The Data Ops team is a fairly new team at Stuart whose goal it is to create closer cooperation between the business teams and the engineering teams. Our job is split in two: one part is short term requests to satisfy the operational needs of each department in each of the different countries in which we operate, and the other is our long term strategic projects.
Some of the things we’re working on
The best way to describe how the team achieves this goal is to provide some examples of the projects we work on!
🏗 The Data Warehouse
Our biggest long-term project is definitely the new data warehouse.
This warehouse will be the one source of truth within the company and will consolidate all relevant data from the internal and external sources we currently use.
🙌 Making data available to everyone
One short term project we’re focusing on is getting offline operational data online and making it accessible to everyone in the company. The data warehouse will meet all of the data needs of the different stakeholders within the company, such as Sales, Account Management, Operations and Management. With such a diverse list of stakeholders we need to gather data from the production database as well as external data sources, such as Staffomatic (driver schedules), Fountain (driver onboarding), Shopify (equipment management), Darkstorm (weather), Salesforce (client data) and Intercom (driver/client communication). We take all this raw data and transform it into the most efficient format for further analysis and insights, all available in one single data source for everyone in the company. So why do we need all these datasources and what are they used for?
⛅️ Supply and Demand Forecasting
The process of forecasting supply and demand is a good example of why we need several sources of data in the warehouse. Supply and demand can be predicted based on historical volume, but they are also impacted by weather and special events, such as Christmas Eve or New Year’s Eve. In order to make a meaningful prediction of future volumes you need to know about events and weather in the past and the corresponding change in demand, as well as which events and what type of weather is expected in the near future. Having all of this data in one source reduces the likelihood of human error while gathering and combining the applicable data, and also minimises the time spent on gathering data, so that more time can be spent on analysis.
💸 Lifetime Value
A benefit of having external sources means that we can perform a deeper analysis on the lifetime value of a client or a driver using data gathered before they began using the Stuart platform. This is meaningful when calculating lifetime value because we can look at what their acquisition cost was in combination with revenue generation and strain on the support team. As a result, we can make more informed decisions about the most cost efficient driver/client profiles and where to source them.
Solving urgent operational challenges
Another aspect of our work is based on urgent projects, which we work on with the business teams, helping them solve difficult challenges.
Anyone who has worked in the daily operations of a company knows that projects appear and need to be solved at a fast pace, which doesn’t always fit within the 2 week sprint format of many tech teams. We therefore have a dialogue with the business teams and try to meet their needs in a timely manner, even if it means putting our long term projects on hold. Most of the data needs of the business teams can be met by the colleagues within the team, but sometimes there are urgent opportunities that require data to be processed in high quantities or at higher frequencies. These needs are impossible for our operations team to meet with the tools at their disposal, and therefore they contact the tech team. In the long term a feature will be added to the product in order to cover the new need, but due to the longer time horizon of building a new feature, the data ops team cooperates with the business teams to build a robust intermediate solution which can be used in the short to medium term. An example of such a project is offline operations.
⬇️ Offline operations
When clients request features that are not yet ready in the product, not providing the feature could mean losing the client completely. In those instances one solution could be to take the client off the app until the feature is ready, which is known as offline operations. However, as the business has grown, the need to make an automatic process for offline operations has become evident. The data ops team is now responsible for automating the collection of this data, storing it in a meaningful way in the new data warehouse, and providing the correct data for payments. Additionally, the Business Intelligence team worked with the data in the warehouse to build a dashboard for Account Management to follow the metrics of these clients. Getting all the data into the warehouse means we have a more transparent and easier way of seeing historical data, and also makes the data available to the whole company. Automating the process makes the data cleaner since human error is less likely. The offline operations project is a good example of how the short term projects can be tied together with our current long term project of building a warehouse with all relevant information that the business teams need.
In Short: We Connect Tech with Operations
From the above examples it’s clear that being familiar with operational processes is essential in order to efficiently build the warehouse and quickly provide intermediate business solutions.
80% of the data engineers, as well as the product manager on the team, have previously worked in Operations and still keep close connections with these stakeholders by sharing office spaces with them, instead of sitting with the rest of the tech team.
The team members have a deep knowledge of how the company works and how data is used within the company and this knowledge makes our communication with business teams easy and efficient.
Being so close to the business teams means we can afford to spend less time on communicating the operational needs of each project, and more time on actually building the product.
The existence of the Data Ops team is crucial because it has made Stuart’s data accessible across the company, empowering all teams to make data based business decisions using a single comprehensive and reliable data source. The team also serves as a bridge between the operational side of the business and the tech team, ensuring that the operational teams have access to and knowledge of the best technology to use, in order to solve their challenges in the most efficient and scalable way.
Like what you see? Join us, we’re hiring. 🚀