(US and Canada) Pascale Assémat, Chief Data Officer of the Consumer and Digital Branch at Le Groupe La Poste, speaks with Kensu CEO Eleanor Treharne Jones in a video interview about the difference between data observability and data monitoring, and the approach and success criteria for implementing observability.

Speaking about the differences between data observability and monitoring Assémat mentions the following:

1. Data monitoring is monitoring the pipeline but data observability is monitoring what is inside the pipeline

2. Data monitoring is finding what is wrong but data observability tells why it is wrong and helps in detecting signals to avoid bigger problems

Speaking about the steps a company needs to take to start considering data observability, Assémat says that they should begin by explaining the consequences of not doing anything. This is followed by understanding the problem to be solved and exploring the observability tools available on the market. She urges data organizations to start small and choose a critical data pipeline for the business that has a lot of anomalies in production. The next step would be to convince and then deploy the tool by integrating more and more data domains.

According to Assémat, the success criteria for observability includes:

1. Fewer anomalies

2. Predictability of anomalies

3. An improved trust level of the business and data teams on the ability to react when needed

When asked why observability is still a new topic, Assémat says that it is already being practiced but is not known by the specific term. There may not be a single tool that addresses the different aspects of observability since the different topics were addressed separately. Observability is structured, has a community around it, and has proper tools today, she concludes.

CDO Magazine thanks Pascale Assémat for sharing her insights and data success stories with our global community.

See more from Pascale Assémat

CDOs Can Write Their Own Job Descriptions

Data Quality Means Nothing Without Context