Q&A with Snackpass' Data PM, Nick Marwell

Happy November! LogicLoop is teaming up with industry leaders at high-growth startups on an end-of-year roundup on operational best practices. 

Today, we’re chatting with Nick Marwell, Product Manager #2 and Data Team Lead at Snackpass. Snackpass is a Series B company offering restaurants self-service kiosks and building social commerce.

Snackpass is hiring! Check out their jobs here.

What are you responsible for? 

I’m the product lead for 3 teams: restaurants, payments and data. Restaurant deals with the hardware and software restaurants use. Payments deals with the flow of funds and accounting. Data deals with the internal analytics and reporting, as well as sales and ops tech.

Which team do you spend most of your time with, and is it seasonal?

These days I spend the most time on restaurants, followed by payments, and then data. Historically, our business was pretty seasonal and we’d see clear declines around summer and winter breaks. Now, we’re increasing demand from non-students even on campuses, and expanding to cities. This is reducing our seasonality. 

We need to ensure things never fail silently

What are your biggest goals and operational challenges today?

I’ll break it down by team. In payments, our biggest goal is to reduce accounting discrepancies when we pay out restaurants. We’re laying out the long-term roadmap for more margin-efficient payments processing. We’re also integrating with services that will make our tax application and reporting a lot more reliable. 

For the data team, what keeps me up at night is that we spent the past two years building for speed, and making decisions with 90-95% confidence in our data. Now we need to raise our internal confidence. We know that slippages will happen when eng ships something, or data gets corrupted. We need to ensure things never fail silently, so detection time goes from weeks or months to minutes. We also really want to integrate ops and sales into the data world. Ops has historically not had anyone technical, and Salesforce has one admin. They only have technical support available for one-off tasks, but we want to empower them to make better decisions. On the operations/support side, it means we have to make them more aware of the problems restaurants are having faster and more often, before they churn. On the sales side, it means linking up our Salesforce, warehouse and production databases to show a consistent view of the customer. 

For restaurants, we’re improving accessibility for our tools across platforms, with an emphasis on mobile. We’re also building out new functionality for the restaurant platform that will give restaurants a differentiated reason to join Snackpass.

What company north star do all these initiatives tie into? 

That’s a great question, at a company level, we want to: 1. Ensure our launches in cities go smoothly, and 2. Grow from just a marketplace and being responsible for 20-30% of their revenue to owning a restaurant’s entire revenue stream. We are really there to be the restaurant’s operational best friend, not just a demand aggregator.

What tools and processes are you using today?

On the data front, we’re excited for our head of data eng to start because right now someone who has good data intuition just looks at the data, can tell it’s off and then we dig further. 

We mostly use Looker for business intelligence, and we have some analytics in Mixpanel. There’s pros and cons for each. It’s on my mental roadmap to consolidate those two one day, but that day is far away. 

How does Snackpass think about build vs buy decisions? 

We’re a pretty buy-oriented company, once we see a pattern emerging. For example, let’s take alerting around Snackpass kiosks that go down at restaurants. We used to deal with ad hoc requests every 1-2 months, outside of DataDog system alerts. We’d fix that with Looker’s alerts which were restrictive, or writing cron jobs which needed engineering. But when the volume of requests increases or if there’s a specific problem to fix, we’d actually take the action to investigate tools that can help.

We need to both fix root causes and invest in better alerting

How do you think about investing in better alerting vs fixing the root cause?

I’ll continue with the example around what happens when Snackpass kiosks are down. It’s important for us to know asap when restaurants are facing an issue, even more than knowing why it happened or automating the remediation. Two thirds of our customers won’t even tell us they’re facing the issue, and then churn. We could reduce that by either improving kiosk reliability, or better alerts. For every hardware issue we fix, we might see ten others down the line. So, the right approach is actually a mix: we need to both fix root causes and invest in better alerting.

What are you most excited for? 

We’re launching a really exciting platform for our restaurants soon. Stay tuned!

What are you personally learning these days? 

I’ve been digging into the rabbit hole of payments these days! There’s a lot of philosophies on whether banks or credit cards or neither will be the center of payments, and I’m excited to see how that shakes out. 

‍

Find out how fast-growing companies are using LogicLoop to set up business alerts & automations without engineers. Start in 5 minutes!