Recycle, recycle, recycle

Leveraging existing work as answers to new questions

One of the safest ways to provide a good answer in data is to give one that's already been used before.

When an analyst has already done the work, they’ve gone through the rigour of framing the question, making sure the metrics are the right ones, that the data is filtered and sorted the right way… There is obviously a large time-saving from reusing old work, too.

We have used this concept for our latest feature in Delphi: Semantic Search.

When a user asks a question, we find semantically-similar work in their BI tool and offer this as a solution to the user. We show these existing pieces of work alongside metadata about who created the work and when it was last edited.

Leveraging existing work also helps provide an organisation with consistent answers. One of the pitfalls of legacy BI tools is dashboard bloat - many pieces of work to answer the same questions are created (and not deleted). Multiple solutions can have slightly (or very) different answers! This leads to confusion and delay, as stakeholders question the data rather than being able to use it.

Of course, if existing work doesn’t answer the question, you can still run a new request as before.

We’re looking to bring on some early design customers soon. If you use the dbt semantic layer, Lightdash or Metabase and are interested in Delphi, reach out to me or sign up at delphihq.com