Using the Elastic Stack for Business Intelligence at Liefery

Founded in 2014, Liefery is a delivery company based in Berlin. They provide transparent and plannable same day and next day delivery services. They are also developing urban delivery concepts for the future. By entering the parcel delivery market, their goal is to create the modern customer experience that people expect. This means predictable, plannable parcel delivery at the terms and at the time that suits the receiver.

We recently met with Simon Stemplinger, CTO at Liefery. He walked us through the company’s Elastic  journey that helped enable their business analytics capabilities with Kibana as a core component. As he and Liefery found out, the Elastic Stack is applicable to more than just log analytics.

Before using Elastic, what issues were you looking at solving?

Simon Stemplinger: We develop and run our own software platform to manage and support our delivery services. Our software spans multiple separate apps and services on multiple application cluster nodes. Looking through log files was quite painful, and therefore it was something we only did when we really had to.

Furthermore, we were looking for a business intelligence (BI) software to help our colleagues in the operational teams (where we manage our drivers and the parcels) make sense of our business data and keep track and improve our operational metrics.

We looked at a BI tool, as well as several hosted cloud-based logging apps, before coming across the Elastic Stack as a self-hosted solution for log analytics. I was familiar with the Elasticsearch DB from a previous project where we used it for a full-text product search and knew that it performed well and could handle large amounts of data. We gave the Elastic Stack a test drive and were immediately excited by the combination of the power of the Elasticsearch database and the flexibility and ease of use of the Kibana visualizations. Kibana has quickly become the go-to tool for our answering all questions about our business data.

So you started using the Elastic Stack for log analytics. What other Elastic Stack use cases have you carried out?

Stemplinger: When we started centralizing our logs using the Elastic Stack, several initiatives were also launched around reporting and service quality management. Most of these initiatives were using Excel. It was very tedious and the user adoption was slow. At some point our engineering team thought: "Kibana is great for data visualisation. Why don’t we just try throwing our business data in there and see how far we can get?" 

I sat down with our Head of Operations and presented him in a demo how to create visualizations and dashboards using Kibana. He was immediately hooked. Despite not being a technical person, he managed to build in a short time quite in-depth and detailed data visualizations and dashboards. We created system health dashboards for IT and operational quality dashboards for our business teams. We then set up large flat-screen TVs with Raspberry PIs attached showing these dashboards during normal work hours. Kibana became our new de-facto quality and reporting BI tool. 

Then we started using the commercial features, as we wanted to widen our user base to all parts of our company. The security feature of the Elastic Stack lets us give certain users read-only access or limited access to certain indices. This allows us to give read-only access to a wider part of the company while giving write access only to certain trained users. In the future we might extend this to allow access to certain data (e.g. financial data) only to privileged users. 

Our Kibana based reporting solution is still widely popular in the organization. In addition we use the opportunity of having both all application log and business data in one database to do all kinds of business alerting (detection of brute force attacks, API issues, pathologic business data, etc.).

What are the future projects you’re looking at developing with the Elastic Stack?

Stemplinger: Having made a name for ourselves as an innovative software-driven company in the urban logistics space, we are starting to offer our software as SaaS software platform to everyone. A number of key players in the industry have already expressed interest and are starting POC projects in various stages. The Elastic Stack enables us to extend our service offering with a powerful analytics add-on that we can pre-configure with help reports and visualizations out of the box, but that is also highly customizable to our clients’ needs.

Can you tell me more about the impact Elastic has had on your company and the team?

Stemplinger: We experienced tons of process improvements. Today, using the Elastic Stack, the business team can save time on their daily review calls on service quality, they don’t need to allocate more time to actually prepare the report and can just at the quality reports in Kibana. In a similar way, the developers can also quickly detect issues using performance dashboards on which they monitor key metrics like the response time, they also enhanced the data with users IDs to track what’s being done by who. 

In a way, Elastic has helped our teams become more responsive and agile. The Elastic Stack helped us overall to get more insight into our data. The result manifests first and foremost in better service quality, less shipping delays, better predictions, as well as more informed meetings and discussions. It allows us to quickly troubleshoot technical problems and gives both tech and business teams a very powerful tool in their arsenal.

Simon Stemplinger (@stemps) is the CTO at Liefery. He is a hands-on developer with a passion for databases, and is a big fan of the Elastic Stack. He spent the last 5 years building up the software of the company; from the first lines of code to a fully featured last-mile logistics software platform and SaaS product.