Although we speak of ” data ” as a broad category, not all data is the same. These are the main classifications that we can establish according to the report.
According to its category
Structured data : structured data is the type of information that we usually find in databases, that is, a text-type file organized in rows and columns. This structure allows the information to be filtered, ordered and processed easily with the usual data processing tools.
Unstructured data : all types of data that do not respond to the aforementioned structure can be found here, such as those generated through social networks, image, audio and video files and a long . Given their format, we cannot process them with standard tools, so we have to look for other solutions.
According to its origin
Zero party data : the data that we obtain directly from consumers through studies and surveys.
Second party data : the data that comes from purchase and sale agreements with different partners. Within this type, we can distinguish between first party enrichment data (to expand the data we already have), those obtained through exchange and those purchased (through brands that put their cookies for sale).
Third party data : data from companies that are specifically engaged in the sale of aggregate data. In this case, we cannot access the source of the data or the treatments to which they have been subjected.
According to your collection process
Deterministic data : data that is collected from 100% verified sources (such as database records) and offers a unique value for the individual.
Declarative data : those provided by the user himself and therefore not verifiable.
Inferred data : those obtained by assigning an audience profile to a user according to their browsing patterns.
Modeled data : those generated by extrapolating from a representative audience sample, for example, the similar audiences of some social ad platforms.
According to the information they provide us
Behavioral data : data referring to the behavior of a user, for example, the pages they have visited or the keywords of the content they have consumed.
Intent or intention data : those related to the interest of a user in a specific good or service.
Who is the owner of the data?
The data ownership is an issue that is controversial and the legislation has evolved in recent years.
The first point to clarify is that the user is the owner of his data , since it is he who generates it. But there are also other actors who interact with the data and who may consider that they have rights over it.
On the one hand, we have the publisher , that is, the actor in whose environment the data moves. Without it, it would be impossible to generate the information, since there would be no place to create it.
On the other hand, we have the advertiser , who makes campaigns and allows enriching the information generated. In turn, the advertiser claims as much data as possible in order to improve its product.
And finally, we have the user , whose information is really only valid when it is added, that is, when it is combined with that of other users (for example, all visitors to a specific website).
At present, there is no stable picture when it comes to data ownership. New blockchain and hash-graph technologies could in the future allow users to sell their data without intermediaries, while publishers argue that users are already receiving value for their data in the form of personalized services. Meanwhile, advertisers are looking for solutions to create their own data extraction systems and gain independence from other players.
What is the value of the data?
The value of a piece of data is determined by the difference between buying an audience compared to buying only an advertising space, and it varies depending on various factors. We must bear in mind that we are still facing an evolving sector and that the value does not depend solely on the data itself, but on the tools we have to process it and on the training of the personnel in charge of this task .
In the world of digital marketing, we have different models to market data :
The sale of databases , email addresses for campaigns email marketing .
The programmatic buying and selling , where data begins to be commercialized. Thus, we go from the personal data of the users to data on behavior, , statistics and a long . Through platforms such as and Ad Exchanges, we can directly buy the audience segment that interests us the most.
The platform data management , which enable advertisers and publishers to access other data sources and reach agreements for buying and selling second party data based on different pricing models: cost per thousand, cost per transferred cookies, tariff models single or monthly …
The collaborative models in which an exchange of data between companies that do not conflict with each other occurs.
Two very important aspects to consider when using or acquiring data are quality and volume . These are some factors that we must take into account to determine the quality of a data:
The transparency on the source of the information and how it is processed. If we do not know this data, it is more difficult to determine its reliability and to predict the result of its use.
The persistence . Cookies have a limited duration and the user can delete them. In theory, it is possible to create cookies with a long duration (even months), but depending on the type of information to which it refers, it is possible that it loses value over time.
The obsolescence (very related to the previous point). Depending on the type of information collected in the data, the “expiration date” will be more or less long. For example, purchase intent information loses value over time, as the user may have already purchased the product. In contrast, information about permanent user characteristics, such as demographics, is much more perennial.
The provenance . The first party data tends to be of higher quality, since we know perfectly its origin and its processing. In return, the volume is lower, since they are limited to the data that we can collect on our own channels. Second party data is also usually of good quality, since it comes from known and theoretically trusted sources, but it is limited by our ability to reach financial agreements with the companies that produce it. Finally, the third party data is the one that raises more doubts about its quality, since we do not know its origin or the processing to which it has been subjected. In return, we can access large volumes of this type of big data.