Top 5 Things To Look For When Onboarding Third-party Data Partners

December 14th, 2017 | Kenneth Wong, Director, comScore Canada

paying online with a credit card

A May 2017 article from The Economist, titled “The world’s most valuable resource is no longer oil, but data” caught the attention of marketers globally. And rightly so. The top five most valuable firms globally that mine and assimilate this data (Alphabet, Amazon, Apple, Facebook and Microsoft), are an essential part of consumers daily lives. Amazon captures half of all dollars spent online in America. Google and Facebook accounted for almost all the revenue growth in digital advertising in North America in 2016.

The ripple effects of data dominance are now engrained into the culture of many advertisers looking to capitalize on their existing treasure trove of first-party data. The logical next steps are, of course, looking for ways to expand their reach with second and third-party data partners. This is where we see the complex data marketplace today.

First, second and third party explained

Let’s briefly summarize the varying degrees of data.

First-party data is data the advertiser owns, such as CRM, web traffic, online/offline purchases, mobile app location or call centre feedback. This data is coveted because the advertiser has complete transparency into and knowledge of its collection and relevancy. The main limitation to first-party data is reach and scalability.

Second-party data is simply what another partner would deem as their first-party data. This data can be acquired through data partners, and typically requires a data or information exchange between two agreeing parties. An example would be an airline and a financial institution issuing a partner credit card.

Third-party data is data collected by any company through either proprietary technology, or by sourcing and bringing together multiple touch points of datasets to stitch to their own. This is then anonymized and sold in the marketplace through direct channels, integrated within demand-side platforms, data management platforms and the like.

Companies typically will fully onboard and utilize their first-party data through re-targeting and look-a-like methods. But what do you do once those resources are exhausted? Complexity again comes into play when vetting and onboarding third-party data partners. Many global companies claim significant reach in Canada, but is this true? Deterministic or Probabilistic data, which method is preferred?

To help answer these questions, here are the top five things to look for:

Data Quality

This can be determined various ways. I would encourage open discussions and a deeper dive into the methodology and technology behind the collection and aggregation of data. Some common questions include:

  • How long have you been in business?
  • How is your data being collected across devices?
  • What steps do you take to anonymize the data? What about privacy and personally identifiable information PII?
  • How does your methodology account for cookie-less environments like iOS Safari?
  • Is your data probabilistic or deterministic? This isn’t a question of which method is better. But it’s important to ask WHY. Unless you are Facebook or Google (with scalable deterministic login data), generally a mix of both would be preferred in order to achieve greater scale while maintaining targetability. With probabilistic methods, the benefit is being able to combine multiple data sources and categorize probability segments so the end customer can choose the confidence level based on a varied audience size. As an example, with demographic segments, with a 90% confidence rate that you will hit females 25-54, the number of IDs = 5MM. But with a 50% confidence rate, it can expand to over 8MM.

Data Relevancy and Unique Selling Point

Is there anything unique that company A can bring to the discussion? Within existing DMPs like Adobe and Salesforce, companies explain their methodologies and unique datasets so you the advertiser/ buyer can do their research and find out where the missing pieces are. It is also important to find out “what’s next”…as in, what is this company working on looking into the future. This directly relates to the question above regarding Data Quality and the number of years in business.

Data Recency

Depending on the product or service you are selling, recency of data may be of utmost importance to your criteria when looking for a data partner. A lot of the time, if you are looking at retail or CPG, the sales cycle may require consumer behavioural data as recent as the last few days. By comparison, in automotive, the stage in which this consumer will likely book a test drive and visit the dealer may stretch over the course of a few months. Again, be sure to ask and get relevant examples specific to your objectives. Also consider at which stage of your marketing mix model (MMM) or purchase funnel you intend to require third party data.


This relates directly to the cost model, which generally varies by provider. There are really two methods: a flat monthly fee or a CPM. As data is largely commoditized, the need for flexible pricing models has never been more important. Ideally, you would want to work with a CPM model (pay as you go) at first. As you ramp up volume and start seeing performance, switch to a flat fee model as a way to level your expenses over the course of the year.

Local Support

This last point is an important one to consider. Will you have someone to call when something goes wrong? Will you have local support available that understands the Canadian marketplace and can account for nuances? Test and Learn is a huge component to really know if the data is valuable to your organization or not.

In an increasingly complex ecosystem with many data touch points, it is well worth the effort to ask the right questions and to build the cost of onboarding data partners into your ROI model to see if it really makes sense.

Kenneth Wong

Kenneth Wong, Director, comScore

Kenneth Wong brings a decade of experience working across agencies, publishers, technology platforms and measurement/ analytics companies. Currently a Director at comScore, Kenneth works with some of the largest global agencies, Canadian content providers and technology platforms. His role has recently expanded to encompass working with clients both on the buy and sell side – activating data, segmentation, and real-time targeting to deliver marketing and media ROI.