Newsletter
First day on the job is challenging. What should your initial focus be?
Over the past few weeks I had the opportunity to talk with great data folks just starting or about to start their new job/role. We spent time discussing what their initial focus should be in order to maximize their impact on Day 1. I wish them all the best!
First day situations can vary quite a lot:
You're the first data person — until now your CTO or an appointed developer extracts some raw data from the database here and there. The finance team extracts data from the payment system in their Excel sheet. Together they built a dashboard they are proud of (should they...?)
You're joining an operational team as their assigned Data Analyst — The Data team starts to organize itself. They hired the right individuals for the right job and works in a federalized model. You're joining the team and are embedded into the Sales teams to help them improve their operations and strategy.
You've been internally promoted "Head of Analytics" — Congrats! Sometimes it starts as an interim position. If it's the case now is your chance! Prove that you can make it work and this interim position will become permanent. If you've been assigned a team, don't forget that your #1 mission is to ensure your team's success by enabling each individual contributor. Your personal operational output doesn't matter that much.
Bringing value at day-1 does not mean say yes to every thing at day-1 exactly. It means understanding the context and gradually bringing your touch and magic over the course of the first months.
Those are some examples but there is an infinity of situations. Some facts stay true in most of them.
Stakeholders whether they are in Sales, Operations, Marketing or Product don't always know much about data. When they do, it's very often the signal that the person who was in your shoes before made a great job!
Do not undervalue them, they definitely know a lot about their job and you most probably don't about theirs. You don't know about the specifics of your company yet, the new market you're in, nor the constraints a specific team faces.
Your first month is a perfect time to ask them to spend time with you. You can run interviews about their processes, what they feel is interesting, what would make them a better operator.
While you won't implement everything they say, you'll remove key blockers. First, you'll build a relation with them. Later on, this relation will be critical in ensuring they consider both you and your work.
Very often, people limit those interactions to a couple of meetings and interviews. One great way to start is to shadow a couple of stakeholders for a few days. They will feel valued and you'll learn a lot with a very limited bias.
When buying shoes, you first expect the shoe store to be opened during their opening hours. After buying them, you expect them to be functional when you need them. It doesn't mean those shoes should be the best in all conditions but you should be able to switch when conditions call for it: hiking, city walks, working from home...
As an analyst, you are the shoe dealer and whether or not you decide to also be the shoe maker, you better provide good shoes for all conditions!
Ok, enough with the metaphor. What does that mean? You should build certainty around the tools and processes you own. It does not mean a unique tool nor a unique process for all data access, but people should know when to use a specific tool: when it's wise to go ask the data team and when it's not.
You'll face different personas with different needs, try to understand them before forcing a new shoe tool. Write down documentation to ensure knowledge is shared around. Also, working with an agile methodology from the software engineering world helps with the transparency of the processes.
Finally, your system as a whole should be as reliable as possible in terms of uptime, quality, speed...
In most cases, a similar question to the one you're trying to answer has already been asked earlier. Sometimes even the exact same one. You should start from there.
While you might think it will be faster to start from a blank page because you won't have to understand all the details of a previous work, it won't. In particular, you'll ensure continuity of service and therefore the "certainty" we were talking about limiting the costly back and forth when delivering the results. Biases from the first analysis might then still exist in the current one, making them even easier to compare.
You'll have time to change things and add your style over time. Right now is not the right moment.
If you're the first data person, you might actually need to invent the wheel though.
Try to use and implement generic tools and process that are easy to setup and easy to maintain for your solo team.
Try to make time for both data modeling, self-serve and ad-hoc analysis in parallel.
Consistency in tools and processes bring the certainty and the trust you need in your first days!