unique cookies. Only 30% of unique users had a lower frequency (less than 4 impressions), and even less than 9% had a higher frequency (8+).

What can we surmise from this?

For brand campaigns where you want to manage the value of the display frequency to one user, it’s a good idea to choose tools that allow you to work with this metric efficiently. In other words, don’t just settle for a reported average frequency value. At Seznam we are skilled at managing this for RTB campaigns.

Did this article interest you, or would you like to know more about the possibilities of the RTB campaigns we offer? Don’t hesitate to contact us via e-mail at rtb@firma.seznam.cz.

RTB team