[protected-iframe id=”67e9f07a16e63f5d7032f6c2eaa9e37b-14960843-33105277″ info=”http://new.livestream.com/accounts/74987/events/2423781/videos/32468371/player?autoPlay=false&height=360&mute=false&width=640″ width=”640″ height=”360″ frameborder=”0″ scrolling=”no”]
Input sound file:
10-17 pm session 2_1001.MP3
Session name: Unlocking The Value In The Internet Of Things: A Smart Agriculture Case Study
Please welcome your emcee, Chris Albrecht back to the stage.
Chris Albrecht 00:28
Almost there. I’m going to waste no time and bring the next panel out. It is Unlocking The Value In The Internet Of Things: A Smart Agricultural Case Study. Stacy Higginbotham is actually going to moderating this one, she’s the Senior Writer at GigaOM. And she’s talking with Lance Donny, Founder and CEO of OnFarm Systems. Please welcome Stacy and Lance to the stage.
Stacy Higginbotham 00:50
We’re going to be talking about connecting farmers in the agriculture industry. This is Lance Donny, Founder and CEO of OnFarm. Why don’t you tell everyone about what OnFarm is doing.
Lance Donny 01:03
Sure. OnFarm takes sensor data, Cloud data that data farmers use to make better decisions about their operation. We bring it together from disparate vendors into one software of service, so they can use anywhere to make those decisions. What that enables them to do is, use less water, less fertilizer, energy labor, and grow better crops. Really why they need that is that, globally, the demand for food has changed; it’s increasing. Farmers will need to grow twice as much as they do today by 2050, and we have no more land, no more water, no more food. How are they going to do that? Almost all of that growth is going to be through technology.
Stacy Higginbotham 01:46
What are some examples of what farmers are doing with your technology?
Lance Donny 01:53
An easy example to understand is weather or soil moisture data. For soil moisture data, there’s sensors that go in the ground and they measure how much water is available to the plant. So, it gives the grower an indication of when and how much they need to apply to keep their crop healthy. So, that data is collected locally at the sensor, and it’s usually sent via a cellular connection back to our system. That’s a great example of using technology to improve how and when they’re applying water.
Stacy Higginbotham 02:24
And then you deliver some kind of system so that they can remotely monitor and manage and cut down on that?
Lance Donny 02:30
One of the big problems in agriculture is that, growers get overwhelmed with the amount of information. For instance, I came from a soil moisture monitoring company previously before starting OnFarm, and we’d have companies with 50 or 60 of these sites. We would give them four, five, or six different charts for every site, every day. Growers aren’t data analysts. They’re in their truck 10 or 12 hours a day, you can’t give them 200 plus charts to make a decision off of. So, we digest that data and bring it into a format that allows them to make a very quick decision about their operation, and a more efficient decision.
Stacy Higginbotham 03:08
I think it’s interesting that the agriculture industry and farmers in general are becoming more and more connected. I think it’s interesting because there’s this overall trend of internet and connectivity pervading obviously, outside of Seminole Valley, but outside of the consumer realm and touches some of these industries that I cover in the Green Tech world; energy, transportation, and now agriculture and farming. So, it seems like it’s finally–
Lance Donny 03:38
People would be surprised how connected farmers are. For instance, in the U.S. and Europe, there’s 14 million farms today that use connective technology. That’s a really big market. There’s about 35 million connected devices, but if you drove by a farm, you probably wouldn’t see any of it, because it’s sensor data and it’s other things; it’s devices connected to pumps and valves and very basic connection blocks for data. That will double by 2020. There will be almost 70 million connected devices just in Europe and in America. And Asia, Africa, all those regions are growing rapidly in their need and use for data.
Stacy Higginbotham 04:17
Do you see the robotic sector moving into this at all? Lettuce Bot was kind of famous.
Lance Donny 04:22
I’ve seen some actually. I just watched something from Samsung the other day that had an autonomous tractor that could go and spray and do those things. I think that will take a long time to get there. What you see in farming though, like in the Midwest in what they call Precision Egg, is tractors that drive themselves. Instead of GPS, they use something called RTK, and they plug in a computer model into the tractor and the tractor plants the seed. It measures to the millimeter where that seed went and the tractor does it all. The farmer sits in the tractor and is kind of the co-pilot, but the tractor really does the driving. They plant 40,000 seeds per acre – an acre is about a half a mile in a block – and they know where every seed went. If you think about that from a data perspective, tons of data. If you want to know how your crop did, if you’re tracking where every seed went and you’re tracking what the yield was off of that crop, and then everything that happened to that seed in the process, that’s just a tremendous amount of analytics.
Stacy Higginbotham 05:31
It seems like as more and more data and connectivity enters that space, it’s that whole issue where – I asked this of Bill Rull yesterday from GE the industrial internet that they’re plugging – is that, there’s less and less need for human labor in that equation. Is that something that is kind of controversial that you’ve found in this?
Lance Donny 05:51
Labor, particularly in California is scarce for agriculture. The labor force in agriculture is shrinking, so there’s actually a need to improve what we’re doing from an automation perspective. But what happens in California in what they call specialty crops – the nuts, the vineyards, the fruits and vegetables – that’s very labor intensive when you harvest those things and manage those crops. If you go to the Midwest, it’s very automated. So, in corn and wheat and those things, you’ve got tractors that do a lot of those components. The labor force is shrinking in agriculture, so we actually have to have technology to fill the gap because it’s not necessarily displacing somebody, it’s really filling a need where they don’t have someone to fill it to begin with.
Stacy Higginbotham 06:36
From a sustainability perspective, I don’t know if you talk about that in your marketing, but it seems like if you’re helping farmers reduce water consumption and energy consumption, that would be a sustainable move as well.
Lance Donny 06:50
There’s a natural overlay. We don’t actually have to talk about it too much because it’s a natural component of what we do. Many people don’t understand the scope of water use in agriculture. About 70% of all fresh water is used for agricultural production; that’s a huge number. So, if we can save 5-10% of water in agriculture, we can fill Shasta Lake every year, just in California. Billions and trillions of gallons of water are used in agriculture just in North America. So, when we make a small change to those things, it makes a huge impact globally.
Stacy Higginbotham 07:28
In terms of sustainability, it would seem like there needs to be more and more food produced to meet the population boom coming so you’re looking to help farmers increase their yield.
Lance Donny 07:40
Diets are changing, as well as population growth. All that production can’t just come from North America. In Africa, a large grow might be two or three hectors, or five acres, and that’s a large grow and they’re growing yams or they’re growing something to the effect. Those growers are looking to places like California to help them determine how to grow better. So, all that production can’t necessarily come from the U.S., but if we can take the models in essence, or the expert advice that comes out of these markets, if we can capture that in analytics and use that in other markets, we can help those markets grow as well, and grow better than they do today.
Stacy Higginbotham 08:22
You guys are a two-year-old company?
Lance Donny 08:25
Two years old, Venture backed.
Stacy Higginbotham 08:25
Venture backed. It seems like kind of a hot space for Venture investment. This in 2013, can you define that?
Lance Donny 08:33
Especially with Monsanto acquiring Climate Corp a couple weeks ago now. If you didn’t follow that or didn’t see that, Climate Corp was a Bay Area company dealing with predictive weather data and insurance around the weather data. Monsanto acquired them for $1.1 billion just recently. I had five Venture calls last week [chuckles]. People go, “Okay, we see the exit in farming, in agricultural technology, and it’s around data”, so they’re interested in data plays.
Stacy Higginbotham 09:06
And you’re going to be introducing a data product?
Lance Donny 09:08
For us, it’s really exciting because we climbed the hard side of the mountain, in that we connected all these end devices that are really hard to connect because they’re small companies disparately placed all over the place. We’ve brought all that data in. The next phase for us is to combine that up with some of the big data. So, weather data and other data in which we can layer those technologies on top of each other and provide better accuracy in our predictive models. So, instead of having a relatively accurate set of 60, 70% weather forecast for five days out for a farmer, can we get to 90% or 95% accuracy? What the means is, if the grower is applying water, we’ve probably saved them some irrigation. Also, if it’s a frost or freeze, like for citrus, we can better predict that event, and they can have three days warning for that event, which for them is huge. So, those are the analytics that we can drive to the farmer immediately that have impact.
Stacy Higginbotham 10:14
You were saying you’re looking at these types of companies because they’re clearing an acquisition target, like with Mansanto. What type of verticals do you think would be an acquisition target for a company like–
Lance Donny 10:28
Monsanto, I think they’re really starting a data war for agricultural chemical companies. In acquiring Climate Corp, Bayer, Syngenta, all the other players in this space, they’ve already started going, “Monsanto’s gone in that direction, they know data is important, we need to go that direction”. So, for us, that becomes a more likely– we’re trying to build a great company. If we get acquired in the process, then that’s a whole different thing, but they certainly step up and have an interest in the data side as much as anything else.
Stacy Higginbotham 11:04
So, you think it was kind of a wake up call?
Lance Donny 11:06
It was. We knew it was there, it was just that “Yahoo” moment – if people remember that from the dot com ages. Yahoo’s big success woke everyone up to the dot com value.
Stacy Higginbotham 11:19
Do you think the agriculture industry has been kind of slow moving in making types of–
Lance Donny 11:24
It’s a sleeping giant. People don’t pay attention to agriculture so much, but it’s everywhere. We cannot live without the agricultural production. So, when you have an acquisition like that, that people see from outside as being huge and get the value proposition of the big data side, people understand it’s more than just how you grow a technology company in a really old, sometimes overlooked space. If we can solve that analytics problem, we can make huge impact in the world and build a great company.
Stacy Higginbotham 12:01
I’ve seen the same thing covering the energy sector. The power grid sector kind of woke up a couple years ago and started buying some smart grid analytics firms. It would seem that agriculture and energy would kind of go hand in hand since it’s so energy intensive.
Lance Donny 12:14
Water and energy are correlated. Moving water in agriculture is the most expensive part of the proposition. It’s not the cost of the water, it’s moving it. So, when you make an irrigation decision, it’s not just the energy. If you’re thinking about pumping water out of the ground, some of those wells are 1,500 feet down, and a crop need three to five acre feet of water annually. An acre foot of water is 360,000 gallons of water. That’s per acre, per year, three to five feet of that. That’s a tremendous amount of water, and so most of the cost is moving it. There’s an associate amount of energy in things like fertilizer that have to applied at the same time. So, the energy nexus and water nexus is huge.
Stacy Higginbotham 12:58
And just a total resource crunch, it seems like it’s coming eventually.
Lance Donny 13:02
It’s a huge problem. Look at California, we’re in a drought now. I’m from Fresno, from the Central Valley, and the growers get in that water allocation every year. The next year’s allocation is no water. So, everything is going to come out of the ground, which is really expensive. If we have multiple of these events, that energy cost is going to continue to grow – we’re seeing that all the time – and the cost of farming goes up, which means all the other economics are more important. How do you manage with less?
Stacy Higginbotham 13:31
I’m going to open it up to audience questions if anyone wants to hear more about big data and agriculture, but I’ve got a question. So, do you mostly focus on the U.S. or is international becoming a big market as well for you?
Lance Donny 13:46
We’re just in the U.S. today, but we’ve received interest from just about every continent, because we grow everywhere. So, Europe, Africa, Asia; we’ve talked with companies in those spaces. And we’re a SAS company, so we can deploy anywhere.
Audience Member 1 14:10
I’m just wondering, what do the farmers do for connectivity?
Lance Donny 14:12
That’s a great question. That’s a problem. There is no rural Wi-Fi, so most of the data today comes out via cellular; 2G, 3G, very rarely 4G technology. So, data bandwidth is a problem and in being able to pull data out smartly is a problem, as well. You’re not going to have this huge pipeline where you can move lots and lots of data, and we’re moving data all the time. So, with sensor data, the soil moisture probe, we move data out of the field every 15 minutes, 24/7. So, it’s just this constant flow of data, but it all goes, for the most part, via cellular.
Audience Member 2 14:58
You talked about climbing the mountain, what were the big challenges for that? What could have made that easier? Is it standards? Is it some sort of translation capabilities?
Lance Donny 15:12
There is no standard in agriculture for data exchange. Every company has their own API and their API looks different. Everything from API’s like XML all the way through flat file drops of data. The challenge is that there are very few really big players in this space and lots of little players. So, we use a technology called Thing Works, which is our back-end or an M2M enterprise cloud application, and we have to create independent data connections to all those different devices and then scale after that. So, that means you’ve got to talk to the company, you’ve got to connect to their API, you’ve got to do a lot of leg work in order to enable that data connection to come in. And then once we do, we can turn a company on in a matter of a couple of seconds if they’ve got that sensor.
Audience Member 3 16:04
I wanted to ask about technologies, algorithms. A heavy part of this is your domain knowledge, but tell me about the tools and technology side of things you did or what was crucial to your getting to this stage.
Lance Donny 16:19
The first stage for us what connecting. How do you make the connections? How do you create insight out of just that flow of data? So, for us, it was about UI and UX design as our first phase, because growers, like I said, they’re immobile. Giving them a chart isn’t necessarily the best way to make a decision off of that data. First phase for us was literally, re-think how information is used in agriculture. When we accomplished that, the next phase for us is, how do your do better predictions? When I say better predictions, what I mean is, if I can tell a grower with a high probability that they need to water this ranch for this period of time in this many days, that’s really valuable for them because oftentimes, it’s a guess. The other thing is, there’s lots of things going on the property, so a pump can only operate and provide enough water oftentimes to irrigate one field; they can’t irrigate multiple fields. So, you might have one pump supporting 16 fields. You can’t just turn it on like you do your house and irrigate your lawn, you have to schedule things. And that means that takes into consideration crop need, it takes into consideration weather and the conditions that the crop is under, and what your capabilities are. So, that kind of analytics are really valuable for the grower and the phase that we’re at today.
Lance Donny 17:47
The next phase is really predictive; how can we better predict weather, pest and disease propagation in the field, and those things that require lots of modeling?
Audience Member 3 17:57
Just a quick follow up. So, what kind of public sources of data do you draw from?
Lance Donny 18:01
We use two sources of public data today. One is, the State of California’s got a weather station network that’s public data. And USGS Grological Services has soil data for all the U.S. that we use. Everything else, we pull out independently from all these operations.
Stacy Higginbotham 18:23
We’re about out of time, so any last words you want to give about connecting the farmers of the future?
Lance Donny 18:31
No, I would say if I was in the IOT space, I would look at agriculture as a huge opportunity for farmers to get information. It makes a huge impact in what they do. There are lots of opportunities for engineering around better sensors. So, one of the problems in the space from the sensor side is the cost of sensors and the reliability of sensors. That has to be driven down. I saw presentations where there are $3 and $5 sensors. Most of our sensors in the agriculture space are in the couple hundred to couple thousand dollar range. Lots of engineering could drive the cost of the sensors down, and that drives the use of the number of sensors up, which makes more information, which makes it more interesting for us.
Stacy Higginbotham 19:16
And OnFarm is also helping better manage world resources.
Lance Donny 19:21
Stacy Higginbotham 19:21
Thank you so much, Lance.