First, second and third-party data is not a tiered system. It’s the quality and transparency that matters, not the party label.
While “first-party” data will only ever belong to the company that sourced it, the data itself remains exactly the same as it’s passed from party to party.
It’s the same data.
Yet there’s a perception that these labels refer to three tiers of data quality – eg: “best”, “second best”, “last resort”.
So let’s clarify a few things.
First-party data:
This is when a publisher (or dating site, online store, ticketing, rate comparison site or any brand that’s able to gather rich, valuable and detailed specific consumer information) collects data about their existing audience.
That information is called “first-party data”. They collected and own the data.
Second-party data:
This is when a publisher shares their data in a one-on-one agreement with an advertiser. That information becomes the advertiser’s second-party data.
Second-party data presents opportunities for audience extension. Even those brands and publishers with their own owned and operated first-party data gathering assets will usually need to enhance their data set with some second -party information.
And they do this by acquiring second-party data in order to reach new customers. Not because they’re settling for any sort of “second best” data – but simply because they need data that they can’t acquire themselves.
Third-party data:
Then there’s the third-party data. Which is the exact same data as it was when it belonged to the first and second parties, only it’s being accessed by, or made accessible to, a third party.
So there are effectively two third parties:
1) Those who buy data directly from second parties, and;
2) Data retailers who make data available to other third parties (ie. DMPs and data aggregators).
Understandably confusing.
The point is: the data itself does not change.
Does a brand new car become worth less between the factory and the dealership?
Raw data:
Another term often misused and confused in the data game.
“Raw data” is just another term for “first-party” data.
It seemingly doesn’t belong to any party and is often perceived as being incomplete if used on its own. The perception is that raw data needs modelling and cross-referencing in order to become useful.
Raw data is actually just first-party data that hasn’t yet been cross-referenced or modelled.
All data starts out as raw or first-party data.
The lowdown:
We should acknowledge that first-party data is awesome and if you are not collecting and using it properly, then you are missing out on a massive opportunity to learn more about your audience.
However, first-party by itself has limitations – especially if you are prospecting for new customers. If you have only first-party data, you have only a limited view of your existing customers, based solely on interactions with your brand alone.
Second and third-party data enhances your view of the world, providing more rounded insights into these peoples’ interactions broadly across the web. This additional information can build profiles of your audience as actual human beings, rather than just silhouettes behind a screen.
In programmatic environments the more data you have to work with, the better. Automated trading with first-party data alone could well limit your scope for new business.
And, finally, data has to be reputable and accurate. But whether you sourced it from your own owned and operated data-gathering assets, or you acquire this data, it’s not ‘second hand’.
It’s the same data.
The credibility of data is dependent on its sources and transparency of collection and definition – not the label.
If you are a buyer, my advice is to check all your sources. It doesn’t matter who you’re buying data from, do your due diligence on the data collection process. The data party label means absolutely nothing.
Trent Lloyd
Co-founder and managing partner
Eyeota