As a common definition, Data-as-a-Service business model is a concept when two or more organizations buy, sell, or trade machine-readable data in exchange for something of value. The DaaS providers are curating, aggregating, analyzing multi-source data in order to provide valuable analytical data or information.
First, data consumption is not stateless. Stateless processes can be performed without consideration for previous knowledge or past transactions. For example, a vending machine is a stateless transaction: a single request and a response. However, the process of operating data shouldn’t be stateless. Data usages have to be bounded by how and where the data was originally captured, its consent or expiration, its ongoing data usage rights, and current or changing global regulations. When companies erroneously abstract data accesses and usages into as-a-service paradigms, those paradigms often disregard the lineage of the data, and permitted uses of data even long after the original use case. Those are dangerous assumptions. For example, when companies capture registration or CRM data for digital personalization, they may have the right to utilize it for an existing session but they may not have the right to use it for other purposes thereafter.. The as-a-Service model fosters a stateless data transaction, ignoring the essential validity of data usages and accuracies of the data itself.
Secondly, data should not be co-mingled. The traditional as-a-Service model allows the client to consume outputs from one endpoint and then co-mingle with the output from another endpoint within business workflows. For example, companies can retrieve weather information from one endpoint, and then flight information from another endpoint to determine the best times to go somewhere. However, for accessing data, this design is unethical and precarious. Co-mingling data requires additional consent from the data producers such as consumers who should be empowered to decide whether they want their data to be combined. For example, someone’s CRM data might be captured at a publisher and their shopping data at a store; businesses can’t simply join these two data elements together in the workflow. Companies can’t simply assume that it was the consumer’s intent to have their data co-mingled with shopping data.
Lastly, when data is aggregated from multiple sources, its storage should not be centralized. Centralized data storage from multiple sources creates unnecessary security risks and violates the modern enterprises’ policy of zero trust data. Unlike Software-as-a-Service, the rationale of upgrading data is not intuitive. Instead, data upgrades should be treated as data operations which are conducted with appropriate consent and data rights. A decentralized data storage also significantly reduces many mismanaged data joins between different data suppliers.
As the public demands an ethical treatment of data, it’s essential for the industry to evolve away from implementing Data-as-a-Service in its current form, instead, let’s focus on building an ethical data infrastructure. Here at Adara, we are constantly researching ways to help companies develop an ethical data practice that earns the trust of your customers, your community and your employees.