Not all data issues are created equal. While some issues may be merely annoying other issues spark hire-on-fire-moments causing your company to bleed $100,000 each day.
At Synq, we work with dozens of companies that are seeing an explosion in business-critical data use cases. A few years ago data was used for ‘finger-in-the-air’ decision-making but today the data warehouse is powering business-critical operations such as a Hightouch sync determining which customers to send emails to, value-based spend allocation to marketing platforms and automated pricing models to decide what to purchase products for.
This has increased the expectations of data, SLAs are now required for core pipelines, and teams are incorporating on-call rotas which were formerly only used in traditional engineering disciplines. dbt has played a leading role in giving analytics engineers the tooling needed to express expectations of their data in terms of deterministic tests.
But while deterministic tests are good at catching known unknowns, we frequently hear data teams refer to their most painful incidents as unknown unknowns; issues they didn’t anticipate but with critical business impact.
Different types of issues you can encounter throughout your data stack