Founder Interview
How Aquabyte Applied Machine Learning to Fish Farming With 3 Customers and a $10M Series A (Interview with CEO Bryton Shang)
- Interview Date
- July 10, 2019
- Interviewee
- Bryton ShangFounder and CEO
Company Metrics at Interview Time
Customers (2019)
3
Series A Raised (2019)
$10,000,000
Seed Round Raised
$3,500,000
Team Size (2019)
30
CAC Payback (2019)
About 2 months
Historical Snapshot
These numbers were reported by Bryton Shang during the interview recorded in July 2019 and are a historical snapshot, not current figures. See Aquabyte’s current numbers.
Key Takeaways
- 01Aquabyte was founded in 2017 and began selling to customers at the start of 2019
- 02The company had 3 customers as of July 2019; customers began with a trial period and then converted
- 03Aquabyte had just closed a $10M Series A at the time of the July 2019 interview, following a $3.5M seed round
- 04The team totaled 30 people, with 12 in San Francisco and 18 in Bergen, Norway
- 05Software was sold as a monthly subscription priced by the pen, and Bryton put an average farm in the $100,000 to $200,000 range
- 06A typical farm has four to six pens, each requiring a single camera costing $20,000 to $30,000
- 07Aquabyte partnered with camera manufacturers and did not carry camera hardware on its own books
- 08Bryton estimated CAC payback at roughly a couple of months, driven mainly by travel costs and local sales team salaries
- 09The company planned to grow from 3 customers to 8 within one to two months of the interview
- 10Bryton described the market as a land grab, saying Aquabyte was one to two years ahead of any competitor
Company Metrics at Time of Interview
| Metric | Value | Source |
|---|---|---|
| Year Founded | 2017 | Founder interview, July 2019 |
| Customers (2019) | 3 | Founder interview, July 2019 |
| Team Size (2019) | 30 | Founder interview, July 2019 |
| Team in San Francisco (2019) | 12 | Founder interview, July 2019 |
| Team in Bergen, Norway (2019) | 18 | Founder interview, July 2019 |
| Seed Round Raised | $3,500,000 | Founder interview, July 2019 |
| Series A Raised (2019) | $10,000,000 | Founder interview, July 2019 |
| Software Price for an Average Farm (2019) | $100,000 to $200,000 | Founder interview, July 2019 |
| Pens per Average Farm (2019) | 4 to 6 | Founder interview, July 2019 |
| MVP Investment | $1,000,000 to $2,000,000 | Founder interview, July 2019 |
| CAC Payback (2019) | About 2 months | Founder interview, July 2019 |
| Planned Runway from Series A (2019) | 18 to 24 months | Founder interview, July 2019 |
Growth Breakdown
Revenue
Aquabyte began selling its monthly subscription software at the start of 2019. The company charged by the pen, and Bryton put an average farm, with four to six pens, in the $100,000 to $200,000 range. As of July 2019 the company had 3 customers. Bryton said revenue was not all immediate: customers began with a trial period and then converted.
Customers
Bryton described deliberately throttling customer growth while the company transitioned from R and D to scaling. He planned to grow from 3 customers to 8 within one to two months of the interview, and to have close to 100 pens deployed by the end of 2019.
Team
The team stood at 30 people in July 2019, split between 12 in San Francisco and 18 in Bergen, Norway. Headcount was described as the primary driver of the company's monthly burn.
Funding
Aquabyte raised a $3.5M seed round and had just closed a $10M Series A at the time of the July 2019 interview. The Series A was intended to fund go-to-market acceleration, international expansion, and new product development, with Bryton projecting the next raise would come around the end of 2020.
Growth Strategy
Word-of-Mouth and Relationship Selling
Bryton credited the aquaculture industry's strong word-of-mouth culture as a key growth driver. Because Aquabyte's team knew many of these farmers personally and had been talking with them over the past year, the company was able to close customers in roughly a month or two rather than through a long formal sales process.
Local Sales Team in Norway
Rather than selling remotely from San Francisco, Aquabyte placed the majority of its team in Bergen, Norway, where the salmon farming industry is concentrated. Local salespeople with existing farmer relationships kept CAC low, with costs driven mainly by travel and salesperson compensation.
Asset-Light Hardware Model
By partnering with camera manufacturers and passing hardware costs directly to them, Aquabyte avoided carrying expensive cameras on its own books. This let the company operate as a pure-play software business from a financial standpoint, collecting subscription revenue from day one of a customer's conversion.
Land-Grab Positioning
Bryton framed the market as a first-mover land grab: once a camera is installed in a pen, a second camera will not be added. He estimated Aquabyte was one to two years ahead of the nearest competitors, which he identified as subsea oil and gas camera companies such as OptoScale.
International Expansion
Starting in Norway, Aquabyte planned to use its Series A capital to expand into other aquaculture markets internationally. Bryton described aquaculture as a fairly international industry, and international growth was cited as one of the three main uses of the new funding alongside go-to-market acceleration and new product development.
Best Quotes
“I had studied machine learning at Princeton and had worked in a number of machine learning companies in the past. My first company was an algorithmic trading firm. So sort of similar to Flash Boys, we were buying and selling stocks automatedly.”
“We have a camera. The camera goes in the fish pen. From that, using computer vision, we can do a number of interesting things. We can determine the weight of the fish, the number of sea lice, as you mentioned, and we do sell it as a monthly subscription fee to these farms.”
“It's in the hundreds of thousands for like an entire farm. And it like, we charge by the pen, each fish pen.”
“So you can take seven thirty seven plane, dangle it by the nose and completely submerge it in one of these pens. So we're talking about the diameter of a pen being the size of half a football field.”
“The camera is actually quite expensive. Now the farmers, they already have cameras in the pens, not as sophisticated as the ones we have, but they're already used to having that type of hardware. What we do is we're actually a pure play software company in that we partner with these camera manufacturers.”
“So we just started selling. We're still in the R and D phase and transitioning to scaling. So right now we have three customers.”
“I think the thing is there's a lot of demand for this product. Part of it is that we've been throttling the amount of customers we have. Actually in the next month or two, we're going to go to eight customers. And then by the end of the year, we want to have almost a 100 of these pens deployed.”
“If I had to just estimate off the top of my head, it's probably a couple of months.”
“I would say so. The thing is, we're still a year or two ahead of any other competitor that's in the space.”
“So we have 30 people. 12 in San Francisco, 18 in Norway.”
What Happened Next
At the time of this interview in July 2019, Aquabyte had 3 customers and had just closed a $10M Series A, with Bryton planning to reach 8 customers within one to two months and almost 100 pens deployed by the end of the year. The company was in the early stages of transitioning from R and D to commercial scaling, with international expansion and new product development as stated priorities. These figures reflect the company's position at the time of recording and are a historical snapshot. Visit the Aquabyte company profile on GetLatka for current numbers.
View Aquabyte’s current profile and metricsFull Transcript
Chapters
- 0:00Introduction and Background
- 0:26How Bryton Landed in Aquaculture
- 2:00The Business Model: A Monthly Subscription
- 2:17Pricing by the Pen: What an Average Farm Pays
- 4:29Hardware Strategy: Partnering With Camera Manufacturers
- 6:40Company Timeline and MVP Investment
- 7:17Why Aquabyte Raised the Series A
- 7:37CAC and Payback Period
- 10:04Current Customer Count and Growth Plans
- 10:42Revenue Targets for Year-End 2019
- 12:22Competitive Landscape and Land-Grab Dynamics
- 13:32Burn Rate and Runway
- 14:26Team Size and Location
- 14:35Famous Five Rapid-Fire Questions
Introduction and Background
Nathan Latka
00:00Hello, everybody. My guest today is Bryton Shang. He's the founder and CEO of Aquabyte, a Silicon Valley and Norway based venture backed company applying machine learning and computer vision to aquaculture, fish farming for biomass estimation, sea lice counting, and feed optimization and formulation. Bryton was named to the 2019 Forbes 30 Under 30 in manufacturing and industry. All right, Bryton, you ready to take us to the top?
Bryton Shang
00:21Absolutely.
Nathan Latka
00:21How do
00:23you land in this niche? That's my biggest question.
How Bryton Landed in Aquaculture
Bryton Shang
00:26It's a bit of a non direct story. I
00:32had studied machine learning at Princeton and had worked in a number of machine learning companies in the past. My first company was an algorithmic trading firm. So sort of similar to Flash Boys, we were buying and selling stocks automatedly. Then I was involved in another company, a series of startups. The last company I was involved with, we were doing computer vision modeling on cancer tissues. That company ended up going through the Y Combinator program, moved out
01:05here from New York, and then ended up incubating this company out of a VC firm out here. The whole idea for aquaculture came out of my old co founder. One of his business school classmates had owned a seabass farm in Mexico, And it had been really interesting to me and ended up traveling all around. And everyone kept on talking about Norwegian salmon. And long story short, ended up living in Norway for a couple of months with
01:34the fish farmers investigating the industry. They have an amazing industry out there, but then developing this thesis of being able to use similar machine learning computer vision technologies, putting a camera in the fish pen and being able to monitor a lot of different things. So over time, learned a lot more about it.
Nathan Latka
01:55And so what's the model today? Is it a pure play SaaS revenue stream?
The Business Model: A Monthly Subscription
Bryton Shang
02:00So we have a camera. The camera goes in the fish pen. From that, using computer vision, we can do a number of interesting things. We can determine the weight of the fish, the number of sea lice, as you mentioned, and we do sell it as a monthly subscription fee to these farms.
Pricing by the Pen: What an Average Farm Pays
Nathan Latka
02:17What's the fee? I guess on average, just the software component of this, what's the average customer paying per month or per year to use the tech?
Bryton Shang
02:25It's in the hundreds of thousands for like an entire farm. And it like, we charge by the pen, each fish pen.
Nathan Latka
02:36Okay. So I guess totally understand that that's what you price against, but what I'm actually asking is the average kind of how many pens on average per company you're selling to. So are we talking like $100,000 average kind of annual accounts for five pens or something different?
Bryton Shang
02:53Something in like the 100 to $200,000 range for an average farm.
Nathan Latka
02:57Okay. And the average farm has how many pens?
Bryton Shang
02:59Four to six pens.
Nathan Latka
03:00Okay. Four to six pens. Very good. And just help my audience visualize that. So how big is a pen? Is it like the size of a typical commercial swimming pool or?
Bryton Shang
03:09So you can take seven thirty seven plane, dangle it by the nose and completely submerge it in one of these pens. So we're talking about the diameter of a pen being the size of half a football field.
Nathan Latka
03:22That's huge, massive. How many... So for a four to... Okay, just for one pen, how many pieces of hardware, these kind of cameras do they have to put in that one tank?
Bryton Shang
03:33It's a single camera.
Nathan Latka
03:33Okay.
Bryton Shang
03:33And the camera navigates throughout the pen watching the fish.
Nathan Latka
03:39So it's almost like when I was growing up and had a pool and we had this little automated thing that cleaned the bottom of our pool that has moved around in the mornings, it would go a little alien. You kind of work like that.
Bryton Shang
03:49Yeah, something like that. So it's not rope winch system and winch moves it all throughout the pen.
Nathan Latka
03:56That's pretty cool. Okay, so I'm always fascinated by IoT plus software plays, especially niche verticals. I've seen a lot of kind of ag tech doing this where there's like orbs and feed grain relative to like how much cows are eating or sensors on the back of Target dumpsters and Walmart dumpsters to help the dumpster people know when to come pick up things like that. And I'm always fascinated by how people are subsidizing the hardware to get
04:19as close to free to their customers so that it can drive more growth. So I'm curious about that for you. How do you handle the cost structure of the physical installation?
Hardware Strategy: Partnering With Camera Manufacturers
Bryton Shang
04:29The camera is actually quite expensive. Now the farmers, they already have cameras in the pens, not as sophisticated as the ones we have, but they're already used to having that type of hardware. What we do is we're actually a pure play software company in that we partner with these camera manufacturers. So from a farmer perspective, they're buying a complete package of the software plus the camera. On the back end, we're actually passing part of that cost
05:02directly to the hardware manufacturer. And that we're not even a part of that. So we're not buying the camera and reselling it on a SaaS model. Just buy that separately.
Nathan Latka
05:15What's it cost?
Bryton Shang
05:17It's like 20,000 to 30,000 for one of these cameras.
Nathan Latka
05:20Okay. We'll call it 20,000 for one of these cameras. And so what you're saying is you will essentially sell your software. So your sales sound something like this. Hey, farmer with four to six pens, it's gonna be between a 100,000 and $200,000 per year for the software. And on top of that, need one camera per pen. They're $20,000 a pop, you have four pens. It's another $80,000. So they're looking essentially at shelling out at least $180,000
05:40for the 80 on the hardware plus 100 on the software first year.
Bryton Shang
05:45Right, but then these cameras, they last for five, six years. And also they already have cameras in the pen,
05:55can potentially retrofit their camera, upgrade it, and so they don't have to buy a completely new camera.
Nathan Latka
06:03Is that, like your salespeople or you, when you're going out and selling this, is that $20,000 per camera cost a big friction point? Have you lost deals because they don't want to pay for that hardware?
Bryton Shang
06:13No, I think that actually ironically the farmers, they're more used to buying hardware than software. And so they'd rather pay a large amount for the hardware and a less amount for the software because they're just not used to that. They're used to buying nets and barges and the bigger, the better, not so much software. So they're not used to sort of SaaS model.
Company Timeline and MVP Investment
Nathan Latka
06:40All right. And put all this on a timeline for me. When did you launch?
Bryton Shang
06:44So I started the company two years ago. We started selling at the beginning of this year.
Nathan Latka
06:51Okay. And I always like to ask this question. So how much did you sink into your MVP right before your first dollar of revenue? We're talking a $100,000, $1,000,000 or more?
Bryton Shang
07:02Probably 1 to $2,000,000 Part of this is that we raised a $3,500,000 seed round. We just closed a $10,000,000 series A.
Nathan Latka
07:14So 13.5 total?
Bryton Shang
07:16Right.
Why Aquabyte Raised the Series A
Nathan Latka
07:17Why do need the extra 10?
Bryton Shang
07:19So that $10,000,000 we just raised is now going to be used to accelerate our go to market and as well as our international plans. So aquaculture is fairly international, starting in Norway, expanding to other countries, and then to build out the next couple of products.
CAC and Payback Period
Nathan Latka
07:37Bryton, do you have a good sense, so to get a new $100,000 or $200,000 per year customer year one, are you comfortable spending that whole first year ACV to get the customer or what payback period are you optimizing for currently?
Bryton Shang
07:50So because we don't take the camera on our books, from a financials perspective, we're making money from day one because we're selling the software. That said,
08:06aquaculture industry is very interesting. It's not a traditional industry where you have to do the complicated LTV calculation. A lot of the industry is very word-of-mouth. And so we know a lot of these farmers personally, and we're selling to them based on one on one relationships. And so we can acquire a customer on the order of a month or two because we've been chatting with them over the past year versus being a long sales team.
Nathan Latka
08:38So what is that?
08:38You're talking like maybe $20,000 to $30,000 that covers flight costs, maybe a meal here or there past meetings. That's essentially what you cost to get a new customer?
Bryton Shang
08:47Well, actually, even though I'm based in San Francisco, most of our team is in Bergen, Norway. And so we have local salespeople there that already have relationships with the farms.
09:01So I think it's even less than that.
Nathan Latka
09:03Well, Bryton, the reason I'm asking is I'm talking fully weighted CACs. You just mentioned salespeople. I mean, there's a commission structure built in there. Otherwise they wouldn't be a call to salesperson and they'd be at another company where they can earn more money. Right? So, I mean, you have modeled some sort of CAC, even though you're not doing any direct paid, like, you know, you're not finding fishermen on Facebook ads. Right? So are you generally talking
09:22like a four month payback on these things or twelve months or more?
Bryton Shang
09:29If I had to just estimate off the top of my head, it's probably a couple of months.
Nathan Latka
09:34Okay. But definitely less than a year you think?
Bryton Shang
09:37Right.
Nathan Latka
09:38Okay, fair enough. And most of that, it sounds like it's going go towards actually the salesperson compensation and maybe a flight here or there.
Bryton Shang
09:45Right, mainly travel costs and having our sales team.
Nathan Latka
09:50Very cool. Okay, so 2017, you start building this thing, put in 1 or 2,000,000 to get the MVP. And then now today, early twenty nineteen, you started kind of onboarding your first customers. Over the past six to seven months, how many customers total have you scaled to?
Current Customer Count and Growth Plans
Bryton Shang
10:04So we just started selling. We're still in the R and D phase and transitioning to scaling. So right now we have three customers.
Nathan Latka
10:18The way, why do you hedge that so much? It's almost like, Oh my gosh, I don't want to share. We only have three, but like you got to start somewhere and three is a great number to start. Why not?
Bryton Shang
10:24Well, I think the thing is there's a lot of demand for this product. Part of it is that we've been throttling the amount of customers we have. Actually in the next month or two, we're going to go to eight customers. And then by the end of the year, we want to have almost a 100 of these pens deployed.
Revenue Targets for Year-End 2019
Nathan Latka
10:42Yeah. So what does that mean in terms of revenue? What's your revenue target by the end of this year in terms of run rate?
Bryton Shang
10:47I think we're looking between three to five.
Nathan Latka
10:49Okay. And does that feel realistic or like a stretch goal?
Bryton Shang
10:53I think that's realistic. Depends on our ability to deliver on the technology.
Nathan Latka
10:57Yeah. And it sounds like today if you got three customers and the minimum ACV is, call it, $100,000, I mean, you're doing like $30,000 or something like that per month right now.
Bryton Shang
11:05Yeah, so it's not all immediate, like there's a trial period and then they convert, but yes, that like once they convert.
Nathan Latka
11:14Yeah, so let me ask you a question. When you model and maybe you don't model this, but when you model lifetime value early on on a company like do you have an idea of what you think a customer is worth to you over the life?
Bryton Shang
11:27That's an interesting question.
11:31Certainly,
11:35mean, we have a model where we've modeled out revenue and churn and things like that. But yeah, I mean, it's something I haven't extensively modeled to give a reasonable answer on that.
Nathan Latka
11:49Yeah. I mean, the reason I'm asking is because I've seen so many of these hardware companies where they actually model the software LTV and it actually is so worth it for them to even give away a $20,000 piece of hardware for free, right? Just to grow distribution super fast because they know the lifetime values make sense. Now you have to bridge the cash gap on that. You raised capital, it sounds like to do that partly, but
12:12I'm just curious if you ever think about that.
Bryton Shang
12:14No, I think that's a good point. I mean, I see your point and I think it could be a similar thing here.
Competitive Landscape and Land-Grab Dynamics
Nathan Latka
12:22Think it only works though if it's a land grab situation, right? It worked for Bird and Lime because that is truly a land grab situation. Get more scooters out, give the hardware away essentially free. Are you competing with anyone else? Is there a land grab for cameras and fish
Bryton Shang
12:22Well, I
12:34would say the other thing is, we're in an industry... It's like once
12:42you get a cable at your home, you're not going to get a second cable provider. And so it's a similar thing. Once you put a camera and a pen, you're not getting a second camera and the pen.
Nathan Latka
12:53It is a land grab first in, I mean, first person to the tanks wins.
Bryton Shang
12:58I would say so. The thing is, we're still a year or two ahead of any other competitor that's in the space.
Nathan Latka
13:08Who's closest would you think?
Bryton Shang
13:10There are some camera companies that are out of the subsea oil and gas industry in Norway, and they've been making these fairly expensive cameras. So I'd say they're second closest to us.
Nathan Latka
13:25What's the name of one of them?
Bryton Shang
13:28For example, Sing Ray or OptoScale.
Burn Rate and Runway
Nathan Latka
13:32OptoScale, interesting. Okay, very cool. Talk to me a little bit more, when you do big raise, obviously you're raising to cover some amount of So you obviously just completed a raise in your mind when you were raising, what are you projected? Like you want to cover and raise for burn for eighteen months or twenty four months or what?
Bryton Shang
13:47Eighteen to twenty four months.
Nathan Latka
13:48Okay. Interesting. Okay. So you raised 10,000,000 to cover eighteen months. That means what you're burning like $400,000 to $500,000 a month right now, something like that?
Bryton Shang
13:58Right, so we're projected to raise at the end of twenty twenty.
Nathan Latka
14:03Okay, yeah, which is about eighteen months from now. Guess my question
Bryton Shang
14:03And
14:06then you want to add in some cushion
14:11at the end, but your numbers are roughly correct.
Nathan Latka
14:15Okay, so you're calling say burning somewhere around $400,000 a month right now though, and your 10,000,000 more than covers eighteen months of runway.
Bryton Shang
14:15Right.
Nathan Latka
14:15And most of that is on engineers or what?
Bryton Shang
14:24It's primarily headcount.
Team Size and Location
Nathan Latka
14:26Yeah. What's the team look like today? How many folks?
Bryton Shang
14:28So we have 30 people. Yeah. Very 12 in San Francisco, 18 in Norway.
Famous Five Rapid-Fire Questions
Nathan Latka
14:35Very good. All right. Let's wrap up here with the famous five. Number one, what's your favorite business book?
Bryton Shang
14:41High Output Management.
Nathan Latka
14:42Number two, is there a CEO you're following or studying?
Bryton Shang
14:46Yeah, that's a great question.
14:50CEO.
14:56I have to come back to you on that. Don't want to answer it all
Nathan Latka
14:56That's okay.
15:00Number three, what's your favorite online tool for building your company?
Bryton Shang
15:07It's not an online tool, but I use Evernote religiously
Nathan Latka
15:07Number for everybody's
15:11four, how many hours of sleep do get every night?
Bryton Shang
15:14I get seven to eight hours.
Nathan Latka
15:16And what's your situation? Married, single, kids?
Bryton Shang
15:19Single, girlfriend.
Nathan Latka
15:20Okay. And how old are you?
Bryton Shang
15:2227.
Nathan Latka
15:23Last question. What do you wish your 20 year old self knew?
Bryton Shang
15:28Yeah, I wish I knew to come out to San Francisco earlier to learn what it's like to start a company and just live and breathe the startup life.
Nathan Latka
15:40Guys, aquabyte dot ai serving three customers. They serve again these fish farms without anywhere between four and six massive tanks. They deploy $20,000 cameras in these tanks and then help these fisher... These fish pens essentially man... Manage things like sea lice and other critical elements of the kind of fish health, right, as they bring them to market. Again, 2017 launch date, they've raised 13,500,000 to date burning, call it, $400,000 per month. Team of 30 folks. Too
16:04early to talk about churn and things like that, but they spent about a $2,000,000 to get their MVP built, launching their first paid plans early this year in 2019. Now looking obviously to scale up from there. Bryton, thank you for taking us to the top.
Bryton Shang
16:14Yeah. Thank you.