Of all your data activities, exchanging data is perhaps the most risky. Before you share data with partners, consider the following threats to your success: privacy lapses, misaligned identities, high costs, vendor lock-in, and long time-to-value.
- Privacy
Some vendors contend that data use agreements are sufficient. They are not. Without proper technical safeguards, privacy infractions happen, particularly when sharing data at scale. That's why CIOs should look for math-based approaches to privacy to complement legal mechanisms.
Privacy-enhancing technologies (PETs) represent an ever-growing group of cryptographic and encryption protocols (math, basically) that let companies glean insight from data without betraying privacy. Look for vendors that employ techniques such as differential privacy, secure multi-party computation, and query thresholds.
2. Identity
Companies have different data models and disparate customer information. Jane Smith may provide one email address to you and another to your partner, for example. Or perhaps she accidentally misspelled her address. Look for data exchange tooling with a strong identity framework, acting as a "rosetta stone" between you and your partners' people data. Effective identity resolution is essential to getting the most out of shared data.
3. Cost
Moving data into a third-party clean room—every month, every week, or even every day—to enable data exchange can be expensive, operationally and computationally (egress fees). Look for vendors that offer a "federated analytics" option. Federation is an analytics construct that lets analysts conduct SQL queries and machine learning without having to centralize data to a single location. This means you and your partners can jointly utilize data across your "walls," clouds, and borders; as if it were all in a single warehouse.
4. Vendor lock-in
Within a fast-growing market, jockeying cloud providers may downplay the inevitability of a multi-cloud market and propose work-arounds that promote further consolidation. You should proceed with caution—clouds' incentives to build seamless interoperability with other clouds are few. Instead, IT leaders should consider leveraging a neutral, third-party software solution to provide data translation and connectivity across clouds.
5. Long time-to-value
Implementing an exchange solution should be as quick and painless as possible, as some data exchange opportunities have a shelf life, and any data exchange will come with pressure to prove ROI quickly. Look for vendors who can abstract away complexity, providing expertise in legal, data ethics, and infrastructure. Not all vendors can contract, implement, and set up partnerships at the same speed, and their differences are measured in months not weeks.
Selecting a provider that focuses on privacy will help you build trust with data governance stakeholders who are—and should be—gatekeepers in the data-sharing lifecycles. Perhaps paradoxically, speed-to-value requires cautious solutions.
Privacy. Identity. Cost. Vendor lock-in. Time-to-value. Learn more about how LiveRamp can help you navigate these risks and make the most of data sharing, safe and sound.
