Discover more from The Analytics Engineering Roundup
Ep 46: dbt Labs on dbt (w/ Daniel Le)
Daniel Le is the CFO at dbt Labs where he has built multiple teams. He is also the former head of FP&A and operations at Zoom, and he helped scale FP&A as the former finance director at Okta.
In this conversation with Julia, Daniel shares his view as CFO on the challenges SaaS companies face and the importance of finance teams creating a holistic view of their business. Daniel gives advice to data leaders about how they can automate business processes with dbt Cloud and use self-service analytics to automate revenue recognition, generate consistent headcount analytics, and more across the organization. Read more about Daniel’s story here.
Listen & subscribe from:
Key points from Daniel Le in this episode:
What was it like on the finance team doing some of the operational work at Zoom or Okta that was really hard?
When I first left hardware and went to SaaS, everyone said SaaS is new and difficult. You have to know gross retention, net retention, you have to do cohort analysis.
Look at every single SaaS company, look at their S1. What are the three metrics? Gross retention, nets, and churn, which is a variable of the above. You have cohort analysis, basically, what is the lifetime value of a customer versus the amount of time? So, there's that bar chart that every banker would have to put together.
But the reality is that getting that data would require you to download your source data or your customer data. You have to do it via spreadsheets. We would have to get it from the source. And sometimes the source is not fully accurate because customers or a definition of a customer is different depending on what source system that you use.
Having to reconcile that and then having to then build these cohort analyses and this massive Excel file that typically freezes or breaks. At Zoom, we had a data team, but then I would partner with the data team and we would have to get the data at a Redshift, and they would have to write these manual queries.
We would build these cohort analyses, and we would finally get into a system, and then when a pipeline would change or some data set would change, it would break, and it'd take a couple of days to reconcile that. Even leading up to the IPO, we would have to do our S1s, and we were like "Okay, we're about to go public, and now we're a public company. We actually have to do it in a consistent way." Then, it would break every so often, and as data analysts would rotate to different areas of the business, getting it in a consistent, scalable way was very challenging.
I had Mark Matteucci, who was my go-to-market FP&A person, he would then learn SQL himself and partner with the data team to figure out how to do self-service. He would have to write these long queries. Luckily, Mark's background was data. So we were able to figure out how to scale that.
But every time something broke, it would take a couple of days for us to figure that out. As our data team scaled at Zoom and it got even more complex, as the business would grow, when we tried to do the analysis ourselves in Excel, it would break because Zoom grew from a couple million to hundreds of millions of users. It ended up breaking our Excel files and some of our queries would have to change. Doing that in a repeatable, consistent way was virtually impossible.
What advice would you give to a data leader about how they can they work best with their CFO and finance teams?
I'm starting to see the role of the CFO evolve.
You have to nail the basic foundations. But in my role here, it's a little bit broad. I own operations, people, legal, security, IT, and then more recently, the data team as well. I've always seen myself as somebody that is here to help the business and to help solve problems. I don't limit myself within my own role. When I was at Zoom, Kelly Steckelberg, Zoom's current CFO, allowed me to get into different areas rather than just doing purely FP&A. Our team would own quotas and we would do both the bottom-up and top-down bookings plan and partner with the go-to-market leaders on those types of analysis.
And I got to partner with Hillary Headlee, who was our sales ops leader at that time, and then I got into our offshore team, and I also ran our procurement team and partnered with our AWS and data center teams. I've always had the mindset of helping solve problems across the company and how to not just limit yourself within your current scope.
If I'm a data leader, I know that data is extremely important to make business decisions. I think that in every role that I've had, I have reevaluated what I've done every single quarter. When the quarter ends, I look back and say "What are my key accomplishments? Have I gotten better?"
I look back at myself 10 plus years ago, and I cringe a little bit because I'm no longer the person that I was, but I feel like I've grown and I've adapted a lot. That's how I always think about things.
I've been at dbt labs for a little bit under two years. We didn't have a lot of these teams that were under me. I've been grateful to be able to bring on some amazing leaders on my teams and we've been able to continue to build a great foundation and a great team. We have to nail our basics obviously, but how do you make sure that you continue to solve business problems? No matter what role you're in, whether you're in data, whether you're in product, whether you're in engineering, you should operate as a business partner in the sense that you have key stakeholders that you work with.
Unless you're very technical and have a very narrow focus, generally you have to interact with more people. How do you make people's lives easier? How do you solve complex problems for the company? But most importantly, how do you continuously improve yourself and the people around you on a consistent basis? That's key.
What would you say is one of the biggest benefits of data being aligned really closely with finance?
I don't know if there's a perfect formula, but I always felt that I have partnered closely with the data team in my roles. Maybe it's because of my background. Early on in my career, I would be one of the few people that would actually partner and have access to our data. I've always felt that data was extremely important to have insights in terms of running the business and as somebody that has led finance teams and continues to lead finance teams, I can say we do a lot of the analytics.
I always felt that the roles and responsibilities of the data team closely align with my finance team. Making sure that we're all marching in the right direction has always been beneficial. Where I've seen it not work is when there are different priorities and a different direction. You never want your finance and ops and data teams to be operating in silos because, ultimately, we're all working to help solve problems across the business, and similar to how we centralize headcount and now we're doing revenue, we want everybody to be looking at things the same way.
I think that in a world in which I've always been operationally data-focused, aligning with finance here was just a natural progression of what I've done throughout my entire career.
Looking 10 years out, what do you hope will be true for the data industry?
I don't know if it's hope, I see it and I breathe it every day, but I would say that we're barely scratching the surface of all the different use cases that we can do with data. I think we've graduated from a phase in which we can store the data and then let's do some analysis.
But, I would say that the analysis that is being done with data is fairly basic today. You talk to finance teams today, they have some access to the data and they're running analysis. But I think folks are doing some basic analytics. I see a world in which there's real-time analysis where we're asked to make decisions faster than we've ever done before.
Just 15 years ago, when I was at KLA, I was pulling data that was a couple of months old, only had a small percentage of the data, and then I was making forward-looking decisions off of that. Now we're in a world where we can actually feed all the data in there. We're leveraging dbt Cloud to actually explore what we can do.
I feel like we're in the early innings of all the different use cases. There are a lot of finance teams that are still doing things in a manual way. One of our goals is to continue to speed up our closes. We close in six business days — a big shout-out to the finance team. We actually improved our close this month to five days. My goal is to get it to one or two days over time.
I don't think I've been at a company, leading up to an IPO that has done it in that timeline. I feel like that should be a baseline going forwards. That will allow us to support the business and innovate even faster.