How CaseWhen helped Buena to transform their data analysis with Metabase.

The Challenge: Before using Metabase, Buena’s’s team faced the challenge of disjointed analytics with no source of truth for their numbers. They had multiple Google Sheets with random analyses created by different people at different times, making it difficult to make confident decisions based on their data.

The Solution: Buena needed a platform where the most important KPIs were always visible, up to date, and trustworthy. That’s when they decided to implement Metabase. They hired Austin Levine of CaseWhen to help them set up their data infrastructure.

The Results: Buena estimated saving at least twenty data analyst hours per month. They were able to build both ad hoc analyses and internal and investor reporting faster and more reliably. The data culture at Buena improved as well; people began using Metabase to find answers themselves instead of reaching out to the data team.

"With Metabase we finally have one single source of truth for all of our company’s most important metrics. We save time on reporting, our analysis is more accurate, and everyone on the team has a clear understanding of business performance. "

Martin Hergert
Head of Strategic Business Development at Buena

Why Metabase?

Before Metabase, the Buena team had many Google Sheets with random analyses created by different people from different points in time floating around the company. There was no source of truth and no one knew where to go for the right numbers. One day they realized that if someone were to ask them basic questions about their business, like : “How many apartments do we have under management in this moment?” or “What’s the current landlord retention rate?” - they would first have to perform an ad hoc analysis, as they didn’t have a dashboard to point to. The team decided that the most important KPIs to their business needed to be visible, up-to-date, and trustworthy. They wanted to become a data-driven company. At some point, a fellow startup founder recommended Metabase to them.

Looker was another option they considered, but ultimately the price tag was a dealbreaker. Metabase’s UI was friendly and clean, making it super easy to get up to speed quickly. And investors loved how clean the design was.

Austin Levine, now co-founder of CaseWhen, an independent Data and Business Intelligence consultancy and a Metabase Expert, was previously an employee of Buena and a one-person data team. Austin helped Buena set up Metabase and create a single source of truth for the entire organization.

The data

• Application data: apartments, contracts with landlords & with tenants
• Interaction data: funnel completion and behavior on their website
• Third-party data: Stripe payments, Zendesk tickets

How Buena team uses Metabase

When the Buena team just started using Metabase, they wanted to create one main dashboard with their most important KPIs, from acquisition to contract signings. The more data the Buena team had access to through Metabase, the more questions and ideas came up for broader and deeper analyses. Eventually, Buena set a goal for each team to have their own dashboard to measure their performance.

Now the Buena team uses Metabase both for investor reporting and for tracking operational work of apartment management. Investors have their own dedicated dashboard, the finance team tracks cashflow and burn rate, and operations tracks the statuses of apartments.

Currently, about 50% of Buena’s team are able to ask questions about their data in Metabase. Most of the Buena team uses Metabase’s query builder, while more technical team members create their own questions in SQL.

“It always starts with an idea for an interesting analysis they want to create. One example is a cohort analysis to measure monthly revenue retention by landlord, to understand if newer cohorts are healthier (i.e., bringing in more revenue) compared to older cohorts. First we think about what the final visualization should look like, then we work backwards from there, thinking about the KPIs we need to create (if they don’t already exist), in which table that KPI should live / how to make it fit elegantly within the existing data model, and working back all the way to our source data, to make sure that we have every piece needed to create that final metric. Once that’s done, then it’s time for the fun and easy part: building the visualization in Metabase. The conditional formatting in the table viz is perfect for a cohort analysis. Setting custom min and max ranges and custom color scales, it’s so easy to quickly see which cohorts are performing well and which aren’t. We ended up building six different cohort analyses for different metrics because the first analysis was a hit.” - Austin Levine, co-founder of CaseWhen, the Data and Business Intelligence consultancy hired to help Buena with their data goals.

The results

With Metabase everything became faster and the quality improved across all of their reporting: ad hoc analyses, weekly internal reporting, and monthly investor reporting. The Buena team estimated at least 20 data analysts hours of time saved per month.

Discovering the Trend visualization type was probably their favorite find. They built charts to show every comparison imaginable: day over day, week over week, month over month, year over year.

Using Metabase also greatly improved data culture inside Buena’s team. Now, instead of approaching their data team with just a question, their employees first try to figure out the answer themselves using Metabase, and share that Question with the data team along with their data question.

Advice for others

Buena’s team set a company goal that Metabase became part of everyone’s weekly workflow. If Metabase didn’t make sense for their department yet, then the next task was for them to work with the data team to build meaningful metrics and dashboards that would help improve their department.

If you’re familiar with SQL, it’s tempting to just create all your visualizations with the SQL editor. But using the graphical editor is a much safer way to protect against future changes in your business logic that would require you to manually update countless charts.