How to Develop a First-Party Data Strategy

They always say the devil is in the details and that couldn’t be more true when we consider the business ramifications of a cookieless future! While it’s clear that wiping out the backbone of digital adtech will have sweeping consequences, exactly how one shores up their digital business can be a little daunting to even the most seasoned digital professional. The main industry messaging thus far has been a pat “develop a first-party data strategy,” but what does that actually mean? Let’s discuss!

what is data? where does it come from and where does it go?? these places!! (I guess…)

First-Party Data is Just Good Business

The first thing to cover is that a first-party data strategy is just good business, even without the depletion of third-party cookies. The Cheetah Digital blog eloquently lays out why; that is, first-party data is no cost compared to third-party, first-party data is more accurate than third-party, additionally first-party data is privacy-proof, and first-party data allows for authentic connections with your customers. Let’s dig into each a little more:

• The cost of third-party data adds up quickly! While third-party data rates range significantly, on the whole third-party data point pricing averages around ~$1.50 CPM, which isn’t insignificant. For a midsize publisher with a moderately active audience business, this can end up costing around $100k/year. Even though most publishers factor this into their rate cards, consider that publishers could just as easily charge the same (or more) for first-party data… without the third-party data costs at all. Additionally, when you consider the wastage, cleaning, etc, third-party data can actually end up even more costly than just the CPM implies
• The real catch of paying ~$1.50 CPM on third-party data is that, as most digital professionals know, third-party data isn’t actually all that good! Resonate blog does a great job of showing how the data is wrong and Digiday does a great job of explaining why. Basically, there isn’t really any incentive for any side of the equation to seek quality… and the way the free market works, if there isn’t an incentive for it… it generally doesn’t happen.
• First-party data should only ever be gathered with full transparency given to users and with the strictest privacy guidelines, inline with GDPR, CCPA, and other similar legislation. Therefore, first-party data done right is inherently privacy proof.
• “Authentic” connections with customers have of course always been the holy grail of marketing efforts and every year it seems like the public grows more cynical, so the authenticity stakes grow higher. The advantage here for first-party data is that the data has been volunteered by your users, so you know it’s accurate and real, and if gathered in a privacy-first manner, you also know your customer is comfortable with you using that data for personalization.

So now that I’ve convinced you a first-party data strategy is just good business, let’s talk about your options for developing a solid first-party data strategy.

magic hands make magic data

What Does the Industry Say?

I apologize for coming out of the gate with what might be bad news, but the first thing we need to talk about is the vital importance of a registered and logged-in user base. I understand this is a sensitive topic for publishers, as historically registering users has been a daunting task to undertake, but it is an absolutely crucial task because all other data you will be using for personalization and analysis will be appended to this user record. Even the New York Times, arguably the best poised publisher in terms of registrations, has had fears of being unable to scale a first-party data business!

I also apologize for what may be further bad news, but there just isn’t a silver bullet for getting your audience to register and log-in. The main tactic these days is to put content behind a registration wall, forcing users to register if they want to engage with your site’s content. This will work for committed readers, however more agnostic readers will still be a challenge. Your best bets here are to make your content as engaging as possible to make users want to register, provide a benefit to your users to reward them for registering, and, if you are lucky enough to have a data science team at your disposal, one of the team’s top priorities should be identifying patterns of behavior/triggers that lead to registrations, so as to be able to actively target those users to encourage them to commit to a registration.

Ok, so let’s assume you have a robust registered and logged in audience, what next? You have 5 main tactics you can take advantage of: mobile Ad IDs, shared IDs, second-party data deals, contextual targeting, and asking users to submit information. Let’s chat about each:

• Mobile Ad IDs: I’m going to casually estimate that somewhere around 70% of most publisher’s traffic is mobile these days, necessitating these mobile users be highly prioritized. One way to do so in the short term, is by collecting and activating against mobile Ad IDs (or MAIDs) as part of your first-party data user profiling. While in the (very) short term this will work, I caution against investing much in this strategy. Why is that? Well, mobile ad IDs themselves are likely to be deprecated around the same time as cookies. Analyst Eric Seufert has some thoughts on this. In any case, it will be fascinating to see what the actual future of mobile advertising will be.
• Shared IDs: Publishers have been working in consortiums for years to share pools of inventory and user IDs, so this technology is not new by any stretch of the imagination. What’s great about this tech is that while it currently functions off third-party cookies while they still exist, by nature they’ll be able to function off first-party cookies when third-party is no longer an option. The downside here is both that first-party cookies are also potentially on the chopping block, and that with the number of companies releasing shared ID technologies, the overall audience is likely to be fractured for some time as companies wrestle out who is a brief contender and which companies are here to stay.
• Second-Party Data Deals: so we’ve talked about first-party and third-party data, what about second-party? As a refresher, second-party data is some other company’s first-party data that they have put up for sale, either in a marketplace or through direct deals. While second-party data has all the advantages of first-party data, I’d say the main issue at this point with second-party data is the overall availability of it. The majority of it is recency/frequency browsing information, which can be used in a variety of creative ways of differing value, but will be unlikely to get you to the trickier-to-replace data sets (ie demographics, financial information, business information, political information, purchase behavior, psychographics, etc).
• Contextual targeting is considered to be the saving grace of a cookieless future, mainly because contextual does not depend on an individual user’s browser behavior. Instead it focuses squarely on the context of a publisher’s content and its suitability for any given ad. Additionally, while contextual targeting can be used very literally (ie targeting beauty ads to readers of the beauty section), with creative metadata deployment, many companies are leveraging contextual targeting (with select other signals, such as device used or timestamp of when content is accessed) in creative ways to determine user psychographics (ie please reference any publisher currently advertising an audience targeting set named something like “mindsets,” “attitudes,” or “motivations”).
• Asking users to submit information is the last, but far from least, option we’ll discuss here. Having users participate in the process of fleshing out their own user profile tends to be the most intimidating prospect for publishers developing a first-party data strategy, but this doesn’t need to be (and maybe shouldn’t be!!) the case. Most publishers fear attempting to provide users with a long form of personal questions to fill out, as they believe that such a practice will alienate users, which is true! But the answer isn’t not asking users to participate in filling out their profiles, rather the answer is in not deluging a user with questions. What to do instead? Progressive profiling. Instead of asking a user to submit a ton of information upfront that might either lack relevance or importance in the moment, ask the right question in the moment and save the information progressively in the user’s profile with the goal of having a full profile… some day. Keep in mind the wealth of creative ways to get a user to participate, for instance you can leverage polls, surveys, or in-app functionality to add to a user’s profile. It’s a long game you’re playing here, but in the end user volunteered information is the most valuable information you have.

So now that we’ve discussed what’s generally happening, let’s talk about your business specifically.

me, pondering all my possible data inputs

What Are You Currently Doing?

If you are currently leveraging third-party data, the first thing I would do is pull a report of impressions and revenue by as granular a data point as you can (this information will likely need to be pulled via your DMP instead of your ad server, unless you only ever run audience segments only comprised of single data points).

Now for what is the easiest, yet most time consuming, part of the entire process: take this report of third-party data points (and I warn there will likely be tens of thousands!!) and code each of the data points into easily digestible categories/subcategories. For example, if you see three data points: “gender: female”, “gender: woman”, and “gender: male”, those can all be combined into a “demographics: gender” category/subcategory.

After you’ve pulled your report and coded out the data to make it easier to look at, review the broad trends. What are your top categories in terms of impressions served/revenue accrued? These will need to be your priorities in terms of replacing third-party data with first-party. To figure out what to do with this new found knowledge, you’ll need to determine…

me tap, tap, tapping into resources

What Resources Are You Able to Tap Into?

Every company is a little different, so the available budget/resources are going to be different at each. That being said, one thing true at every publisher is that developing a solid first-party strategy requires executive level buy-in as executing the strategy is going to require a wealth of teams, from Revenue Operations all the way to Editorial.

The ideal set up for a company would be to have full use of: a CDP, a DMP, GA360, a flexible CMS, a dedicated ad product team, a dedicated data science team, full involvement from editorial, a wealth of content both in terms of depth and breadth, and a substantial number of logged in users.

So if that’s where you want to be, where are you currently at? I’d recommend next doing an audit of your adtech/martech stack to look at what platforms you have available for use. To start with: does your company employ either a CDP or DMP? Without either of those, moving forward will be very difficult as you won’t have a great way of storing/managing/activating data. From there you can consider what other systems/data you can plug in. Does your site offer paid subscriptions? If so, a payment processor like Stripe could be in play. Do you only have access to GA360 and DFP? Well, great! Age, gender, and browsing behavior is still accessible to you. Take a lay of the land and put on your most creative thinking hat to think about all the different imaginative ways you can leverage the data at hand. Don’t sleep on the metadata tagging of your site’s content either, as in the end it may turn out that’s actually the best weapon in the average website’s arsenal.

me, deciding what data listener to evaluate next

What to Do Next?

We’ve covered the value of a first-party data strategy, the industry take, as well as how to investigate your specific situation. The only other relevant topic we haven’t discussed yet is the timeline, and the time is NOW!!

Safari started deprecating third-party cookies in 2017, Firefox started blocking third-party cookies in 2019, and Chrome has stated it will remove third-party cookies sometime around January 2022… which is all to say, that as we are nearly a year away from this cookie apocalypse, there is no time to waste! While developing a viable first-party strategy and getting full company buy-in will be time consuming in itself, it’s important to remember that not a single one of the strategies listed here is retroactive. Meaning: you only start gathering data post implementation and, considering scale is a crucial factor here, the more time you have to gather data and profiles, the greater your potential scale will be.

For those about to rock on this task: I salute you. It’s no small accomplishment, though if it’s any consolation, pretty much all other publishers are currently going through the same process at the same time. If you’re intimidated, I recommend just taking a deep breath, reviewing best practices, and getting to work! Remember you’re doing noble work as this isn’t just for your audience targeting business, but rather to benefit the business as a whole. Tealium projects 5 company wide benefits for this project:

• Getting your “data house” in order to comply with current and forthcoming privacy regulations
• Overcoming the devaluation of third-party cookies thanks to changes to browsers
• Creating a sustainable and constantly improving customer database
• Setting the stage for customer experience insights from machine learning
• Delivering a consistent customer experience across the entire organization.

If all of this so far has been a refresher, here are some more advanced tactics to optimize your first-party data strategy.

Either way if you’re struggling with the process, I’m here to help. Please reach out for a data strategy consultation today!