How Metabase helps every team at Bdeo get insights from their data.

The Challenge: Before Metabase, Bdeo.io used Redash, but many of the non-technical team members struggled with Redash’s user interface and its limitations.

The Solution: Bdeo needed a platform that could run SQL, but that could also create interactive queries for non-experts. They wanted to build dashboards that told good stories, and wanted a wide choice of visualizations and formats to help tell those stories.

The Results: Metabase made it easier for Bdeo teams to access their data and make better decisions.

"With Metabase, we have drastically reduced the time needed to visualize our client’s activity, and generated data-driven insights for the data & product teams. "

Javier Cubelos
Head of Visual Intelligence at Bdeo

The Visual Intelligence technology developed by Bdeo helps insurance companies automate certain tasks in their underwriting and claims management processes. This allows insurers to shorten their processes, save time and costs, and speed up decision-making to offer a better experience to policyholders. Currently, 48 companies in more than 20 countries use Bdeo’s technology in their processes. Bdeo has a team of 70 people, about 50% of whom are technical teams (data and tech).

Why Metabase?

Bdeo generates a lot of data, from customers, product, artificial intelligence models, etc, and all this information is stored in a relational database to be able to perform queries on structured data. They were looking for a platform for querying and visualizing their data so they could answer questions, spot trends, and monitor the behavior of their data (operational, financial, marketing, and data science).

The data

• Application data that tracks how people engage with their products.
• Global usage of their products, features, and services.
• Results of the damage detections to monitor their AI systems’ performance.
• Third-party data to analyze its usefulness and deep dive on it.

All this data is directly anonymized, formatted, and propagated from their transactional databases (DynamoDB) to their relational analytical databases (PostgreSQL) following an event-driven approach (triggered with any modification and using AWS Lambda for data propagation).

How Bdeo team uses Metabase

With Metabase, Bdeo built dashboards for:

• The data team, in order to monitor their AI systems’ performance.
• The product team, to understand how people engage with the product.
• The marketing team, to extract general knowledge and information.
• The growth & customer & finance teams, to monitor client usage.

The main teams that make the most use of Metabase are the data, product, and customer success teams, but the marketing, sales, and finance teams use Metabase as well.

Questions include:

Data team

• How does the performance of our different models evolve over time??
• Which of our predictions does the user modify? Should we consider retraining?

Product team

• How are the clients using our different product features?
• What are the main pain points of our current product?

Customer success:

• How does client usage evolve in both pre-production and production?Can we anticipate any problems based on these trends?

The data team makes greater use of SQL queries, as raw SQL offers deeper customization for more complex results; however, the query builder is very helpful for simplifying the query for other teams to explore.

The results

The main advantage of using Metabase is that it has connected the different teams through all the information generated by the company.

“Many of Metabase’s features are remarkable, but we could highlight the variety of visualizations and their customization available for queries.”, - says Javier.

Metabase offers a dynamic SQL query creation interface with the use of variables, snippets, and options to review table information interactively. This SQL editor has allowed people outside the data team to become more interested in how to query data and improve their SQL skills to meet their primary needs.

Javier’s Advice for others

1. Pay attention to all the features of Metabase’s product before getting crazy generating queries. For instance, the SQL snippets can help you avoid constantly repeating parts of the queries, and the variable filters can help you generalize queries and make them cross-functional.

2. Metabase allows you to organize questions and dashboards with collections, so a collections structure by area and project will make it easier for people to find what they’re looking for. You can also mark collections as official, which is important to differentiate validated results. You should put the same effort in extracting data insights as you do in organizing, sharing, and communicating them in an understandable way for the different departments involved in your company.