[Auren Hoffman] Welcome to World of DaaS, a show for data enthusiasts. I'm your host, Auren Hoffman, CEO of SafeGraph. For more conversations, videos, and transcripts, visit safegraph.com/podcasts.

‍Hello, fellow data nerds. My guest today is Will Marshall. Will is the co- founder and CEO of Planet, a company that owns and operates hundreds of satellites, producing images of the earth. Planet has over 100 million in revenue and from over 600 customers. Welcome to World of DaaS, Will. 

[Will Marshall] Hey, pleased to be here. Awesome. 

[Auren Hoffman] Now Planet first launched its satellites in 2014. And within four years, you were collecting images from the entire world basically. To me, this seems incredibly fast. What's the non-obvious thing that made this possible? 

[Will Marshall] Well, I would say the answer to that question is agile aerospace.

[Auren Hoffman] What's agile aerospace?

[Will Marshall] You've probably heard of agile software, which means release early and often, test out the software with customers or in a sandbox, and then iterate as quickly as possible. That leads to fast development. We've done the same with aerospace. So over the last five years, we've averaged four launches a year, roughly every three months or so as they don't often spread out evenly like that. But roughly every three months on average, we've got a new launch. And we put up a late next generation of satellites in each one. 

[Auren Hoffman] How many satellites per launch?

[Will Marshall] That ranges from a couple to 88. That one was a world record, but generally, think about 20. And when we launched 20, we also put up a next generation sensor or a next generation radio and a next generation hard drive.

[Auren Hoffman] You're testing it out and seeing if it works.

[Will Marshall] Yes, and the following fleet will be made of all of those satellites. Also on the software domain, we're constantly updating our software, you'd be surprised about that. We have interns that come to our company, and we give them the keys to the satellites and they're like “What? You just don't upload software to all of them at once and break them all”. It's pretty hard to break our satellites, but you know, unlike something on the ground, you can't go up and press the reset button. Generally, we iterate the software on board really rapidly too. Every week we're uploading new software, making the satellites more efficient, both in terms of the operations and eventually processing of the imagery up there. So you can imagine how we are increasingly turning space, more like the software domain, and certainly, most importantly, bringing space up to date with Moore's law. Taking all the latest generation in the electronics that are in your phones and in your computer and other consumer electronics, and putting that in and and by doing that keeping space strapped to Moore's Law, as I call it.

[Auren Hoffman] So you have 400 satellites up there, that actually could be 400 different satellites with slightly different software and different components and you just keep iterating to get better and better. Is that right?

[Will Marshall]  Yeah, sure. I mean, we do it in fleet. Maybe a bit more like if you get a few 100 people together, they have slightly different models of iPhones in their pocket or Android phones. But yes, sure. There are different models and it's our job to make sure it's backwards compatible. So if you ever used data from our first satellites, you can use it from the latest satellites, but it's also our job to constantly iterate. And again, just like you don't want a three year old phone in your pocket, you don't want a three year old satellite in space because it becomes obsolete. Our next generation satellites that we launched last year produced five times more data than the previous generation per day per satellite, roughly at the same cost. Our job is to make more capability in that box.

[Auren Hoffman] Okay, that makes sense. Now, I see Planet as a data company, not an analytics company and not a space company. I see your data is kind of these persistent images of the earth. And then of course, you have this proprietary way of getting the data, you launch satellites. Do you see it similar, or do you see a different way of describing Planet as a company?

[Will Marshall] No, exactly the same. Of course, we have satellites and we are a satellite company, we're of course a data company, and we're also a software company as we build a lot of software to process all the data. But the most important thing is that we service data to our clients. We service imagery and analytics to agriculture clients, we service that to mapping clients, to government clients. That's what our consumers and our clients use, right? They don't buy satellites, they don't buy software, they buy data and analytics feeds. I actually liken it to Bloomberg. Bloomberg terminal services data for financial services, of course. The great thing about Bloomberg is that margins are high, because when you sell a data feed a second time, that costs are very low, and it's high stickiness, because your data is embedded into their workflow. The only difference is twofold. One is that we have a proprietary data set that mainly aggregates public source financial data. We obviously have a unique 200 satellite fleet that images the whole planet every day, that's a unique data set. Not only is it hard to get that data set, because you have to launch all those satellites, but you can't go back in time. Years of data, even if you were on that system. And actually, that archive of data is what enables us to build machine learning on top, which is incredibly important for analytics. And the specific difference with Bloomberg is that we service many different vertical markets, right? So we think of it Bloomberg++, but Bloomberg is a good way of thinking about it. People do get confused about our business model. They will get very excited about the satellites, we're really a data company. When people get confused about that I say, if you think of Planet as a satellite company, then you should think of Google as a server company. Because yes, they have lots of servers, and that's the hardware back end. But just like servers are back end to Google, satellites are back end to us. Our satellites are a little bit more sexy than Google's servers. Some people do get excited about them. I do too, I'm a space geek. But in the end, the data services are much more like a Bloomberg. 

[Auren Hoffman] If I’m a data consumer, let's say I'm a data scientist, one of the hard things about using imagery data is that the data is very large. How has the sophistication of your customers evolved over the last few years to use that type of data?

[Will Marshall] It's a great question. This is a really fascinating point. I mean, we have 25 terabytes a day of imagery, over 3-4 million images a day from the satellites, that is a vast data set. That has always value, starting and helping people recover from floods in Germany right now, like people using it for security and finding out new threats, like where's all these fleets of ships, protecting coral reefs, improving agriculture. But if I'm real, a lot of those users can't yet get value from the data because it's so much. Google gets value out of it for maps, because they have sophisticated geospatial teams, big ag companies get value out of it, because they know how to process satellite imagery. Governments get value out of it because they have big teams doing that. But the smaller insurance company, the little hedge fund, the developer can't. So we're working hard to build the analytics to go up the stack to help that. We just hired Kevin Weil, and he was the lead product for Twitter, and then Instagram. And at Twitter, he uses an analogy that they had the data firehose, the Twitter firehose, right? And they were selling it to people and they thought this was the best thing since sliced bread. And yes, once again, Google got value out of it, and governments got value out of it. You can't get value out of 500 million tweets a day in 100 different languages, and so on. What they started doing is adding certain analytics like, “okay, we'll just tell you how often your brand as a company has been mentioned and the sentiment around that brand”. That was something useful to many more people. So somewhat analogously, we have a firehose of data coming down from our satellites. As we add certain analytics, many more people will get access and value from that data. And so that's what we're building. So with time, I think it will move away from the geospatial experts, enterprises, to the folks with just data science, and then maybe to the folks with not even data science.

[Auren Hoffman] At the same time you're adding all these analytics, there does seem to be another motion going on in a macro environment in that these data scientists are becoming more and more lethal, because they have more and more tools around them. So the data scientists today even those that may not be able to use your data, just because the tools around them are getting so much better and so much more powerful. In three years, they may be able to use your data. And so you're going to have a larger group of buyers regardless, even if you didn't go into analytics. Do you see it the same way?

[Will Marshall] Definitely. I mean, look what has happened in computer vision in the last five years, it's incredible, right? Computer vision has been one of the most successful areas of machine learning, extracting information out of imagery. It was a pioneer for pulling out cats and dogs from pictures you put on Facebook, or YouTube, or whatever. But the same technology underlying those neural networks can be used to extract out roads and buildings, or ships and planes, or trains or trees, whatever it is out of our imagery. In fact, I gave a little TED talk about this, where I likened it to building a Google index of what’s on the internet and made it searchable. Basically, with machine learning and computer vision, we can extract out all the objects in every picture. If we do that for the whole world and do that every day, we basically build up a database of where everything is on the earth over time and that is then searchable. A bit like Google has figured out how to search the internet, we're figuring out how to search the physical planet.

[Auren Hoffman] And even if you don't do that, you could have customers that do it or partners that do it. Okay, interesting. Now, not knowing that much about your market, if I was going to try to create a competitor to Planet, let's spitball for a second, I would try to insert low flying planes and drones with a camera to try to collect data continuously. Why would my crazy scheme not work?

[Will Marshall] Planes are tricky because of area codes. So planes travel a couple 100 miles an hour, whatever they do. That's about 100 times slower than our satellites. If you look at the swath width, they cover about 10 times less at least. So that's 1000 times less imagery per day, per plane. So to cover the same area, you're going to have to have 100,000 planes or 200,000 planes to cover the same areas as 200 satellites? That's a lot of planes. Separately, you're gonna have to go and get permission to fly those planes in every jurisdiction: in every county, every city, every country.

[Auren Hoffman] North Korea is probably not allowing me to fly my plane there.

[Will Marshall] Nor will China, nor will India, or many other places, whereas with space we are actually outside the territory of countries. That territorial airspace goes up to 100 kilometers, and thereafter its space. It became international law that anyone can fly over anyone's territory and take a picture from space. It's partly because you can't very well fly out to Russia and turn left, you're gonna go over it. They can't just say you can't take a picture of that. The US and Russia during the Cold War said “well, okay, we’ll let each other do that, we only let each other fly planes” and then it became international law that anyone can fly a satellite. That also means North Korea could, in principle, fly a satellite over the US and there's nothing to stop them from doing that. So we're allowed to do that, whereas with your plane idea, you'd have to go and get permission from everyone. 

[Auren Hoffman] Okay, got it. Okay, that does seem quite hard. Okay. Now, okay, you sell it...? 

[Will Marshall] ...expensive? Funny point, people think “well, aerial imagery or drones, it's gonna be much cheaper”. Yes, the actual cost of the drone is cheaper than the cost of the satellite. But when you factor in the area coverage, the unit area, our imagery is far cheaper than collecting it by drones. So for example, drones for agriculture makes no sense to me. In almost all agricultural cases, our resolution suffices and enables precision agriculture. And you know, the cost wins because we're much much lower. There are certain applications of drones for imagery like Buildings Science where you need to know brick by brick. We haven't got the resolution, so it might make sense to use drones, but you have to think about the application. Certainly if it's wide area coverage, satellites are pretty good. 

[Auren Hoffman] Absolutely. Now when you sell really valuable data, but of course, there's lots of other data like in the geospatial data world, like SafeGraph where I work sells geospatial data, there's companies like CoStar and Corelogic that have used spatial data, how do you see yourself partnering with these hundreds of other companies that also provide data?

[Will Marshall] Well, I want to collaborate, I want to mix and match our data. We launched last year our first data fusion product that married our data with the Landsat data and the Sentinel data from NASA and the European Space engine, respectively. And it gave the sort of radiometric accuracy of those satellites, which is really good at radiometric accuracy, like the only photons turning down the electrons, CCD, and then accurately merging. And yet, with our resolution, which is much better than those systems, and temporal resolution, which is much better. And now we have this color balanced, radiometrically accurate daily image, that's cloud free at three meter resolution using those. There's one plus one equals three, or one plus one plus one equals five in this thing, but it's always synergetic, you get more value than the sum of the parts. So I'm excited about merging with other geospatial or even non geospatial data sets.

[Auren Hoffman] The great thing about combining data is the value of all the data sets goes up as they get joined together. That's why these like join keys on the data are so important. A similar question for all these analytics companies, you're doing analytics, but then there's also companies like Orbital Insight, Data Robot, there's even like Accenture and Booz Allen, how do you see these companies and working with them? Do you see them as partners? Do you see them as competition? How does that work long term?

[Will Marshall] I very much see them as partners. I can imagine an ecosystem of folks building apps and services on top of Planet’s data. There's 1000s of applications of our data. There are pretty mainstream applications of our data, but there's a long tail, and each requires bespoke applications, maybe some novel data, industry knowledge, and we need to package them together to service those vertical markets. And we want to work together with folks that innovate on top of that data. Planet is not a monopoly over all the best ideas. And we want to see that happen. We work with Orbital Insight, we work with those companies to get access to our data to enable them to add value for their users. We think it is absolutely helping to build the market. The market opportunity is huge here, we don't need to think about this too competitively, because there's just so much. We see Planet as being part of the two biggest transformations in the global economy, this transformation to a digital economy, and sustainable economy. So digital transformation is where industries like agriculture, where we improve the crop yield by 20% or 40%, by doing precision agriculture, and I can explain that a bit more in detail if you want. But like every industry, Big Data and AI is hitting them, and they're enabling them to become more efficient, and 20-40% improvements in an industry of that sort of scale, whether that's agriculture, or transportation, or forest management, or civil government is huge, right? It’s a multi trillion dollar transition. And then it's a transition to a sustainable economy, which is also a multi trillion dollar transition as all the governments and companies of the world are trying to measure their emissions and measure their environmental or ESG targets. They have to measure natural capital. Basically as a global economy, especially as we're rebooting out of COVID trying to reboot into a sustainable economy. The first thing you have to do is measure natural capital. We've been presuming it's free, trees are free, we can come down now and don't have to pay any tree any money. We just cut it down. We assume we can just pollute stuff into the atmosphere into the rivers at no cost and that obviously has to stop. It's completely unsustainable, and it's going to kill us if we don't stop.

[Auren Hoffman] First, what are the benefits of Planet? First, it kind of has to be measured in a way, right? 

[Will Marshall] That’s my point. With this digital transition and the sustainable transition, the first step is to measure this stuff, you cannot manage what you don't measure. You just can't. We have the data that is pretty foundational to helping those two transformations. I'm not saying it's the “be-all end-all” data set that solves all of the world's problems, but it's pretty foundational to assisting those to multi trillion dollar transition. 

[Auren Hoffman] Back to your business, we met at your office a couple years ago. And afterwards, I took a tour of how you’re building all these satellites. It's super cool. And then I came out of there and I'm like “oh, man, this is totally unfair, these guys have a totally unfair advantage when they're recruiting people”. How is the coolness factor that helped you recruit talent?

[Will Marshall] People do love space. And across the spectrum, whether it's people in our sales and marketing to, of course, that people are building and designing the satellites, and everyone in between. So we can't just rely on that. I mean, we care about giving fair wages to people, it's a bit hard to compete with the likes of Google and Facebook and Apple here in the Bay Area. But we do hire with competitive salaries in relevant verticals, markets and spaces. And we just don't take that for granted. We care about our employees. But yeah, for sure it's an attractor. So we've never had a big problem hiring. Hiring hasn't been our biggest challenge.

[Auren Hoffman] Well, that's great. I mean, you're probably the only company that can say that, so that's really interesting. Now data companies generally have a lot of trouble attracting a Series A, Series B, because in the beginning, the margins look really bad, because it's like you'd have this fixed cost that the data that in some ways that cost shows up in your COGS. In Planet's case, their startup costs were even higher than most data companies. What advice would you give to other companies who have these high fixed costs, create core data in telling that story?

[Will Marshall] Well, speak to the right venture capitalists. Not many of them understand that. But actually to your gross margins point, we have very high gross margins. With the incremental cost of selling that image to the second customer, is close to a very round number. It's just the compute cost of serving the customer, it's almost nothing. So the irony is, yes, it costs a lot to put out costs. Yeah, that's a fixed cost, and thereafter, it's gravy. It's more gravy than you think you might go “well, it's like AWS, per the servers”. But actually, any particular compute instance, can only be used once. Whereas an image can be distributed to 1000s of users, and almost no incremental cost. Actually, it's much better than an AWS type situation. In terms of margins, data businesses are much better. I think they're fantastic businesses. You can look at it as an expensive CapEx or you can look at it as a high barrier to entry, and they are really great margins. 

[Auren Hoffman] There is a story that you have to tell in the beginning that people have to kind of come on board on this vision, right? 

[Will Marshall] Yep. 

[Auren Hoffman] All right. Interesting. Now, okay, speaking of data businesses, I admire ZoomInfo, Henry Schuck, the CEO was on World of DaaS recently, and I noticed that ZoomInfo was actually featured in Planet's recent investor presentation. Now, outside of those two companies, there aren't a ton of data companies that have gone public recently. What companies do you admire? What do you learn from, what do you look to as you're building your business?

[Will Marshall] The company that I look to the most is actually Bloomberg. I mentioned them already, that analogy that they're providing data feeds and those data feeds are pairing customers to make smarter decisions. I think that's what Planet is doing, fundamentally. It's a very different backend, but the frontend actually doesn't look so dissimilar. And I certainly think the margins and the growth and the stickiness are analogous. Anyway, there's no perfect analogy, and of course, they didn't recently go public. 

[Auren Hoffman] They started 40 years ago, right?

[Will Marshall] Yeah. But that's not the be all and end all of company of course. It's just a milestone and they chose not to do it that way. But I think they're the best company for us to admire. 

[Auren Hoffman] Interesting. Okay, cool. Now, you're a student of data businesses. We actually met because you sent me a cold email, wanting to discuss the DaaS Bible blog that I wrote. Not many CEOs send cold emails out to random bloggers. What other resources have you found valuable in helping you create such a successful data company?

[Will Marshall] Well, apparently yours. I look around, I mainly get advice from advisors. I mean, I've seen your blog, I do look at blogs and try to read books and articles that are relevant. To the point of many of the things that you've said that data as a subscription service are really rare businesses in comparison with SaaS, everyone knows that. We have these discussions in our boardroom about this, because people keep on pushing us to that cookie cutter of a SaaS, because that's what most people know. It works, but actually, there's differences with DaaS companies, which is, for example, they much more lend themselves to be horizontal players that are relevant to lots of vertical markets, whereas SaaS companies, you are more likely to be in one vertical market.

[Auren Hoffman] Your SaaS company, you would have a solution for agriculture. We're taking all these feeds and stuff like that. And yeah, that's exactly right.

[Will Marshall] So there's similarities and differences. And of course, I already mentioned, one-to-many is an important feature of DaaS models. Anyway, there are differences. And I think, the trial of challenges is a less or less well trodden path then the SaaS, there's more of that to follow. But I stick to my guns, we're primarily a data company. We are also a software company, because as we add analytics, it enables, per the earlier point about going up the stack to enable solutions, we do grab more value and enable users that wouldn't otherwise be able to use it. But in my sense, the core value proposition is the data. And I actually believe that's the case for a lot of companies, but you don't think of it like that. I think Google is absolutely a data company. I mean, people say it’s a search company, you can say it's a video company, you can say it's an AI company, you can say it's a server company to the earlier point. But I would say it’s a data company. Google is not leading the world because it has the best search algorithm, not anymore, because it has the best data about what everyone wants. And that data set is unrivaled. Right? You'll see them open source all their AI technology. But they're not an AI company in the sense that their product is an AI. In fact, they give that away for free, which tells you something that's not the asset that they think is most valuable. If they thought that was the most valuable, they would not be putting TensorFlow and all these modules online.

[Auren Hoffman] They had their best engineers develop TensorFlow, and they're like “Oh, yeah, we'll just give it away for free”. Of course, the reason is, what does it matter, they have all the data, they give it to anyone they want.

[Will Marshall] The asset is the data. The Economist quipped that data is the new oil. I think there's challenges with that analogy. I don't think data is dirty. Also, oil is only one to one.

[Auren Hoffman] Once you spend it, you can’t resell it. 

[Will Marshall] However, in this in a couple of senses, it's a good analogy. One is, it powers lots of industries. And others, you have to refine it before you can do that, dig it out of the ground and give it to people. And it's extremely bad, it is commoditized. And people talking about data being commoditized. That doesn't mean it loses value. Tell the world that the oil is not valuable and you're gonna have some laughs. No, data is extremely valuable, even as it gets commoditized. And so I do think data is empowering lots of industries, just like oil. 

[Auren Hoffman] You mentioned AWS, is compute a good analogy as well?

[Will Marshall] Again, only one person can use one compute instance at any one time. So no, I think data stands on its own and being a one-to-many, and not one-to-one.

[Auren Hoffman] Interesting. Now, one interesting thing about Planet is that a lot of Planet's revenues come from the government. And, you're lucky when you started Planet there was already an existing use case. But I still imagine it's very difficult to sell to the government. What's the secret? What advice would you give to others who want to sell into the government market?

[Will Marshall] Assistance, wanting to sell with the government to the government? I think a lot of tech firms don't think that they want to, they think they are bureaucratic or they disagree with some policy. I actually think that's the wrong attitude. I think leaning in and being a partner. Governments are trying to do services for the citizens and it's our job to try and help them, not just one government but any government. So persistence and wanting to work with the government I think is important. We just today announced that we're expanding our contract with NASA. It's really exciting what we do with them, we provide them data to help monitor what's called key climate variables, the underpinnings of our climate models. So they do a lot of research on data like glaciology, how fast are the glaciers melting, and hydrology. And the announcement today was about expanding our contract, and they want to provision our data to everyone that's a researcher under NSF grants. So there's 280,000 NSF funded researchers, and they will now all be able to get access to our data. That's important for multiple reasons: one, for science and advancing our understanding of systems. Secondly, because they generate all of our use cases that underpin our clients. So actually, it serves Planet very well, I mean, NASA is paying for this so that's great for us. But also, agriculture wants to move to understand and leverage satellite data, they don't just do it on the fly. They base it ultimately on scientific studies that show that our data can enable precision agriculture, or sustainable agriculture practices, that improve their crop yield, decrease their fertilizer use, decrease the impact on the environment of that so it enables them to advance the industry. Same with forestry and forest management. And so academic papers actually underpin our business. And so this is really great for us. But anyway, serving the government is something that we think makes a lot of sense.

[Auren Hoffman] Is there some insanely crazy use case that people would appreciate or wouldn't have been able to think of? Agriculture is a really interesting one, one that people would grok and understand, but there may be some other, random one that maybe people would have never thought of before?

[Will Marshall] One that's happening a fair bit now, but I wouldn't have thought about even just a year or two ago, is permitting. So local governments are using our data to enable permit enforcement. So planning for houses, we will check whether they've got planning permission and then send fines to them if they're not. Humboldt County uses that data to track marijuana. It’s legal, but you have to have permits. 

[Auren Hoffman] There's a classic one of Greece where they have a tax for pools. And they show that there were way more pools in Athens and people paying taxes for this.

[Will Marshall] This is interesting and I was just to administer this this morning, this is like the speed cameras for them for permitting. And they don't want to send the person out both because it's more expensive, but also it makes it too impersonal. And then people try and persuade you, “well, I just had to have this swimming pool for this, or just to have this extra amount”.

[Auren Hoffman] Or in certain countries, there might be bribes or other types of things.

[Will Marshall] Here’s your bill in the mail, you were speeding. Yeah, you build a building where you are meant to and it's impersonal, but it's actually efficient. And it does the job. I think that there's gonna be a lot of that. We've seen people do some incredible stuff with our data. I mean, we work with Google on maps. And you might think, “well, that's an obvious application, keep maps up to date”, but the sophistication of what's behind the scenes there is insane. So they use their cell data to figure out if there's a new road or building because people are suddenly driving through a field and they're like “what the heck”'. That automatically triggers a lat long task request to our high resolution satellite, that high resolution takes a picture, and then computer vision extracts out the road, or that new building, or that new train station, or whatever it is, and then updates their maps. And they do that tens of thousands of times a quarter. And it's completely automated. Humans are completely out of the loop. And that's just that's how your maps stay up to date. It's just crazy. But it is a really sophisticated system. So people might think “oh, maps get updated”, but that's an incredible system. So yeah, there's a few years. So those are some of the novel ones, but a lot of it's quite down to earth. I mean, that agriculture use case, on the infrared band we can literally tell crop type and crop yield, and every three by three meter pixel of every area of every farmer. 

[Auren Hoffman] And you can tell the crop type, because you can get the temperature coming off of it, you can see a little bit of what's under the soil, and there's other kinds of things that you're getting from different spectrums.

[Will Marshall] Yeah, so our near infrared spectral band picks up chlorophyll and we basically can measure biomass. And if we trim that over time, we can quickly tell the crop time. But the scale of it, agriculture represents 25% of the landmass of the earth. So back to your plane thing, it’s really hard to find planes or drones over all of that. We can fly over all of that every day and enable the farmer to do more precision agriculture every day, of course.

[Auren Hoffman] And the great thing about some of the spectrum things is when you're getting imagery, if you have clouds, you can't see it that day. Whereas a lot of these other types of spectrum stuff, you can still get through the clouds, which is really nice.

[Will Marshall] Well, we're getting that way. I mean, actually, most of our spectral bands can't go through the cloud. But what we can do is burn through the cloud by having a really regular frequency. We just announced these hyperspectral satellites, which will have 400 spectral bands, we are going to be launching them in a couple of years here. And they will be able to pinpoint methane and co2 point source emissions, like here's the gasoline, here's the house that has that gas on, here's the waste facility. 

[Auren Hoffman] I know, most satellites, the ideal time is, say, 11am to 2pm that they want to look at, but there may be some interesting things that are going on at night that are very different about these particular places or something. Do you think about time of day as a kind of a particular variable?

[Will Marshall] Absolutely. Most of our fleet scans the whole world at about 11am local time, no matter where you are. So if you're in LA it’s 11am, if you're in Tokyo, it’s 11am. And that's because they stay fit with respect to the Sun and the Earth rotates underneath it, so local time of day plus or minus. And that's really useful because then you have consistent shadow angles and things like that, which enable machine learning and analytics to be easier. We complement that with a sky set system that we can zoom in on a particular location up to 10 times per day, which is useful for a lot of applications where you need more rapid revisit. 

[Auren Hoffman] So let's dive into SPACs. Just because it's such a great business, Planet has  many, many different options. And if you want to go public, there's like eight different ways to go public, and you chose the SPAC route. Why did you choose that route, as opposed to the other ways to potentially bring liquidity to your shareholders?

[Will Marshall] Well, firstly, we're excited about going public. I mean, that's the first thing that motivates. We think that Planet is ready to go public. And the world really needs Planet. Let me explain that. So we're ready to go public, we've got a mature satellite fleet producing a unique data set and now we've got a mature business on top of that, with over 100 million in revenue last year, now it's public knowledge. And it's growing in a healthy way. We've spent the last year preparing our executive team and other functions, financial, legal, and so on to be ready to go public. It was always the right destination for our company and we were ready. Then the SPACs became part of course in recent times. And it's an efficient mechanism to go public. And it's one way you can say a little bit more about your future and your plans. And I think that's good that within reason there's some challenges with SPACs, but we felt, of all the companies that made sense to use that mechanism, it was Planet. Pretty much every SPAC under the sun wanted to partner with us, so we were lucky to have the choice. And we chose dNY Technologies, because we really loved how they saw the business, understood us as a data company, understood how we were enabling the sustainability transition, the digital transformation of the global economy, source in that big perspective, had done SPACs before. This is their fourth one. So they’re experienced in that process, they really cared and loved our mission. And so we thought that was right. And then we managed to attract really great investors into our pilot as well, BlackRock, which has a big focus on ESG and ensuring their companies care about the environment and sustainability. And we can actually help them. Marc Benioff is putting in a last check. He is excited because he wants to dedicate the next 20 years of his life to biodiversity. He's been looking at every startup under the sun that is doing anything to sustainability in the environment, and biodiversity. He said either they are using Planet or they need Planet’s data. Basically, as he put it, all roads lead to Planet when it comes to sustainability. Basically, our data enables that sustainability transformation. And we have companies like Google coming back, existing investors and so we have a powerful set of actors with us and we always like that, because we want to build a company with great investors.

[Auren Hoffman] Alright, a couple of couple personal questions for you. Now, before Planet, you were an accomplished scientist at NASA. What are things that one could only really learn from working at NASA? 

[Will Marshall] Systems engineering.

[Auren Hoffman] Okay, what do you mean by that?

[Will Marshall] That is putting together really complicated projects that take planning, require thinking about requirements, thinking about how the systems interlock. So any one of the subsystems of our satellite are really complicated. Our radio systems transmit 1.8 gigabits per second or 1000 kilometer range from a radio that's a few centimeters in size, to a dish that's on the ground that we built ourselves that's a few meters wide, somewhere on the ground. Those radios are really difficult. There's 12 subsystems of a satellite. If you open up a satellite, like when you open up your phone, you will see these complex motherboards, those reaction wheels, power systems, batteries, compute systems and camera systems and all that. We have pioneered every subsystem to be a really state of the art system, much better than anything you can buy online. And then, just like Apple does with its phones, we have integrated those into a common box, that's really hard. So they all work together, the cameras work with the reaction wheels, that work with the power systems. And then those satellites then have to work with the other satellites.

[Auren Hoffman] Why do you learn that so much at NASA? Is it because as opposed to working at Qualcomm or working in some other place? Is it because they only get one shot to put it in space so they really have to go through the precision? Or is it because they've got 50 years of thinking about these hard problems? Like, why is that in the water and NASA so much? 

[Will Marshall] So where I was going with that point about all the different steps, and there's several more, is that putting all those things together takes really thoughtful planning, it doesn't just happen. This is where the agile aerospace mantra sort of breaks down, because actually, you have to plan it out at some significant level. Unlike software, where you can just go well, let's just keep iterating. 

[Auren Hoffman] You also have to buy all these components.

[Will Marshall] Yeah, yeah. And that's why the satellite has to work the first time. If it doesn't, you've got a brick in space. We have to have all these circuits that enable us to reboot it and the circumstances. The satellites have to work together. We have to have Mission Control Systems.  At NASA, we would have 200 people doing controls around every one satellite, round on the clock or whatever, we do the opposite as we have two hundred satellites controlled by one guy. It’s not quite like that, there's a few. Generally, they're building the software that automates the Mission Control while actually controlling the satellite. But that systems engineering takes a huge amount of planning and a huge amount of understanding. It's not for the faint of heart to plan that out. And so very few institutions on the planet have done that. NASA has done that with the Apollo project, the US government did it with the Manhattan Project. There are projects like this, but it's that kind of scale. We've undertaken a minor Apollo project, we launched the largest fleet of satellites in human history, we had these ground stations around the world, we had to create our own technology ourselves. And we do that all with a 500 person team. So it's a lot in a small area and our NASA training, to cut it short, enabled us to undertake complex projects like that.

[Auren Hoffman] Interesting. Now another kind of personal question is I love the Tom Hanks series on HBO from the Earth to the Moon. I don't know if you've ever seen it, but I just absolutely love that series. Is there a favorite space movie that you have?

[Will Marshall] I like Star Trek. I don't often get that much time for movies, to be honest. But Star Trek is very good. I love how they always encounter different cultures. It's actually a philosophical program. Yep. And people think of it as science fiction. It's actually much more about morality and sociology and I find that interesting as well.

[Auren Hoffman] Yeah, absolutely. Okay, cool. Last question we ask all of our guests: if you can go back in time, what advice do you wish you could have told your younger self?

[Will Marshall] Gosh, so many things. I mean, stick with it. I have quite a lot of self intrinsic motivation, but you know, it can be tough at times when you have a lot of people saying “you can't do it, no you can't do it, no you can't do it” and it's like, you have to go “is this something I've missed, or what?” and stick with it. If you’re having doubts, I really like what Gwynne Shotwell says “no assholes policy in SpaceX”. “Life's too short”, I would tell myself that I learned some lessons there.

[Auren Hoffman] If you find someone who is super talented, but has got a difficult personality, or can maybe treat others in a way that maybe shouldn't be tolerated, do you just let that person go? Or do you say, “Okay, we're gonna put them in a box and only interact with them through an API?” Or how do you think about that?

[Will Marshall] I love that, an API, you can try that, and I have and it doesn't work. Teams work together because they’re teams. And unfortunately, even if somebody is wickedly talented as an individual contributor, there's no use if they can't work with everyone else. So I look for people that care about our mission, who are smart, and are collaborative. And if you have those three things, everything else fits into place. So that's my philosophy on that.

[Auren Hoffman] This has been great. Could you tell our World of DaaS listeners a little bit more about where they can find you on the internet? 

[Will Marshall] That's quite easy, it’s planet.com.

[Auren Hoffman] Okay, perfect. That's easy. All right. Well, this has been awesome. Thank you so much for being with us on World of DaaS.

[Will Marshall] Hey, no worries at all.

[Music playing]

[Auren Hoffman] Thanks for listening. If you enjoyed this show, consider rating this podcast and leaving a review. For more World of DaaS (DaaS is D-A-A-S), you can subscribe on Spotify or Apple Podcasts. Also check out YouTube for the videos. You can find me on Twitter at @auren (A-U-R-E-N). I’d love to hear from you.