How to tell data stories with Peter Walker, Head of Insights at Carta
Welcome to the Canvas Podcast, where we bring business and data leaders to talk about how to make data easier for everyone. Today, I am super excited to have Peter Walker, Head of Insights at Carta on the show.
Peter, thanks so much for coming on. Do you want to start by telling us about yourself?
So I've been in startups, basically my whole career. I started at a media intelligence startup back east out of DC called PublicRelay, and built a couple of data teams there especially focused on Tableau analytics and actually embedding Tableau visualizations within our product.
I got a little burnt out on the data side. So I went over to the marketing team, and ended up running the marketing department at this startup as well. I then ran a data visualization team at the COVID tracking project at the Atlantic, during the pandemic, which was an amazing experience. And then I've been at Carta for about 18 months now in this insights position where my whole job is to make Carta data more useful.
What initially attracted you to data in the first place?
So in a roundabout fashion, to be honest, I wasn't a data-centric person during my university career and my degrees in history in PoliSci. But I knew I wanted to join startups and I tried to find something in college with a couple of buddies and then that fizzled out and I was looking around at startups.
It seemed as though getting a little bit of SQL skill is a great way to uplevel yourself when you're trying to get hired in those early positions. So I just took a crash course in that, and that proved really useful. I think it pushed forward though and became more of a passion when I saw how many business users, especially at early-stage startups, just have no data experience at all, and the way that holds you back without being able to touch that side of the business.
It pushed me to move forward and I've dug more deeply into the visualization and business intelligence layer. But certainly, I could imagine the utility of going super deep into infrastructure and other data platforms as well.
What was that data experience like and what prompted the move to go work on the operator side of the house?
I doubt that this is a completely unique story, but at a growing startup that's trying to do a hundred things but they only have resources for 50. It feels sometimes like the analytics teams bear the brunt of that. And quite honestly, I think some of it has to do with the way they intake requests.
Like literally just looking at a stream of data requests that are coming through Sack or are inputted into a Google form or something it's so it was impersonal at some point and it felt like I was running around being a little robot, building things for other people instead of getting to look broadly at business goals.
And so it was great that my CEO at the time was like I hear you - come and try your hand at something else. And in my case, it was product marketing for a while. And I really got jazzed up on that. So that sort of kept the passion going for this startup for quite a while.
I can imagine a lot of people nodding along saying yeah. Being in the queue all day, every day, just answering data requests from across the company, it feels difficult to see the forest through the trees.
What were some of the learnings that you took coming into an operator role as a data practitioner and what were some of the common threads between the roles?
Yeah. I think that starting out in an analytics role and then spending a lot of time on the marketing side of the house, there are some pretty clear commonalities.
You're both beholden to the overall goals of the business. Sometimes you feel like you're in a less strategic function occasionally, but the interesting part about having that data piece under your belt, when you go to an operator role, is it just allows you to move much more quickly on your day to day task. So for instance, if you are a marketing team and you're trying to put together a personalization pitch for this big campaign that you're working on.
If you know SQL you can pull all of the user information that you've had over the last six months and look over that data yourself and formulate your hypothesis without having to wait for someone to pull the data for you, and then go back and forth on whether or not it's clean or presented in the right way.
At Carta, I'm working with an amazing group of marketers that are really data-centric a lot of the time, but oftentimes I will hear marketers say we wanna spend more time on persona building, or copy, or content, or big strategic ideas that have nothing to do with data.
But then they're always beholden to other teams to bring that data to the forefront for them. I expect that certainly SQL and then maybe some other data forward capabilities are going to become part and parcel to a lot of different roles across the business.
Maybe it's actually typing that SQL in, or maybe it's being able to leverage low-code tools so that you can avoid having to do SQL yourself, but either way being able to pull your own data and understanding the relationships between data tables just makes you so much faster. And because you're faster, you get to do more projects.
And because you get to do more projects, you become more experienced at a faster growth rate. So I couldn't recommend it highly enough. I wish honestly, I learned SQL even earlier in my career.
You have a super interesting role at Carta. Can you tell us about it?
Definitely. I'm jazzed about what we're doing at Carta every day. My role as Head of Insights is I have a little team within the marketing department. That whole goal is to make Carta data more useful.
So what do we mean by that? It takes two separate forms. First and foremost, we get to take all of the wonderful work that our data science team is doing. And we get to showcase the data set that Carta has access to to the wider ecosystem and hopefully we provide some value for that along the way.
So a great example of that is we just released our compensation report for the first half of 2022. That's data from two and a half thousand startups or so about what they are compensating right now. How much of your payroll goes to which functions? How is remote work impacting hiring?
We have 30,000 or so cap tables on our platform. So this kind of reporting and data sheet structures are things that we've been trying to pump out into the world. But the second phase of that is what I'm really interested in, which is how do we make sure that Carta data is helping offline conversations?
It's one thing to put out a report and get excited, because people are sharing it on Twitter. It's quite another to make sure that Carta data is being referenced where it should be. So say you're a series A company and you're looking to raise your next round. We wanted to make sure that Carta is there to help you with the benchmarks you need for your industry, stage, and the kinds of companies like you who have been raising lately.
And so making sure that Carta becomes known as this data source is my overall purpose at the company. And then there's a lot of different side projects of helping marketing become more data-centric and utilizing data as a business user across the business. So I sometimes feel like I get to be a cheerleader for our data teams and they've been all this work building these pristine models.
And then I come in and add this like a thin layer of visuals and people are like, oh, that's awesome. Now we can go and use this data, but really it's their work that has been the whole pyramid behind it. So it's been an awesome time and Carta is a really cool place to be at the moment.
What makes your data visualizations shine? What are some tips and tricks you can share?
Yeah. I think this is where a lot of data practitioners will focus on just the immediate MVP use case of a chart. So as I mentioned, I think I've indexed most highly in my career on business intelligence tools and then data visualization as a subset of all of the data practitioner work that you can do. And one advantage of being in marketing is that you get this better feedback loop from customers about what they're actually responding to.
And often times it seems pedantic, but it's just because the chart isn't good. Or the chart is done in the wrong way, or the chart is confusing or the accesses don't make enough sense to stand out amongst all of the data content out there. It really isn't that hard, as long as you're willing to put in the time to build good-looking charts and then keep the reader’s interest with maybe a variety of different visualizations and understanding where they're best used, but something that I've seen certainly within Carta and then other startups as well.
For instance, if you're building an internal deck and you're taking charts from Excel or Google Sheets, spend a little time with those charts to make sure they're clearly labeled, and change the colors to something that makes sense to uplevel the branding just a little bit.
This can really pay dividends because people will actually focus on the thing that you want them to focus on, which is the takeaway from that piece of data. Oftentimes like what I'm doing I'll build, say a visual and Tableau, and then I'll export that visual from Tableau into Figma so that I can get a little bit more precise with the layout of whatever that visualization ends up being.
And I honestly find those 10 or 15 minutes that I'm spending totally change the responses I get from both execs and external audiences. It really is worth it having great-looking charts as simple as that sounds.
I think that just to double down on that for a second, the exploratory part of our work that I get to do basically every day is some of the most fun.
And honestly, I think it's some of the most value-add that I can provide. For the data team, for instance, it's always gonna be their third or fourth priority to go through this model and pull out different parts of it and try different charts and see what looks good and what would be easily explainable to an external audience and all that kind of stuff.
It's not the best use of their time, but when we're trying to make a point to broad audience - be it founders externally, HR leaders, or whoever's reading this, we have to assume very little about their knowledge of our dataset and also of this data in general. So the more that we can spend time building the right charts at the right cadence hopefully and eventually also add some interactivity to make them feel a part of it.
That's a really useful time well spent for us. And so again, this is like a marriage between marketing and data that has been part of my career. If you're a marketer and you're listening to this - go and try your hand at some of these data visualization tools. I think you'll pick them up pretty quickly and also you'll be able to speak to the business teams in a way that makes a lot of sense to them.
When does the relationship between business and data teams work best and when does it fall apart?
It's pretty easy at times as an operator to treat data teams as a resource that you just either request out of or that you are basically a customer of. And you're just either going to be really demanding for all of these little things that they on the data side might think you'd be able to self-service or you basically just take what the data team gives you.
So when I've seen these relationships work best is when they're strong relationships, not just at the senior executive level, but all the way down the teams so that marketers or biz ops or other operations functions can go to the data teams and involve them early on in projects and say, “Hey, we are looking to redo the way we do attribution as a marketing team. Can you be a thought partner with us to walk through all of the different models we have to how to best integrate different systems?”
Just involving them early on and not just coming to them with a list of demands is a much better way to work. And then on the data team side, I think oftentimes data teams don't do a wonderful job of telling their own stories.
Like it's difficult for a person who knows nothing about data to understand what a data scientist does. So I love the fact that at Carta, we do these things called show and tells where different parts of the organization will come up and just walk you through a project that they recently did walk you through the outcomes, what went well, what didn't and when the data team does that for Carta, I always learn something new about, oh, that's how they're thinking about the business.
This is how they're trying to connect these data sets on the back-end. It's incredibly helpful. And I think it's something that I would take to any other company that I've been at. Allowing data teams to talk about their work day to day more broadly and more often to people who ingest their in products, I think is super helpful.
We can get into the way that business intelligence tools like Looker and Tableau intermediate between these teams and the way that's useful and the way that it's not, but when it comes to personal relationships that's some of my best advice.
You worked on this amazing COVID project for the Atlantic. Can you tell us about that?
It was just a surreal experience, honestly. Everyone remembers where they were in late March, early, April 2020 when everything was shutting down. I did as many data nerds would just go looking for data sets to visualize and Tableau during that time, like what's the best stuff that we know about what's going on with the pandemic.
And I stumbled across this data set of COVID tests. That's being run by two reporters outta the Atlantic Robinson Meyer and Alexis magical. And they. Are forming this little thing called the COVID tracking product. And I'm like, oh, this is cool. This will be useful to get just a steady stream of COVID data that I can then mess around with in Tableau.
I start putting all these charts that I'm building and Tableau stuff online, and it starts picking up, people start referencing it daily and I'm like, oh wow. Now I feel like I have a responsibility to keep doing that. Then they actually reached out and said, Hey, we'd love to bring some of this expertise in-house, cuz they were trying to put out more and more charts.
And it just, it was very clear that what you wanted at the beginning of this pandemic was charts about change over time, how cases are going, deaths, tests, positivity rates, all the things that we're super familiar with. Now having three years in this pandemic.
So I came onto the project. It's an entirely volunteer project, save for 3-10 paid staffers from the Atlantic. And it's a 400-person slack that no one has ever met each other. We're all in different parts of the US across the world, honestly. And it's just this, such an amazing spirit within the chat. Like we've all talked about and probably, hopefully, some of us have had experiences where you feel super committed to the work that you're doing, but this was totally different.
You were barely going outside. It was basically all anyone was talking about and you felt that you were providing data to the wider world and to America as a whole. So it was a pretty amazing sort of motivation that you felt every day. And the nice part that I could bring to this was they had wonderful tech teams running the back ends and great architects about building infrastructure and dbt projects and all these kinds of places where we're hosting data, what they needed was how can we create 40 to 50 charts a day that we can then run through super quickly and pick whichever ones are the most interesting and say, this is today's story and put it up on our site and put it up on Twitter and let everybody know - here's what you need to know. If you can't follow all of the headlines, like here are the two or three charts about what's changing most quickly.
And I'll give a little plug here to Tableau, like that's where this product was incredibly helpful. Because you can build and iterate very quickly with the drag and drop model. So it allows even non-technical people to be very quick at chart building. So I built a little team there and in the end the tracking project just an amazing feeling of the impact that I don't know if I'm gonna match again in corporate America, but I was so happy and thankful that they brought me on so that I could help in a small way there. It was amazing to meet some of them afterwards, too.
The other part of that project that I thought was so interesting was, and this is something I've tried to take to Carta as well, was that instead of saying, we're gonna raise a bunch of money from philanthropists and keep this going. It was actually designed to shut down as soon as we felt like the data that was being put out by the government was good enough. So it lasted for about a year. And then we were like, look, the CDC has good data right now. We're ready to close it.
I love that orientation. I haven't found it very often, but I think about it within a company as well as saying, what is the quickest path or what is the part of this business that we need to build?
We’re either automating ourselves out of utility, or we're only gonna spend this up for a certain amount of time and then we're gonna kill it once it's served its purpose. I feel like there are a lot of zombie things that can exist both within companies or other organizations. So that mentality of like, how do we find a way to automate ourselves out is a really useful one.
Speaking of changing market conditions and remote work - with the reporting that you're doing for Carta, what's been surprising with the recent insights that y'all have been pushing out?
I think probably well there are things that sort of match the headlines, definitely layoffs that startups are up to and it is getting more difficult to raise rounds. The venture community is not in this heyday of 2021. What's been interesting though, is within those stories, there's still real strength within the venture community.
I think I was just looking at this morning, like Seed and Series A valuations within Carta data are down 5% maybe. There are fewer rounds happening at those valuations. So certainly funding is harder. But I don't think there's been quite as much retrenchment as perhaps you've seen in the public markets. So that's a note of hopefulness within our compensation data. There's such cool stuff happening when they're on remote work right now.
And there are all these interesting compensation questions about where you build a team. Do you compensate them equally regardless of where they live? Are you trying to geolocate salaries? Are you allowing people now, who have been at a core headquarters to go off and live elsewhere? And our data is pretty inconclusive.
Everyone's applying a different strategy, but I'm excited because I think what'll happen over the next couple of years is companies themselves will take on an orientation. Like we are an in-person company, or we are a fully distributed company, or we're this kind of company.
And then workers with tech talent will allow themselves to just discriminate based on their own preferences and say, this is the thing I value most in a work environment. So I'm in an index on these companies and I'm excited about having that choice for more and more people within tech, but then hopefully beyond tech.
How can people learn more about you?
So if you're into startups, a VC investor, or even just a startup employee, we've got a lot of great data to help educate you on what's going on in the market at the moment.