Time is Running Out: Get Your Owned Data Strategy Figured Out
If publishers don’t know it already, they’re going to realize it more and more as we move closer to Google’s removal of third party cookies from its Chrome browser:
Owned data about your audience is the only path forward for publishers.
We cant rely on third party sources for user data. It’s why many publishers are quietly panicking about Google’s change. I imagine some publishers are holding out hope for the Privacy Sandbox to come up with some innovative way of providing all the same benefits of 3rd party data while still respecting user privacy.
I’m skeptical of that approach and I think it’s not smart to wait for Google to come up with a solution. Additionally, any of the solutions that Google is pushing forward are likely only going to work with Chrome. Firefox and Safari would still need to buy in, which means publishers are still seeing diminished results across those browsers.
Publishers need to take ownership over their data and start figuring out how they are going to get the information they need about their audience. The fact that they ever gave up that responsibility just baffles me.
Brian Morrisey over at Digiday said it best in my opinion:
I always said, I call it the original sin of internet media was separating the audience data from the media impression because it automatically, it inevitably commoditized ad inventory. When you can just chase a cookie around the internet, the pendulum swung too far to simply the audience data versus the media environment.
So, let’s agree that the next 3-6 months needs to be about coming up with a data strategy. I want to walk through the two categories of data that you should be collecting to help ground that conversation.
- Zero Party: A relatively new term that some marketers are throwing around that means data that the individual has explicitly given you.
- First Party: This is data that we can gather about people even without them explicitly telling us.
We’re going to want to collect both.
It starts with email
But first, let’s be very clear…
All of this starts with an email address. That needs to be the central piece of data that you then build an ever growing profile around. You can collect all sorts of data, but if you don’t have that email address, it’s irrelevant. Here’s why…
Data for the sake of data is pretty useless. However, if you can use that data for communication and promotion, it becomes valuable. The email address is a constant that, once you have it, is highly unlikely to change.
Ask yourself… When was the last time you changed email addresses? If it’s your work one, it’s when you got a new job or started your company. If it’s personal, it likely hasn’t changed for years.
That’s powerful for a publisher and makes it possible to grow with your reader as their needs and interests evolve over time.
The other reason we use the email address as the hub of data collection is because it doesn’t matter what device you’re on. One of the deficiencies of cookies is they’re device specific. Therefore, if I am reading something on my computer and then read it on my phone, it’s hard to conclude that the person is, in fact, the same person.
Because people’s email addresses don’t change and they typically use it on their computers and phones, that can act as the means of connecting all their consumption.
Whatever strategy you come up with must start with how you can get as many people as possible to give you their email address.
Since we’re talking about people giving you their email addresses, it makes sense to start the conversation with the data that they may explicitly giving you.
The goal here is to have people give you specific pieces of information about themselves. They’re making a choice.
I find that there are two primary tactics for getting people to give you information. You either offer them a newsletter that they want to sign up for or they register. If you’ve been reading for a while now, you’ll know I’ve been advocating for publishers to build out registration for close to a year now.
The goal of zero party data is to understand core information about them. Who are they? Where are they? What do they do? Why are they here on the site? And then any additional qualifying information within those categories. Here are a few examples:
- A Media Operator (in an ideal world): First name, last name, company, media company type, job level and function
- Financial Website: First name, last name, individual or professional investor, company name and job function(if professional), stocks in portfolio (individual) or AUM (professional), etc.
- Travel Website: First name, places they’ve traveled, preferred airline & hotel, do they travel for personal or business (you could dig into professional questions if so), etc.
The list goes on… The type of publication that you’re running dictates the data you are asking for. A travel website, for example, might care less about professional information (unless it has a large business travel audience).
I don’t believe that media companies should be collecting as much data as humanly possible. Instead, I believe that we should be collecting the right data. Therefore, a travel website asking for assets under management is probably not really necessary.
A very critical part of this exercise… This data needs to be structured. That means you should be avoiding text bosses as much as humanly possible. Why? The variations in data can make segmenting and building cohorts far more difficult. Subtle changes to titles would make an otherwise common title become unusable in your database.
One topic of much discussion is how to get the data. There are two schools of thought.
The first is the long form. We’ve all seen it. You go to register for something and they ask for all the data that they want from you.
The second is progressive profiling. This is where you ask for as little as possible, but with multiple touch points. For a newsletter sign up, maybe it’s name and email. For a premium piece of content, maybe it’s another two fields. For a webinar, it’s another two data fields.
Both approaches have their pros and cons.
With the long form, you are getting all the data up front. It’s really great to get complete profiles from every person. On the other hand, the common belief is that the more form fields there are, the lower the conversion. But even that has come up for debate. Some people who work in conversion optimization have said that length isn’t a problem so long as you test which fields are causing a drop off.
With progressive profiling, you aren’t getting caught by the long form field problem, so you can theoretically get more people engaging. However, this strategy requires that you convince the user to come back for more bite size chunks multiple times. If they don’t, you’ve got incomplete profiles on them.
Honestly, I think the best approach is to use both. A simple tactic could be to collect the email address for a newsletter and then redirect the user to a full form. Most people likely won’t fill it all in, but if 10% do, that’s still something. Industry Dive does this. They are guaranteed to get a newsletter subscriber and then also some additional data from people that decide to fill it in. Then you can start progressively profiling the rest of the people so you’re capturing data over time.
When I say first party data, what I’m specifically saying is contextual data. We want to know what the user is actually engaging with on the site.
Registration is incredibly important for this strategy. As the publisher, you want to be tracking every place a user visits on the site. But tracking an anonymous visitor doesn’t help you do anything with that data, hence why we started with the zero party data capture.
However, before you can start with this exercise, you need to get your structured data set in order once again.
The straight forward way of building this out is to rely on your site’s tagging structure. You’re likely already tagging each story, so it’s not hard to extend that data into your user profile.
The downside with this is that top level tagging can miss nuance. For example, if it’s a large finance site, it might use a keyword like “stocks.” There’s some context there, but it doesn’t tell you what kind of stocks.
That’s why it’s really important that your tagging structure is well defined and that you can add to it over time.
Once this is done, you’ll now start appending specific tags to your user profile. If you ask any marketer, this can be a gold mine and really help improve performance.
How you can use it
If you’ve done all of this, what you are left with is a user profile that has information about who they are (zero party) and about what they want on your site (first party).
From here, there are a variety of ways to use this information in a way that meaningfully impacts your business.
The simplest is with your marketing. With all this structured data, your marketers can create segments on any and all of the data points you’ve collected. You could send targeted campaigns to people with a specific job function that read certain topics. The list is small, but your messaging is focused.
This is table stakes and we should all be striving to get to this point in our organizations. It’s critical.
The Holy Grail of this blended zero party/first party data strategy is your ability to deliver deep insights to your advertisers. Part of why advertisers have loved 3rd party cookie programmatic advertising is because they can target specific people no matter where they are on the internet.
But as we move toward a world of contextual targeting, layering in data about the people that saw the advertiser’s message is invaluable. Now the conversation isn’t just, “we delivered your campaign to people that care about X.” The conversation becomes, “we delivered your campaign to Y people that care about X.”
The addition of “Y people” into that sentence makes you that much more important to the advertiser.
B2B publishers have been doing this for years now because they’ve been required to from their advertisers. Consumer publications are going to need to learn how to connect the two sides to succeed.
If you really want to make your head spin… Take the first party contextual data, layer on zero party personalized data and then add an additional layer of data from your revenue server that tells you how that person should be monetized.