From Marketer to Analytics Engineer
Especially in B2B companies, there are those on the Go-to-Market and Operations side of the house who work with data and eventually wonder: is it possible to switch to the Data and Analytics team? The barrier to do so is less imposing than it seems, as we learn from this interview with Mitchell Wright, Senior Analytics Engineer at Vercel and Polytomic customer.
In our interview below, Mitchell describes his journey from Marketing and Operations teams to becoming an Analytics Engineer on Vercel’s Data team.
Polytomic: Thanks for joining us, Mitchell. Let’s begin with a basic question: what is your current job title at Vercel and what does it mean?
Mitchell: Pleasure to be here. My current title is Senior Analytics Engineer at Vercel. The responsibilities I have are mainly around data modeling though I also work further up and down the data stack. For example, I also work on analytics, dashboarding, and miscellaneous BI stuff. Further down the stack I also do some engineering work on our data pipelines.
But like I said, my work mainly centers around taking the data in our data warehouse and modeling it to fit the use cases of various teams within Vercel.
Polytomic: When you say ‘various teams’, can you define that? Are you focused on one or two teams in particular, or does your work involve all teams within the company?
Mitchell: Definitely the latter. For example, I recently worked with the Finance team to automate a bunch of monthly processes around closing our books, as well as assisting with assembling data and reports for Vercel’s board meetings.
I’ve also worked with the Engineering team on product metrics to help measure the customer impact of Vercel’s various products.
Right now I happen to be doing something completely different: I’m working with the Marketing team to improve our attribution models and help them measure the impact of their spending so that they can make better decisions on allocating their budget.
Polytomic: Based on this, it’s clear that you’re steeped in your role on the Data team. But you weren’t always on this team - you joined Vercel to do other things. Can you describe the journey that led you to this?
Mitchell: I think it would be illuminating to start from my time before Vercel. My career started in digital marketing. I was working at companies like Qualtrics and Gitlab, helping manage ad spend, run digital webinars, and other typical B2B marketing work.
I then joined BloomTech and stuck my head into Operations: I spent time with the Marketing and Sales Operations teams, automating processes and generally making those teams more efficient.
My entry into Vercel happened through a former coworker who was there as an overwhelmed one-man Marketing and Operations team. He hired me to help automate and improve various workflows for the sales and marketing teams. This work involved looking at and moving a lot of data.
Note that in the midst of all this we did not have a data warehouse at Vercel. Any time someone needed product data or an analysis of it they’d have to hit the production database. This was obviously not a good idea at our scale.
So the company needed a data warehouse. My boss at the time assigned that task to me as I was already mired in systems work. We set up Snowflake and Polytomic to stream updates from our production database into it. This was a turning point in the data journey for Vercel: our executives now had insight into all sorts of product metrics to help them decide what to improve for our customers. It was time to assemble a professional data team.
We hired a data team and a few months later I requested a transfer to them. Note that my team was not even in the same org chart as them. I was, on paper at least, a foreign body. But my past work with systems, data, and the data warehouse they were now in charge of was encouraging to them. They assumed that, with minimal guidance, I’d be an asset to them in short order.
Polytomic: It certainly seems that calling you Vercel’s first data person (despite your job title at the time saying otherwise) would not be unfair. Out of everyone at the company, why did it fall on you in particular to set up the data warehouse? Did you just do it without being asked? Or did someone assign the task to you?
Mitchell: The spark that started all this was that we (the Operations team at the time) were struggling to get access to product data. As a product-led-growth company (i.e. self-serve user signups) this was important. At our scale, you could only feed so much event and product data into Salesforce before realizing that a company-wide analytics solution lies elsewhere.
Any of the engineering teams at the time would have been more qualified to set up our data warehouse, but in classic startup fashion our team was the one with the problem and all other teams had their own problems to deal with. So I took it upon myself to implement the solution.
There’s a general comment about startups that’s worth making: in a startup you’ll always have more opportunities to do things that you normally would not be able to in a larger company. Because everyone has so much to do they care less about who tackles problems - they just want them solved. People are less precious about role definitions and boundaries.
I think my transition to the Data team would have been a lot less likely in a larger company.
Polytomic: How would you describe the differences and similarities between working in Marketing and Operations and working in a Data team?
Mitchell: One obvious difference centers around work culture. Marketing teams tend to be meeting-heavy. When I was in marketing I felt like I was always in meetings and discussions with many people talking. In contrast, on the Data team, you’ll often do a bit of talking upfront to spec things, but then after that there are days where I have no need to speak to anyone. Instead, I spend my time on reviewing and approving pull requests as well as thinking about, modifying, and testing data models.
On similarities, the major one is the iterative nature of the work. Running marketing ad campaigns is an iterative process: you’re constantly tweaking things based on the performance of your ads.
Working with analytics data models is also iterative work. You’re always modifying them to better service the requests from your coworkers who use them.
I suppose these are examples of a larger commonality in all work: you put something out there for people, take input and feedback, then iterate and repeat.
Polytomic: People who switch fields often have things to say about advantages and disadvantages that their former experience gives them in their new field. Do any come to mind?
Mitchell: The main advantage is significant: domain knowledge. For example, when I work with the Marketing team on attribution, I already understand their problems without anyone having to educate me. We can immediately get to work solving problems.
The other advantage may be more subtle but is still important: knowing the importance of marketing your work internally.
A lot of the work that data teams do, as hard as it can be, is often in the background or downright invisible to other people. You need to market your work internally for two reasons:
- You may have useful tools and services to offer your coworkers, but they’ll never take advantage of them if you don’t educate them.
- Especially in larger companies, if no one knows you did something then you’ll never get credit for it.
There is a self-promotional aspect that technical people, including data teams, shy away from. Working in marketing teams caused me to be aware of this right at the beginning when switching to the Data team.
This is something that the Data team at Vercel certainly spends time on. Every time we put out models or infrastructure with wide applicability, we make sure to announce them to other teams and highlight use cases they could fulfill with our work.
On disadvantages, nothing major comes to mind. Perhaps the biggest was how rusty I was with SQL! I took a SQL class as a student but my use of it was superficial. This was something I definitely had to progress on, though that happened quickly. This also allowed me to naturally embrace tools like dbt.
Polytomic: One issue with data teams is that, especially with go-to-market or finance data, they don’t quite understand the data they’re working with due to a lack of domain knowledge. Do you have advice for the new data analyst who’s thrust into a project involving business data they know nothing about?
Mitchell: Absolutely. Understanding the domain you’re asked to analyze is essential. There are two ways you can do this.
First, there are many online resources you can consume to educate yourself. Whether it’s articles on the web, YouTube, or even getting overviews on topics from LLMs (ChatGPT and so on).
Second, and I think this is underappreciated, shadowing your coworkers is always immensely educational to understanding their role and domain. This is something I used to do when I worked in operations: I would sit next to coworkers and watch them do their jobs, thus understanding the inefficiencies in their workflows. I think data people can follow the same advice and end up with both knowledge and empathy towards the domains whose data they analyze.
Polytomic: Let’s turn to those on the go-to-market operations side who are flirting with joining a data team. Any advice for them?
Mitchell: I suggest simply getting more technical then applying that knowledge to your work.
There are so many online resources that can help with technical topics and learning SQL. Especially with ChatGPT, once you have some basic understanding there are a ton of opportunities to have ChatGPT write a basic SQL query for you. You can then apply that query to your data, see results, modify it, see different results, and so on. You can even get ChatGPT to explain the SQL to you line-by-line.
Once you have some basic level of proficiency, look for opportunities in your own work to apply this knowledge. There are always more efficient ways to pull data, combine it, and improve workflows. You’ll eventually build proficiency and confidence in your ability to work with data.
Separately, I want to emphasize again the benefits of working in a startup. Startups give you the opportunity to take on work beyond your assigned responsibilities. As companies grow larger, they often switch from hiring generalists to needing specialists. In that environment people’s roles become more rigid, making cross-department transitions harder to pull off.
Polytomic: Sage advice indeed and a good place to end. Thank you kindly for making the time to chat with us, Mitchell.
Mitchell: My pleasure!
Do you need to sync data to or from your data warehouse? Book a demo with Polytomic to see how we can help.