processed per day
reduction in engineering hours
to get data ready to be consumed
Situm turns any mobile phone or bluetooth low energy transmitter into a precision tracking or navigation system for more than 6,000 buildings around the world. If you’ve ever used indoor navigation in an airport, mall or office building, or if you work in an industry where high value assets are monitored around the clock, there is a good chance that Situm has provided the technology to make this happen.
Tracking assets and providing navigation means a lot of data. Every day, Situm ingests more than 300 million geolocations from around the world. With double digital month-on-month growth of the number of gelocations ingested, Situm’s CTO, Adrián Canedo was facing some major challenges.
Situm’s tooling & databases were not going to keep up with their growth forever. At least not without a big investment - both in terms of dollars and engineering resources. Originally built on PostgreSQL, and later on MongoDB, the engineering team found themselves spending more time fine-tuning and maintaining their database. This was taking resources away from product development and in-turn slowing down the release of new features. Not to mention customer tickets: since Situm didn’t have a simple interface to easily explore customer data, customer support was regularly reaching out to the development team when it came to solving customer tickets.
But more importantly, speed was an issue for Situm, when it came to the speed of development. It was difficult to materialize large volumes of data. And, that made building real-time data-intensive products difficult. The results weren’t real-time. Situm saw an opportunity to serve their customers with new real-time analytics and reports. But to do this, they would need to build complex ETL pipelines and preprocess the data first. Changing a query would require a complete do over too.
For Situm’s customers, time is money. And delays can be costly. One customer, a security company, wanted reports which analysed in real-time the number of users within a geofence, user trajectories, heatmaps, or time spent in a zone, so that they could take action and respond to incidents on the fly. Other customers needed real-time reports to gather insights about customer flow or for Covid contact tracing.
When Situm made the move to Tinybird, the results were instant. Literally. Adrián and the team at Situm use Tinybird to speed up geospatial analysis, and create incremental materialized views for their customers. Using Kafka, data ingestion lag has been reduced to less than a second. Scalability is no longer a concern as visualisations and reports involving billions of geolocations can be computed on the fly, without the latency of their previous stack. Tooling complexity is no longer an issue, and Situm can focus its efforts on development rather than maintaining database clusters or other infrastructure.
With an intuitive user interface, the customer success team can easily perform data exploration and generate new prototypes and reports with milli-second latency. There’s no need for complex pipelines, preprocessing, or in-depth SQL knowledge. Client reports can easily be tweaked by adjusting the API Endpoints, which can be created in a single click. With the customer success team able to get back to customers faster, customers are happier - and so are customer success as they are independent and are self-sufficient. The development team can concentrate their resources on development rather than tickets raised.