In Digital Advertising, Focus On Reaching People Over Devices

When it comes to digital media activation and performance, you may be focusing on skewed metrics. But it’s not because you are doing things wrong – it’s a byproduct of how the ad ecosystem is structured. 

For a long time, the only way to get an audience into an advertising platform was through an onboarder. Because of the way these legacy solutions approach data matching, they were built to report on the number of cookies, mobile ad IDs (MAIDs) or other device IDs “matched” to a given user. It’s important to note that this does not mean the matched device IDs are active, it just means there was an association made at some point. 

As different device IDs came to market in the media ecosystem, advertising teams became too focused on maximizing reach over data quality. As a result, onboarding solutions optimized toward larger and larger device pools, which led to a dilution in the quality of the data onboarded into advertising platforms. 

The good news is there are now ways to optimize sending data, resulting in improvements to targeting efficiency and precision. This, in turn, lets you focus on the number of real, addressable customers in your audience, as opposed to the number of devices that may or may not accurately represent them.

We’re talking about a paradigm shift toward better quality audience targeting so that your targeting actually matters, rather than pursuing a big reach number that may not actually represent reaching people. This shift leads to both stronger business results and improved customer experiences in paid channels. 

Reaching devices (match rate) vs. people (addressability)

Let’s start with the metrics that represent the two different approaches to understanding audience reach and quality. 

Match Rate (Device-based): The total number of online IDs that can be matched to offline customer profiles. The online IDs can include third-party cookies, CTV IDS and MAIDs from the past 90+ days. The matched IDs at this stage do not necessarily mean they are usable by the advertising platforms. It simply means there was an association between a person and an online ID made by the onboarding partner. The process goes like this: 

1. Onboarder matches your first-party data to their third-party graph, usually denoted as a pseudonymous ID

2. The onboarder then connects device identifiers found in the advertising ecosystem to the pseudonymous ID 

3. Once the links are known, they then distribute the full list of device IDs, without deduplicating, to the ad platforms 

The last step is critical to remember because it demonstrates how a single customer can be mapped to several IDs, leading them to be counted multiple times as successful matches without clarity on how this match is happening.

In contrast, direct activation from a first-party data management platform takes advantage of new data supply chain efficiencies and opens up a transparent approach to calculating a truer addressable audience metric:

Addressable Audience (Person-based): A group of people that can be targeted with marketing or advertising outreach in a particular channel or platform. Durable identifiers, such as hashed-emails or authenticated IDs, associated with each customer mean that when they are matched in an ad environment, they are addressable in that platform’s advertising bid stream.  Advertisers gain transparency in this approach.

Crucially, over-indexing on match rate and failing to pay attention to audience addressability means that advertisers are not getting a clear picture of whether they are actually reaching more people.

And, important note, Match Rate and Addressability are the most common terms for these concepts, but just to keep us all on our toes in an already-confusing ad landscape, some platforms use different terminology. It’s always a good idea to look into what the metrics in th eplatform you're using actually refer to. 

So – let’s see how this plays out in practice.

Example: Higher quality first-party data foundation beats higher device count

In our work with Amperity users, we often run tests by delivering the same audience into an ad environment like Google, Meta, or The Trade Desk via two different paths to see which leads to better outcomes:

1. Amperity (first-party data) → Onboarder with cookie-based ID graph (e.g. LiveRamp, Hightouch Match Booster, etc)  → Ad Platforms

2. Amperity → Ad Platforms

Here are some typical results, directionally based on work with multiple Amperity clients across verticals:

There are two numbers related to matching in the ad environment, Addressability highlighted in yellow, and Match Rate highlighted in teal. 

As noted above, Addressability represents the percentage of customer records uploaded into the ad environment that successfully connected to an active user (not devices!) in their ID graph. Result:

• Percentage of records that reached an active user when passing through onboarder: 55%

• Percentage of records that reached an active user when served directly from Amperity: 100%

Match Rate represents how many devices (not people!) from the uploaded audience can be served an ad in the respective inventory channel (i.e. Search ad vs. YouTube ad, etc). The percentage is derived by dividing the number of devices matched by the total size of the initial segment – so, for example, for Search, 10MM / 12MM = 83% Match Rate through an onboarder and 9MM / 12MM = 75% Match Rate through direct activation. 

Remember, though, that onboarders distribute device ID lists without deduplication, which means a single customer can be mapped to several IDs and thus double-counted.   

These outcomes reflect what we typically see: going through an onboarder results in a higher device count, but a much lower quality within the ad environment. This leads to two conclusions:

• The onboarder ID graph has variable quality

• A higher rate of matching devices does not mean more people are seeing your ads

The takeaway is advertisers must expand their definition of what it means to reach an audience.  It is critical to move beyond device counts and account for key considerations like: 

1. Quality of audience data in ad platforms 

2. Transparency in the matching process 

3. Time to activation 

4. Performance of the data uploaded by the activation / onboarding partner  

Functional differences between device-based & person-based

As the example above shows, there are two distinct approaches to measuring the quality and actual reach potential of your target audiences. Let’s go a bit deeper into the differences between orienting around devices vs people: 

So why distinguish between the two?

We often find our brand advertiser clients emphasizing the importance of match rates without fully appreciating that onboarders are using device-based match rates as their primary metric. It means advertising teams are not getting the full picture and instead relying on a black box approach to mapping valuable first-party data to associated devices. 

Because of the 90-day lookback, device-based match rates can be deceivingly high, but still fail to deliver on a brand advertiser’s reach and performance objectives. While device match rates shouldn’t be fully dismissed, this match rate alone is a poor indicator for how your campaign is likely to perform.

Also, device-based metrics depend on third-party data like cookies and MAIDs, which means they are becoming less effective and more subject to consumer privacy regulation. Currently, 40% of cookies are relevant for less than a day, and another 15% last for less than 30 days, with the reliability of these identifiers continuing to dwindle. As the ad ecosystem evolves, it’s essential to anchor around metrics based on first-party data, both for its superior accuracy and reliability, as well as its inherent compliance with heightened privacy regulations.

Anchoring around person-based metrics is part of a bigger trend

Setting your media activation and measurement goals around reaching people rather than devices fits into the broader pattern of how brands can evolve with the new ad ecosystem. This is one more facet of the shift from third-party to first-party data strategies. Successful person-based addressability depends on having a high-quality first-party data asset, which at this point should be a key part of any organization’s ad tech approach as third-party data infrastructure continues to deteriorate. 

It used to be the case that your ability to reach people rather than devices was limited, because first-party data was hard to work with and there was no choice but to go through onboarders. That’s no longer the case. Advancements in customer data tools for media and ad tech now make it possible for teams to bypass onboarders, leading to far more control while also getting better results. Winning in paid media today means owning your data instead of renting it, really understanding the quality of your audience data foundation, and streamlining the path between your data foundation and ad platforms.

Improvements in measurement are a part of this too – stay tuned for an upcoming article about centralizing and owning ad performance measurement. 

But even aside from how person-based metrics are part of where digital advertising is headed, it just makes more sense – why go after devices when you can reach real live people? Bots and agents may be so hot right now, but the future is still all about human customers.

Check out our on-demand webinar about evolving paid media and retail media operations.