Since Google’s latest U-turn, publishers have followed one of two schools of thought: Make use of the first-party data infrastructure they’ve already invested huge amounts of time and resources in building. Or, as we’re increasingly seeing, stitch together deterministic and probabilistic data to target their audience.
But there’s another way. Publishers can create their own bespoke privacy-compliant identity graph, using only deterministic data, to unlock a whole new level of targeting, audience engagement and revenue potential.
So what are the publisher benefits of building an identity graph?
1. Data enrichment: understand your users to drive engagement
Email, web, mobile, apps - you name it, publishers are facing an ever-growing range of channels through which to reach their audience. They’re at a crossroads where they need to move beyond a reliance on publisher-focused data in order to know more about their readers.
The versatility of an identity graph means publishers can connect every type of user demographic, preference, and behavior - accessing hundreds of millions of signals across a fully-consented customer database.
So whether they need address-level data for mailing initiatives, email addresses for nurture campaigns, or mobile numbers for telesales outreach, they can start to build a full device graph complete with cookies, HEMS, MAIDs and IP addresses for different targeting mechanisms.
Surprisingly, many publishers still can’t tell you who their best customers are across their properties. An identity graph can show who is consuming what content, which in turn informs future editorial and/or advertising strategies. One of the best ways to build a picture of an audience is to take the identity graph and overlay data from a behavior graph to see who the most engaged customers are, including what that audience cares about outside of the publisher's property.
2. Resolution of unknown to known: create new audience segments
The ability to identify the total addressable audience, and create new segments, is a foundational component in resolving unknown visitors. After all, why have an identity graph to learn more about your customers if you are not also resolving your unknown visitors to known?
The key value proposition here is that publishers have the ability to identify and resolve these users at a significant rate, with most publishers resolving consistently between the 47% and 58% range. For one publisher, having an identity graph helped resolve 47% of 100 million unique users in the first month alone. That's 47 million users they now know have visited their website, which allows them to create a whole range of new and interesting segments to target.
What’s more, by using a solution that offers identity resolution on top of data enrichment (with no revenue share set against it), the more a publisher can monetize these new audience segments, the higher the ROI.
3. Audience growth: leverage your data to extend your audience
Once you’ve built a picture of how your readers interact with your site, and which new segments to create, you can leverage this insight to model look-a-like user profiles. This enables you to identify, target, and acquire new customers that look and behave like your best customers.
Again, not only can publishers see what their customers are consuming while they’re on the site, but they can also gain insight into the types of content they might be consuming offsite. They can then compare actual interactions against perceived interactions, and model that data to find their next-best customers.
4. Monetization: turn audience data into revenue
Having your own identity graph enables you to leverage, bridge and/or map to the demand side and drive revenue. Rather than relying on data collaboration partners who only offer publisher-centric data, or huge walled gardens who charge onboarding or per-volume fees, the secret lies in the ability to leverage a tailored yet dynamic referential data set that can be adapted and forward-licensed as required throughout the lifetime of the data partnership.
Where to start (and test) an identity graph
The first thing to do is ask: Do you need an identity graph? For some publishers, manually collecting customer data is sufficient to start with. However, as they begin to scale, they soon realize that having an identity graph provides greater flexibility to leverage the data across the enterprise, enabling the organization to more easily analyze, enrich and append the data set.
Using a trusted data co-operative who can aggregate multiple first-party data sets from a wide range of platforms and partners including marketing, sales, CRM, and HR is key to building a 360-degree view of a publisher’s audience.
Setting up an identity graph is as simple as inputting a couple of lines of JavaScript from the data provider, so for instance, Prebid partners can simply use it within the Prebid container. Then, when an anonymous user comes to the website, they can perform a match to see who they are.
Then it’s time to test the data available. To really get under the hood of how the graph will cater to their individual organization’s use cases, most publishers carry out a minimum of two tests: an enrichment test, and a cookie sync.
Enrichment test: Using a sample of say, 100k records, publishers can match their example 100k customer or prospect emails against an identity graph to gain additional attributes e.g. company, seniority, job titles missing emails, mailing address or wealth scores. Determining match rates helps to enrich their existing database, identify any data gaps, and start to understand revenue opportunities based on the enrichment.
Cookie sync: Publishers can also perform a match across a pool of 2 billion cookies on devices in the US alone, allowing them to create new identifiable first-party customers. Using the cookie sync to resolve unknown to known users also helps publishers with forecasting. One publisher - who focuses on direct-sold premium rather than programmatic or longtail - found their display eCPM recently decreased to an average of $20. Through surfacing a new premium segment of users to target, based on both the content they consume and their specific demographic subset, the publisher was able to quickly restore their eCPM.
As publishers look for more innovative ways to build audiences, they shouldn’t have to settle for inferred, probabilistic data. Instead, they can leverage observed, deterministic data from trusted, consented, and continually refreshed sources, without having to ask the question: where is the data coming from?
To learn more about how to level up your identity graph based on 360 million consumer records across 125 million businesses, as well as identity resolution and behavioral graphing, reach out to the 5x5 team today.