How to: Data Science at a Startup
Startups can be amazing environments to learn quickly and apply what you learn in ways that provide your business a direct impact and value. My biggest challenge (and opportunity!) in my current role here at Gesture is to prioritize tasks that make the most sense for the phase Gesture is currently in along the data maturity curve and I thought this could serve as a good resource for others that are currently in this situation.
1. Embrace being a Generalist:
Life at a big mature company is different from life at a startup. In larger organizations, you have clearer delineation of roles and responsibilities and you more or less have a good understanding of what tasks you’ll be working on when. At a startup where you’re one of few “data people”, you’re stretched into a generalist managing everything from data analytics, data engineering, warehousing, modeling, testing, deploying, visualizing and reporting. That’s a lot to get familiar with and you’ll get the sense you’re constantly putting fires out. This can be rewarding but it can also be overwhelming. Many startups might not necessarily be in need of a tried and true data scientist. Rather, they’d likely be looking for someone that’s familiar with lots of various aspects of data and analytics to be able to wear the hats mentioned above.
2. Find the low hanging fruit:
Ultimately, small startups need someone who can make sense of the data they do have and help them leverage it appropriately, build out the data systems and warehousing, create dashboards and empower the business leaders in marketing, sales, finance and operations to be data driven, run some data experiments and allow the company to get a feel for data science, help grow the business and will take on the time and effort to skill up and take certifications for the business use cases that can really help the start up the most. This is an intuitive process. The better you are at listening, the more intuitive you are.
3. Push for data driven decisions:
The biggest impact you can make at the beginning when you’re in an environment without many peers is to help the business gather and use insights on a regular basis. By helping the business understand where it is now and how they can make better informed decisions using the infrastructure they already have, you’re finding a way to create legitimacy for your role and your department and create trust within the organization as you promote a data centric company culture. As the go to data resource, your startup wants you to be proactive in asserting your recommendations. Who better to do that in the best interests of the company?
4. Do data evangelism:
Becoming the company’s biggest data advocate while you help prepare the business for being data driven in the future is key. Building solid systems and helping make decisions about how to store and use data will allow you to add in more sophistication over time. Dashboarding and reporting may not be the really exciting stuff you thought you’d be doing as a data scientist but this will allow you to evangelize data importance throughout the organization and leadership team, and it will directly contribute to the health of the company on a daily basis. Having a good grasp on your data and leveraging it appropriately doesn't just help the company operate, but it's crucial in pitching the company to investors as well.
Once you can reliably collect and store data and have helped your business transition to being data driven, be self sufficient with gathering insights from the data, optimized existing processes and understand its place in the market: you can finally start using some fancy-shmancy algorithms or build a warehouse maybe!
Thanks for reading and good luck out there!
Irene Bratsis, Gesture Data Science Team