What’s the frequency distribution, Kenneth?

I have some unsolicited advice for anyone working in marketing on the client side who’s responsible for oversight of programmatic media spend in any way: At the conclusion of every campaign, make your agency pull and present (1) a report that details a campaign’s reach against the total audience size used for the campaign, and (2) a frequency distribution report for the campaign alongside the frequency settings that were used to execute the buy.

You’ll learn a great deal about your agency’s competence and whether your dollars are being invested in ways that are aligned to the best effectiveness evidence that’s readily available from a number of reputable sources around LinkedIn and the Internet these days.

Unfortunately, most agencies aren’t especially good at executing programmatic buys. Instead of working to construct audience segments of the right kinds of people that are appropriately sized to the media budget, it’s common to see massively scaled, weakly- or un-targeted segments with no priority assigned to them. And instead of using well-reasoned minimum and maximum frequency settings, buyers regularly allow demand-side platforms to serve ads without much human guidance. Taken together, this too often results in a frequency of one exposure for 80+ percent of reached households for the entire campaign, which is typically four weeks or more.

In terms of both brand growth and advertising effectiveness, Binet, Sharp, and Ritson have consistently and compellingly in recent years demonstrated the need to prioritize reach, particularly among category buyers who don’t buy your brand; however, establishing and reinforcing brand links in consumers’ minds does require a level of repetition, so frequency shouldn’t be an afterthought.

Note that I’m not advocating audience segmentation schemes that resemble micro-targeting in any way. I’ve yet to see any credible research that would lead me to think it’s effective to chop a market into small segments in order to serve tailored creative to each of them. I’m sure someone can quote an exception or two, but as a rule, it’s best to use enough segmentation to help narrow delivery in a logical way without excluding people who will be in the market two or even three years in the future, budget allowing. Again, reaching people beyond your current customer base is the key to long-term brand growth.

Anyway, here are some (hopefully) useful guidelines to use as a check against poor programmatic planning:

1. For brand-focused campaigns, one ad per week is generally fine and for activation-focused campaigns, 2-3 ads per week is most effective.
2. For brand-focused campaigns, reach as many category buyers (including your brand buyers) as your budget will allow. Again, talk to everyone who buys the category if possible–Remember that most people don’t perceive huge differences between brands, so ad creative should be working to build awareness of distinct assets through emotional (i.e., System 1) appeals.
3. For activation-focused campaigns, target known brand buyers and people who are currently in-market to purchase. Here, creative may appeal more to reason/logic and convey helpful information about incentives, product attributes, and the like.
4. If you have data for your current buyers that can be used for marketing purposes, it can be employed to create lookalike (e.g., statistically similar) audience segments. By using information about your current buyers to help identify similar people who shop the category, your impressions can be better prioritized if you don’t have enough money to reach the entire category.
5. If you’re working to calculate a media budget during annual planning, make your agency partner size addressable media segments for your brand buyers and category buyers. Use these base figures in tandem with effective frequency levels to determine how many impressions you need to buy for a campaign and set the budget that way instead of using a historical or arbitrary figure.