What Are Known Audiences? Leveraging Data for Media Optimization in the New Age of Privacy

I. Introduction

Being an ecommerce marketer has become significantly more difficult recently. Whether it’s COVID, supply chain issues, or the introduction of stringent privacy laws like GDPR, CCPA, and the game-changing iOS 14.5 update, marketers need to be taking a more sophisticated approach than ever in order to evaluate the effectiveness and react accordingly. As a result of these changes, brands are increasingly focused on building and maintaining a direct relationship with their audiences, rather than relying on vendors like Facebook to determine who receives advertising.

If this is something you’ve been thinking about, we want you to understand three key things as you embark on this journey:

1. Audiences built using your own (first-party) data present a significant opportunity to engage with prospects and customers that YOU deem important to your brands.
2. Advertisers have an additional opportunity to directly evaluate campaigns built off this data.
3. Learnings can be applied to optimize your campaigns going forward.

II. First vs Third Party Data

As you begin to build out this strategy, it’s important to understand the differences between the categories of data. There are two primary types of data to leverage: first-party and third-party. For the sake of this article, we’ll refer to these as “known audiences” - they’re both groups of individuals you can specifically target but they have important differences.

First party data

• Represents all the data you directly collect from your audience, including current customers, prospects, and fans.
• Highly valuable given the relevance for your brand, filled with detailed customer information that you mine over time, enabling you to form closer, more personalized relationships with these individuals.
• High level of effort to build these first-party data sets which takes time and effort (often years) and you need to be intentional about scaling these data sets in order to maintain their quality. But the rewards you reap, in terms of insights and audience response, can be invaluable.

Third party data

• Data collected by parties who do not have a direct relationship with your end customer.
• Common sources of third party data are: non-profit organizations, academic institutions, government agencies and data “co-ops” who package this data up and sell it.
• Readily available and, depending on the size of your target audience, it is generally scalable, limited mostly by your willingness to acquire it. However, unlike your first party data, it is often more expensive and their value needs to be vetted through testing

First party data sets are best at driving loyalty, if repeat purchases are common, or for cross-selling if you have several relevant product categories. Given the direct relationship with your brand, these audiences are by definition relevant while third party audiences tend to be most effective when there is a strong fit for your product/service. A good example of this would be a diaper brand broadly targeting a national audience. Given that those without children aren’t in the market for diapers, this would be much less efficient than targeting a 3rd party audience consisting of individuals who have young families.

As a marketer, your job is to build a systematic program to grow the volume of high quality first party data while also efficiently taking advantage of all of the various third party data sets that exist. Critical to this is the implementation of a robust testing program so you can evaluate what’s really working as you constantly look to optimize the campaigns you’re activating with this data.

III. Known Audience Testing

As you evaluate the effectiveness of any media, it’s important to isolate the true incremental impact of media, separating out the conversions that would have occurred as a result of general brand awareness/loyalty, seasonal demand, promotions, etc. vs the conversions that would not have occurred had that media not run. While there are multiple ways to go about this from media mix modeling to multi touch attribution to site side analytics tools, at Measured we believe implementing experiment-based incrementality measurement is critical which can then help calibrate other modeled approaches you may be taking.

This can come in various forms (i.e. geo testing) but when your goal is to measure the effectiveness of known audiences, we recommend implementing a randomized control trial and evaluating the results based on an intent to treat analysis, which is considered the gold standard in measuring a causal relationship. This approach splits a population into “treatment” vs “control” groups and then evaluates the relative differences in conversion rates between the audiences. This is the same methodology used when evaluating effectiveness of drug trials and, while theoretically simple to implement, it requires a robust data engineering capability to avoid “contaminating” the results of the experiment. It’s important to build a reliable system that enables you to implement these tests efficiently, all while ensuring that the final results are actually statistically significant.

If this all feels overwhelming, Measured can help you build a systematic test and control testing platform when evaluating known audience campaigns, whether these are catalog, email, SMS or social campaigns on Facebook, Pinterest or Snapchat.

IV. Optimizing Campaigns

Once you have implemented a known audience testing program and are confident you’re accurately measuring outcomes, the next step is activating the insights. When implemented correctly, these tests should enable you to answer three specific questions as they relate to optimizing media strategy:

1. What is my overall incremental Return on Ad Spend (ROAS) or Cost per Order/Acquisition (CPO/CPA) of a tactic?
2. Which audience segments within my data set are most/least responsive?
3. What media strategy is most effective at driving conversions?

Think of the three questions as a pyramid of sophistication. Initially you want to understand the effectiveness of a tactic overall. Once you’ve established that baseline, you can then start to understand how segments within it perform. For example, in a recent Measured test, one of our clients uncovered that their audience tends to either repurchase within the first 12 months or not until 2-3 years later. Taking this information into account they were able to modify the targeting of their direct mail campaigns to exclude the cohort who last purchased 13-24 months ago, dramatically reducing wasted spend in their catalog program and significantly increasing the ROAS.

You can then take this type of testing further to optimize your campaigns and some examples of that are:

• Testing for alternatives to expensive direct mail campaigns (which can have CPMs upward of $100) by splitting an audience into different segments who are exposed to different contact strategies. i.e. direct mail only, email only, email + direct mail, email + Facebook, etc.
• Testing different offers in email to understand if the higher conversion rate you see from a discount really offsets the revenue lost as a result of offering it in the first place

The options here are nearly limitless but the important part is first implementing a systematic measurement program that allows you to confidently (and efficiently) answer these types of questions.

One thing we have often seen is a hesitancy to implement tests like the above on platforms like email due to their low relative cost and it’s not uncommon for us to hear some version of the following:

“Email is basically free so I don’t think it makes sense to test it because we won’t stop sending them.”

We believe this is a mistake. While there may not be a significant direct cost of running these campaigns, there is a meaningful opportunity cost of not optimizing them. Keep this in mind - if your email campaigns drive $5M in annual sales, then for every 1% improvement in efficiency you will see an additional $50,000 of revenue. In this realm, small changes lead to big outcomes.

It underscores the importance of treating media as a profit center, rather than a cost center. There is a limit to how much money you can save by cutting budgets, but there is no limit to how much more you can sell by more intelligently optimizing your campaigns. Measured exists to help you both identify wasteful spending AND identify new growth opportunities.

V. Conclusion

Every interaction with your audiences is an opportunity to learn, adapt, and improve your strategies. The keys to effectively doing so are: maximizing these interactions, collecting relevant data, and leveraging this information in a way that allows you to to quickly react and to build more effective, targeted campaigns.

We invite you to explore how a data-driven approach can transform your campaign management, optimize your tactics, and ultimately, boost your returns. Remember, understanding and effectively utilizing different types of data is critical as you start to focus on building a more direct relationship with your customer base.

At Measured, our mission is to unlock the power of incrementality measurement and optimization so every marketer knows where to invest media spend to drive maximum business outcomes. We spend every day thinking about incrementality and building a platform that allows our clients to leverage incrementality to their advantage. Specific to these first and third party data sets, our Known Audience Testing solution provides a robust framework to efficiently design and deploy statistically significant tests that allow you to unlock the hidden growth opportunities that are hiding in plain sight. Please feel free to get in touch if you think we can help!