website templates free download

What is fourth-party data and which is its role the customer service and martech evolution?

Which additional value-added cases could be built combining fourth party data with others?
Is the CDP entitled to use fourth party data to improve the end-customers 360 view profiling? 

April 25th, 2022

When discussing about customer data, privacy laws or usage of data for activation, the initial well know classification is:

  • Zero party data is the data that a customer intentionally and proactively shares with a given company like: communication preferences, product interest in case of retail, maximum risk/volatility assumed in case of wealth management, etc.
  • First party data is the data that a company collects directly from its customer’s activities (digital channels, usage, CRM, purchases, etc.) and that can be used with consent.
  • Second party data is basically “first party-data” from other companies that is shared with consent by the end-customers with the company.
  • Third party data is any data that is collected by other entity that does not have any direct link to the visitor or customer; it could be social media profiling, website browsing tracking activity, etc. The DMPs have a leading role collecting and processing this type of data. It can be used with/without consent (but should be consented) to create new segmentations, targeting ad-hoc audiences, etc.
CRMs and, most recently, CDPs are using those types of data in order to build the so-called “360º customer views” and also to create sophisticated segmentations with more conversion efficiency.

The questions that emerge now are: could additional types of data be used to verify how trustable are the other types of data? Can they enrich the 360º customer view and identify new types of audience/segmentations?

In our opinion the answer is YES, and our approach is the FOURTH PARTY DATA.

Fourth Data Party is time-series databases of non-customer data, but general data (weather, stock markets, sentiment about certain topics, levels of advertising on a particular vertical or competitor) that crossed with customer actions (first party data) provide additional profiling use cases:
  • Understanding the drivers or concerns of each end-customer, by measuring correlation/causality between their actions and the level of those external general events. It can help deliver a more accurate, timely and personalized CX. 
  • Design new types of audiences/segments based on customer behavior regarding general events, so they can be activated by specific campaigns when they could be more relevant.
  • Verifying the accuracy of zero party data. The accuracy of zero party data is not fully reliable. Customers sometimes do not fill with total accuracy the surveys/forms collecting zero party data or even in the case that they filled them in with total accuracy, they could change some of their views along the time leaving the “their zero-party” info outdated. For example, if a brokerage end-customer declared he was a risk-appetite investor and suddenly in a market downturn he checks his portfolio with anxiety, via many channels, perhaps his circumstances have changed or he was not as risk averse as he thought.
All these cases and more to come will be key not only for marketing users but also in the Contact Center, since both automatic interactions (chatbots) and agent-based conversations (chat, voice or videocalls) can benefit from a better customer knowledge, some kind of mindset-insights map.

There is plenty of public information available about general data (weather, traffic congestion, stock/market evolution, public transport occupancy, airlines passengers, etc.) that can be obtained through public APIs (paid and open-data). There are also companies providing sentiment analysis about multiple topics and measurements of advertisement pressure / media presence / reputational level of companies. Thus, a myriad of behavioral experiments and related external variables can be introduced to enrich end-customer info and the 360º views, validate zero-party data and create new kinds of audiences, based on “potential causes of an action”.

The optimal way to introduce this data for analysis is: 1) API-based connectors for easy data collection, 2) #nocode flexible data-model that allows the inclusion of both objective numerical data (i.e. stock variation) and more subjective info (sentiment about a given topic) with several degrees of granularity.

Finally, a last question to be considered is “who” (which IT element) should be in charge to provide new use cases using fourth party data. We tend to be agnostic and several options could match the same goal:
  • A holistic/general CDP could collect fourth party data and create the internal logic to make the analyses and enrich the profiles. 
  • An “unbundled CDP” approach where a DataWarehouse / Datalake or a primary CDP can delegate in an “specialized CDP” this task. To do that, the primary CDP would share the zero and first party data, while leaving to the specialized CDP the collection of the fourth party data and the creation of the framework for behavioral experimentation. Finally the specialized CDP would return the enriched profiles and new types of segments/audiences to the main CDP.
Your thoughts? Let’s discuss! 

You can follow Whenwhyhow in linkedin, twitter & instagram