The Data Gap for 1st-Party Data and UIDs

First-party cookies, or some form of Universal ID, will be the end of the third-party cookie deprecation problem. Right? Is that not what dozens of companies are telling us? Well, it IS what they are telling us. Today I want to discuss why that is a half-truth at best. This is one of those topics that really frustrates me in how narrowly people are thinking about the ripples from cookie deprecation. To start, we need to discuss what a Universal ID is and how it relates to first- and third-party cookies.

First-Party Data and Universal IDs, Defined

First-party data are those data collected from a primary source — your website, data that say Experian or Equifax owns, registration data, and the like. They are, in one regard, essentially 1:1 data. Okay, hold that in your mind and let’s think about the idea of a Universal ID (let’s call them UIDs). A UID is an anchor that you can attach several of these first-party data sources to — it’s called “appending data.” Most of the time they are attached at the household level (called HHs, or sometimes called zip-11). So, a good UID can take in many of these sources and create a really robust picture of the individual or HH.

Good so far? Good. This one is a little nerdy, I know — sorry about that — but stay with me because, depending on what you are selling, this could be one of the biggest impacts to your advertising that has ever come across your desk.

Going back to the picture of a UID, I want to share an example of how clear this picture can get. Meet Josh Goodin (though PII [personally identifiable information] is usually obscured in one of a few different methods). He identifies as: sex – male, gender – man, age – 40-something, race – Caucasian/Choctaw, disability – yes, veteran – yes, income – $XXXX, credit score – XXX, total debt – $XXX, debt types – home loan, student loans, auto loan, married – yes, children – yes, ages – 16, 12, 12, address – 123 Blah Blah BLVD, Fairhope, AL 36532, homeowner – yes, XXXX square foot, paid $XXX for the property in 2019, vehicle owner – yes, garage – VW, Honda, Lexus.

Okay, let’s stop here — this is plenty for our illustration. Wow, right? That is a lot of useful data for marketers. If you have your own data to attach to this, it gets even better. First-party data can be wildly useful. As an example, let’s say you are Johnston & Murphy, and you know that Josh buys a new pair of dress shoes/boots every Fall. It’s Fall, and we have his physical address and his e-mail, so we send him the Fall catalogue and we send him an e-mail showing off our new shoes for the season. Useful. Also, if you are targeting just at the demographic layer, it is super useful as well — e.g. people who make more than X, live in X area, are parents, between ages 25–54, etc. These data will certainly help marketers reach their target markets, but there is a key aspect that has been left out.

The Piece Everyone Leaves Out: Intent Data

What has been left out, you ask? Intent data. Intent data are those that tell you what someone is doing in the moment, unlike UIDs and first-party cookies which are fixed data sets about the individual. Reread the data points above. Which ones tell you that Josh is in the market for a new car in the next six months? How about that he is planning a vacation to Key West this summer; planning to buy a new high-end watch soon; looking for a new retirement planner; that he is unhappy with his current cell service? And that list can go on and on.

Here’s an example: It is clear from the demo-based data that Last Key Realty in Key West could identify Josh as a target — right income, lifestyle, interests, etc. What those data do not convey is that he is actually planning to go to Key West this summer. How much more valuable are Last Key’s ad dollars that are spent on the specific intent of Josh to go to Key West vs. generically someone who might think about Key West as a vacation spot? Last Key is a small vacation provider and cannot afford to just target everyone who is a vacationer. Intent signals are what tell you that someone is going to take specific action on a purchase; intent signals are what we will mostly lose in the deprecation of third-party cookies.

Who Gets Hurt the Most

So, you can start to see our problem, right? You can see that without third-party cookies, our ability to hyper-focus our ad dollars will be drastically impaired. It is not true for every industry, and it is not true even for all retailers in a given industry, but for some the impact could mean it is between two and ten times more expensive to reach the same number of purchase intenders. Industries that come to mind are Travel, FinServ, Auto, and High-End Retail, to name just a few. I believe that most organizations are not talking about this coming loss because they don’t understand it completely. Making that understanding gap wider are the UID purveyors who are basically shouting “Look here, look here!” pointing at their ID so as to capture business. And because they can’t do anything about the loss of the intent data, they do not want you focused on that part.

I will say that I have come across some UIDs that I like; that I think are going to be very valuable post-cookie. But you must understand your business and media buying, and think about what the impact to your own ability to reach consumers will be. For the agency people reading this, really think about your clients and how to broach this conversation with them. I know that it is not an easy conversation, but the harder conversation is explaining a 4X increase in cost per conversion overnight when we lose cookies.

The 2.5 Ways to Deal With It

The picture is probably bleak enough now, right? Maybe you are wondering what we can do about it? There are 2.5 paths we can go down. Let me explain.

Path 1 — Single Sign-On (SSO). SSO is where a user logs into a website and that website uses the data to connect the user’s actions now to their previous activity. Retail benefits from this greatly. Here is an example: I like Banana Republic button-downs, so every time I go to their site I log in with e-mail and password (SSO). That tells BR who I am and allows them to bring up data I have given them in the past like size, address, credit card info, etc. By logging in, I have given my tacit permission to be tracked. Gap Inc. owns BR, Gap, and Old Navy too, and my data will persist through that entire network as I navigate around it. This is useful for sure, especially for customers who are “regulars,” but again, it only helps some business types. For some industries, like say Auto, it will be a very difficult path to get login data.

Path 2 — Walled-garden publishers. Next, there is one other place to get that intent data: a type of media provider that retains intent data — publishers. Publishers can be a single site like Bloomberg or Runner’s World, or they can be a group of sites under one umbrella like Google or Yahoo (these are usually called providers vs. publishers). The larger collection of sites is what is valuable here. These collections of sites are often called walled gardens. A walled garden is simply a publisher or provider who owns a lot of sites and collects data about its users from those sites. Some walled garden examples are Google, Meta, Pinterest, Yahoo!, etc. Using Meta as an example: You are logged in and moving between Facebook, Instagram, WhatsApp — Meta can see all that you do. They can use that data to class you as a type of audience — maybe you follow hiking trails and hikers all over the country. With that info Meta can class you as someone with a propensity for outdoor activities. See how it works?

There is a limitation to the usefulness of this data, however, and it goes back to the name “walled gardens.” The thriving “garden of data” inside these publishers is easy to see, but the limitation is the “wall.” What “wall” means is that data cannot transfer between the different owners of the gardens. You might know that Josh Goodin is into outdoor activities from Meta, but you can’t connect that person to the Josh on Yahoo! You can’t follow him through the hundreds of sites in the Yahoo! network serving him ads for a new backpack that he needs if you are The North Face. This leads you to a place where you really need to be able to engage with all the walled gardens to buy media that is taking advantage of the intent data. It also means that you will duplicate advertising to individuals because you can’t tell it is the same person. BUT duplication is a lot better than no intent data at all.

The Google Problem

One thing to take away from this walled garden talk: If you or your agency is relying exclusively, or almost exclusively, on Google, you probably are going to have some trouble in the future. Let me explain. Think about the way I just described the walls of the gardens — Google is a walled garden and Google has a mountain of data, oceans of the stuff, but Google is shockingly top heavy. That mountain of data is like a funnel — in other words, the utilization of that data is almost non-existent. What does that even mean? Well, let’s think about Google’s opportunity to activate data in the context of their own garden: Google.com, YouTube.com, Google Discovery, Google Maps. That is the whole list of places you can run ads inside the Google garden. Now, you can buy other sites THROUGH Google, but because those are not owned by Google, you cannot push Google’s data into those sites. Worse, I would argue that the need for intent data is far less on Google.com, as generally the user is declaring their in-the-moment interest by what they type into the search bar. To restate: If you or your agency is relying on Google tools to do all your media buying post-cookie, and you benefit from intent data, you have a serious data shortfall inbound.

The Half-Path: Unified Garden Access

I said earlier that there were 2.5 ways to deal with the need for intent data. What is the .5? Well, there are a very, very few companies working to give you access to all the gardens through a single point of access. It DOES NOT solve for the duplication problem, but it does mean that you can use all the gardens’ data inside the respective gardens. It requires more than just being able to buy media inventory on other sites — it requires a technical integration inside each garden. It is only .5 because it is the same thing as buying from all the gardens individually; it’s just giving you a way to do it more centrally. You could do it yourself by making and managing a couple of dozen media buys with each of the publishers and providers. The companies working on this access are just going to make it a lot easier for you. I am aware of one that will allow for AI optimization across the providers relative to the performance of media too, so maybe that one company gets a .75 instead of a .5.

What You Should Do

This was a hard article, I know. This cookie thing is not going to be fun, but arming yourself with information can help you navigate this complicated topic. My hope is that it gives you the info you need to push on your vendors. Make them tell you exactly what they are doing about these problems. I say all the time that it is too late for the answer to be “we are working on it.” What? If they say that, find a new partner.

If you have read this and you have questions, feel free to reach out. More coming soon — thanks for reading.