Frequency in digital media serves as both a tactic and a means of measure. The media community has long believed in the concept of the “Rule of 3”*. Some define this as 3 exposures required before a user takes action or remembers a brand. Others define this as a best practice where any more than 3-4 exposure per day becomes too intrusive to users. Traditionally, in digital media, frequency is measured as a number of impressions delivered to a user usually within a 24-hour period.
Less is More
One school of thought in the media community is the concept of “less is more”. This segment of the industry (largely subscribers to the “Rule of 3”) believes that a strong impact can be generated through stricter frequency caps, and that more aggressive caps do nothing more than produce wasted spend and eventual immunity to the ad message. Promoting the perspective that frequency capping should exist centers around two central ideas: Intrusiveness and effectiveness (diminishing returns).
Privacy is a concern in online usage and specifically, online advertising. While advertisers appreciate the measurement and targeting capabilities that can help in reaching a very specific audience in the digital sphere, users do not want to feel as if they are being watched or followed, or that any of the personal data (including browsing habits) is being utilized commercially. Therefore, frequency capping is a tactic that can decrease the likelihood that a brand will be perceived as overly aggressive or outright “big brother-ish”.
There is also data that suggests that repeated exposure to a banner, past a certain amount of views, won’t serve to drive increase in conversion rate or awareness, but rather causes users to become immune to the repeated messaging. Therefore, the practice of frequency capping can help to limit the amount of exposures a user is subjected to in a given day.
More is More
On the other hand, there is another segment of the industry that subscribes to a belief and supporting data to suggest that the “correct” frequency may actually be higher, or perhaps capping should not be employed. Such considerations revolve around goals of brand awareness.
A 2012 study performed by Dynamic Logic and DoubleClick called “Better Brand Engagement with Display Advertising” points to the fact that users who were exposed to more than 2 impressions per day had stronger brand recall than users who had more limited exposure.
Considerations for Frequency Cap Determination
There are many factors that can influence whether advertisers choose to cap frequency in online display advertising, and if so, at what rate. At Geary LSF, we tend to subscribe to the “Rule of 3” as a general best practice and starting point, however, we employ testing to define the true effective number for each client and buy type. Below are the primary considerations that should be factored in when determining frequency capping practice and rate:
Unit Type: Certain classifications of online display are, by nature, already somewhat intrusive. Therefore, frequency caps should be a very real consideration for advertisers if engaging in remarketing strategies, utilizing pop-ups, interstitials, pop-unders or any form of “screen takeover” units. Providing countless exposures to users for these types of buys could quickly serve to promote a negative brand image – one of intrusion. On the other hand, more traditional banners in buys that don’t necessarily guarantee “above the fold” page positioning may benefit from a higher frequency cap to capture user attention (particularly as frequency cap metrics are often based on impressions served, and not impressions viewed.)
Buy Type: The pricing model can also dictate the need and rate of frequency capping. For example, in a CPM cost model, advertisers are charged on impressions served. Therefore, with higher or no frequency caps, advertisers are likely to be paying for multiple impressions to the same user, who after a time may become immune to seeing the same message. Whereas, if the media is being purchased on a CPA or CPC model, then frequency considerations may not be as prominent, as ultimately, the exposures themselves are “free” – the advertiser is charged only when someone clicks or converts.
Buy Scale: Another consideration to frequency comes in the form of how much inventory is being purchased, and across how many networks. An advertiser who has a somewhat broad target audience, and is purchasing display on one or two networks can stand to be less concerned about ensuring certain frequency caps, as they are not likely to be reaching the same users over and over. Whereas an advertiser who is engaging across multiple ad networks with overlap, may be more prone towards a conservative frequency cap to reduce “wasted” spend.
Program Goal: Studies such as the DoubleClick/Dynamic Logic one referenced above have shown that brand recall is higher based on a higher amount of ad exposures. Therefore, a campaign that is trying to achieve brand awareness should consider less restrictive frequency caps. Direct response advertisers, on the other hand, are looking to drive a certain action or conversion, usually at a target cost per. Therefore, in these instances, repeated serving of impressions to the same users who don’t choose to interact drives inefficient spend. For DR purposes, we recommend an initial testing period where data can be collected to determine what frequency is most likely to produce a conversion rate, and at what point, a diminishing return is reached.
Creative Assets: A final core consideration for determining frequency settings should center around messaging strategies and volume of available creative assets in the rotation. If, for example, an advertiser has only one or two banner concepts in rotation, a lower frequency cap may be in order so that the messaging doesn’t go stale quickly. Whereas advertisers with multiple messages and assets in rotation (particularly if testing creative) may choose to relax their frequency caps under the theory that even if users are exposed to their ads multiple times, those ads have a fresh message or different look and feel each time.
Technology is redefining how we think about frequency
The display space is constantly evolving, and therefore, I project that in the very near future, frequency metrics will be redefined, or potentially removed from digital display as a consideration.
Real time bidding (RTB) in display is a recently new technology that is the foundation of DSP’s (Demand Side Platforms), and is continually being adopted by various ad networks and exchanges. RTB functions off the concept of the right price, for the right impression, to the right user, at the right time. In this auction based environment model, each potential impression is given a value in the form of a bid based on how likely it is to produce the desired conversion or action. The advertisers with the highest bids (meaning highest value for that impression) are given the exposure. Under this concept, the technologies will collect data to determine things like what time of day produces better conversions, what composition of demographics is more likely to convert, what sort of content or behavior drives propensity to act, and also, as related to frequency, which exposure or whether new or returning visits are likely to convert. Theoretically, RTB can do away with the need to try and define the “right” frequency for display buys, as it will, based on data, adjust the value for impressions or frequencies that are not likely to produce a desired action.
Additionally, many traditional ad servers can allow advertisers to define sets of rules for when ads should show, and/or what message should be delivered. This includes the ability to show a different message to a repeat viewer, or after a certain amount of exposures, cease showing the ad to someone who’s been exposed but has not interacted.
Historically, frequency has been calculated off served impressions, in which no one can really be sure whether users actually “saw” the ads. For example, an impression is counted if the ad on the page loads, however, that ad may be below the fold of the page, and the user never ended up scrolling down. More recently, the idea of “viewed” impressions has been introduced, and is starting to gain traction among some networks and ad servers. This concept relates to using an added metric based on the likelihood that an ad was actually seen by a user. Therefore, if the ad was served below the fold, and we know the user never scrolled down the page, it would not be counted as a viewed impression. Therefore, it is highly possible that as viewed impressions become a standard measurement of display, frequency will be redefined based on who saw the ad, and not just that they were served one.
What is the ideal frequency cap to employ for online display advertising?
The reality is that there is not one global answer as to what the ideal frequency for online display should be. We can utilize best practices and widely held beliefs as directional guidance, however, the type of media and how it’s bought can certainly influence the “right” approach to frequency. So like with many other tactics and strategies in digital media, testing, measuring, and then determining is the best answer for any advertiser, as it will allow them to arrive at the means of what is truly right for them and their buy types.
Footnotes:
* Rule of 3 background: http://mediaplanninginstitute.com/basic-knowledge/effective-frequency/