Forward Food Tech Podcast

#3 - Jim McDougall: Machine Learning Drone Systems Improving Profitability for Fruit Growers

January 26, 2021 ForwardFood.Tech Season 1 Episode 3
Forward Food Tech Podcast
#3 - Jim McDougall: Machine Learning Drone Systems Improving Profitability for Fruit Growers
Show Notes Transcript

In today's episode, we're speaking with Jim McDougall, Co-Founder and Commercial Director at Outfield.

Outfield enables growers to rapidly image fruit crops with self-flying drones, then receive yield forecasts and productivity maps from a cloud-based machine learning system. 

Jim, along with his co-founder Oli, founded Outfield to make the global fruit industry more productive, more profitable, and more sustainable.  

Jim is Outfield’s Commercial Director and sales lead, engaging with companies and organisations across the fruit and agritech sectors. Before Outfield, Jim worked as a risk consultant advising private equity firms on mergers and acquisitions, with a focus on environmental and social governance. 

Outfield are based in Cambridge, UK and are currently focussed on raising their seed funding round. Enjoy the conversation between Rob and Jim!

Iliyana Dimitrova:

Welcome to the Forward Food Tech podcast. In today's episode, we're speaking with Jim McDougall, Co-Founder and Commercial Director at Outfield. Outfield enables growers to rapidly image fruit crops with self-flying drones, then receive yield forecasts and productivity maps from a cloud-based machine learning system. Jim, along with his co-founder Oli, founded Outfield to make the global fruit industry more productive, more profitable, and more sustainable. Jim is Outfields Commercial Director and sales lead, engaging with companies and organisations across the fruit and agritech sectors. Before Outfield, Jim worked as a risk consultant advising private equity firms on mergers and acquisitions, with a focus on environmental and social governance. Outfield are based in Cambridge, UK and are currently focussed on raising their seed funding round. Enjoy the conversation between Rob and Jim!

Rob Ward:

Hello, Jim, welcome to the Forward Food Tech podcast. This is Jim McDougall, and he is a co-founder at Outfield. Outfield is a really interesting business. Can you tell us what you do,please, Jim?

Jim McDougall:

Yeah, absolutely. So Outfield has been around for a few years now. And what we do is we make the fruit industry more efficient, more productive, and reduce the waste and the environmental impact of growing fruit around the world. The way that that works is we set our growers up with drones is their own drones, and they fly them themselves using maps that we send them. The technology's incredible now, it's all automated, and the drones more or less fly themselves. Then they send us those images and we tell them how many fruit they've got, or how many blossoms they've got, and give them back productivity maps to their orchard, allowing them to make better interventions forecast their yields and sell more fruit.

Rob Ward:

So it's top fruit?

Jim McDougall:

Yeah, so, we focused, focused and focused again to get to apples. So we're doing apples and pears now. But the same rules apply in vineyards, in citrus fruit and in stone fruit, and we're doing trials on all of those at the moment.

Rob Ward:

So as a strawberry grower, I remember going out to count flowers in rows to try and do our crop forecasting. And it was it was pretty awful. And from a level of accuracy perspective. Farmers before they were using your system, what was the level of accuracy on forecasting their top fruit yields? The thresholds, what was that?

Jim McDougall:

So if you ask any one grower, they'll tell you it's more accurate than the average. No one's below average is often the case. But the marketing and distribution desks will tell you that the average yield forecast is off by about 20%, which means that they're working to try and sell fruit about 20% less than what's probably going to come through. So they aren't left short at the end of the season. Because if they oversell and can't hit the quotas that they've sold, they ll lose their customer. Even wor e than that, in New Zealand, the last three years because the cl mate has been changing out t ere, it's becoming harder and h rder to predict, their lobal forecasts for Jazz apples cross New Zealand has been off b 30%. And that's just very h rd to

Rob Ward:

That's fascinating! So you're looking at plus or minus 20%. So it's a 40% potential difference. And when is that, when are the deals struck? I guess it's before they're picke

Jim McDougall:

It is, yeah. So they start negotiating contracts early in the season. And obviously, every grower's different. They'll have different setups and different marketing desks so some go straight into one big company or some distribute through marketing desks, but they'll be finalising contracts through June, July and August, but they won't know what they've got until they get them in the packhouse in October. And there's various factors that affect what you actually get not just how many fruit there are, but also what size they are. So who you can sell them to, the quality of those fruit and the ones you have to reject. Actually 30% of the fruit grown in the UK doesn't make it off the farm because it'll either be under spec or under sold or you know, will never be taken out of storage.

Rob Ward:

As someone who has been in horticulture for most of my life, being under selling, that is actually, should be a criminal offence. And the marketing desks should be put in prison for it, in my opinion. You know, the amount of times that happens, because it's the concern to 'get it sold', pushes prices to a situation to then possibly get the right price or more likely to undersell it. So you were previously looking at a potential variance of 40%. What level have you got to now? What's the best forecast places you've got to actually, from your system, to actually then a picked crop? How far have you got with that?

Jim McDougall:

With a caveat that for some crop structures and for some varieties, we're consistently under 10% error now, which is huge for these guys, because it means that not only can they manage their orchard to that level and get the right number of fruit per tree so they get the right size and the right management. But then exactly as we were just saying, they can be more confident in those forecasts. So they can sell more at a premium price as well, by selling them earlier in the season.

Rob Ward:

So I'm gonna ask you how many tonnes of apples are produced or top fruit produced in the world and if you took a 20% off that, how many tonnes off that would be? Actually don't even bother but it's a lot.

Jim McDougall:

It is a lot, yeah!

Rob Ward:

So what we're looking t is accurate forecast ng. And preparing for this cal , we also talked about other th ngs it does, the mapping echnology, and therefor the precision manageme t of the crop after that. Ju t go in through that because t's really interesting.

Jim McDougall:

Yeah, for sure. One of the biggest value points of this obviously is the sales point. It fits really nicely with the growers management of the crop. Now, I don't know if you've ever been in an orchard with a grower but if you stand there for about two minutes, they will habitually start doing things to their trees. They'll start picking up apples that they don't want to be there or pruning them to encourage growth. Growers do this innately because they know that you can manage a tree to get the perfect yield on a year. Now, a tree will change by up to three fold year on year because they're perennial. So you have to be able to manage individual trees, you know, targeted in a specific way to maximise what you're getting out. Now, there's no way of doing that at the moment really, because one orchard will have 5000 trees in it. And if you go to one farm of orchards, they'll have half a million trees. It's apple trees as far as the eye can see. And a farm that size will have five kinds most of the around. So you can't possibly know what's happening on every row in every corner of the orchard. And even more than an arable because the perennial tree crops are big 3D structures, you can only see about twenty trees at once. And then you forgotten what the first one looked like, you can't see the whole orchard at once. So that's where Outfield comes in. And the first product that we started providing was giving people productivity maps for their orchards. Telling them how many blossoms they had per tree, and then how many fruit they had per tree so they can work to modernise that orchard. At the moment, they'll spray to thin blossoms or to stay in the fruitlets, or to fix blossoms to increase the number per tree. But they'll do that in a blanket spray across the orchard. What we can do is we can tell you spray these five rows not these five, or now we can tell you spray this tree and not the next one, spray less on this tree and more on this tree. Now, that said, that precision spray bit is coming in this year. I don't want to over promise that but we're doing trials in it right now, it's very exciting.

Rob Ward:

This is a lot more than yield mapping. This is almost like a 'sat nav for pruning'! I'm sorry to use a name that probably doesn't work. But yeah, I guess, it is as a farm, you know, a farmer mindset which we all have here, is that, is this over complicating a problem or is this a way to really make a significant difference to the bottom line. I can imagine any farm that is using blanket treatments for anything will not be doing that within five years. So this precision techniques across all forms that has been helped by this 'sat nav' sort of thinking, you know, the computer-aided science is making it possible. It's the level of chemical inputs you can reduce or even much more importantly, the value you can get by that extra detail because it is digitalized. It then becomes efficient to be able to manage it but you couldn't process that as a human being. Having 50 years of experience and labour is a big one in horticulture, or getting hold of good trained labour.

Jim McDougall:

And I've got to say at this point as well, the guys who run the orchards do a phenomenal job of doing huge credit to the growers, because they do such an incredible job of predicting and refining their orchards based on data that is mostly gathered by ISOs. It's a hell of a skill, but there's definitely space in there to improve on it. And that's where new technologies come into it.

Rob Ward:

So just to recap here, we're looking at going from a 40% potential variance in yield forecasting, but also a third of fruit is left on the farm because of quality issue. To turn it on its head, the level of inputs you're putting in and the then, I presume, the and I'd love to talk to you about how your data interchanges with other providers of usages, and how you intermix. And what we're seeing with other ag tech innovations is that the ones that really got these interchangeable platforms, enabling it, making it very easy for other people to use that system. I.e. other, hypothetically, truck spraying or a robot arm that could do pruning or whatever the other is. Because you're not doing that, but you're enabling it to happen, they are the ones that are succeeding. So the ability to be very good at one thing, but also very easy to change other areas, it seems to be the way forward. Talk us through how that works. Maybe an example of a technical partner and all you do is you knowledge share between them, you know, one data centre to another.

Jim McDougall:

One that I know that I can talk about very openly is the Precision Orchard Dosing project, which is an Innovate UK funded research project that Ourfield are quite heavily involved in. We're into the third year of that now and what that's looking at is taking blossom mapping data and using that to precision apply fruitless thinning agent with the view as we were saying earlier to get the right number of fruits on every tree. If you look along any one row the orchard at blossom time you'll have a tree that's covered in blossoms and then you have another one that's very sparse, one that's medium, one that's sparse and so on. They're incredibly variable. And if you can go in and you can apply the right amount of blossom fixing or blossom thinning agents, you can just get the right number of fruit from the tree right from the get go. Outfield doesn't do spraying equipment. We we have no idea how to do any of that. I do have my mechanical engineering degree but it's not my not my area of expertise at all. What you need in order to do any of those things is you need to know how many blossoms are on every tree and you need to know where every tree is accurate to, you know, a few centimetres so you can spray in the right place. And that's what Outfield's system does. With a really cheap drone, in 20 minutes, you can get a survey of your entire orchard, high GPS accuracy, knowing where every tree is informing the mapping for those precision systems, and then also telling them what to apply and where. And the other bits that we need to feed into that is the plant science knowledge to know how much to spray, in which place. So the partners on this project are NP Seymour who builds bespoke spray equipment down in Kent and also NIAB who are doing the plant science aspects and a few others, Hutchinson's with their agronomy background, etc.

Rob Ward:

So it's all very UK based but I hear you're now looking to move abroad and reach out to there. Tell us quickly about that.

Jim McDougall:

Yeah, sure. So we realised early on that in the UK, the season lasts six months, and then everyone goes off into something else for a while. So we made a move quite early in this to also work in South Africa. Same time zone, easy to get to and it means that we're operating all the way through the year because we can jump between the two seasons. The demands and the pressures are very similar internationally, though. And so the fruit industry, everywhere in the world, there are obviously subtle differences. South African trees are four metres tall, whereas UK ones are two metres tall, but the same pressures are there and the same technology is needed to digitise and then optimise the orchards.

Rob Ward:

So how many trees, how many hectares do you think you'll be working with next year or this year?

Jim McDougall:

So last year, we had beta test users in four different countries, we covered 600 hectares which is about 300,000 trees. This year, we're looking to do about 10 times that at least so we're trying to get into the millions this year. And obviously all that data continues to improve our system as well, which is great for us.

Rob Ward:

The big question is what's the bottom line here? What's the financial impact you're having?

Jim McDougall:

So the biggest impact that we can have is improving the productivity for our growers and improving how much they can sell of what they're already making. Growers work in the UK, at least, between a minus two and a plus 9% profit margin. On a really good year, you'll make 9%, on a really bad year, you'll lose money on your orchard. So what we can do is, we can take a 50 tonnes per hectare grower up to 55 tonnes per hectare, just by making sure they've got the right number of fruits on every tree, and then we can tell you in advance what to sell. And then you can sell it for at least 5% or 10% more. And as you're speaking to the marketing desk, probably a bit more

Rob Ward:

What's that per hectare then in cash contribution?

Jim McDougall:

Yeah, sure. So we're looking at taking a profit margins by about 30%. We're looking at adding about 1,000 per hectare in value and it will cost you, at the moment, about 100 per hectare to use our system. And as we're adding new features now...

Rob Ward:

You beat me to my next question, thank you! What cost a 1, they should get 10 back?

Jim McDougall:

Exactly.

Rob Ward:

You know, it's a ten to one ratio, whatever currency you're using. Brilliant. Okay, let's move on what got you into this? What got you, where did this come from? And what other industries have you been involved in to inspire you to then use this into horticulture?

Jim McDougall:

Yeah, brilliant. So before Outfield, I was working in the private equity space doing environmental social governance, which really means looking at the risks that people are taking on when they're buying oil rigs, or chemical facilities or farms. And one of the biggest aspects of that was environmental damage. And we were looking at farmland in the Ukraine, their chemical survey and realising that the data that they had was usually just in a book, if you say to someone, what did you spray last year, what were yield, I just grab a book off the shelf behind them. And, and there's such a huge gap in terms of technologies available there. At the same time, the other founder, Ollie, working in aeronautical engineering space, and he was very aware, he's a huge drones enthusiast. He was very aware of what technology was available, and what can be done with it, particularly around the data capture, and also the machine learning aspects, understanding what's in those images. So we put our heads together and realised that in terms of all the problems that we're seeing coming up in the world, food security, environmental damage, looking at feeding the growing population, that there needed to be more efficiency here, and that we could, you know, find some solutions. So we first started working with NIAB, here in Cambridge, the National Institute of Agriculture and Botany. Plant science researchers who were looking for innovative and technical solutions to be able to do more in their space. They know that they had a big labour shortage because they'll send PhD students or post-docs who know more about plants than I ever will into the field for six weeks at a time to go and count the leaves on plants. And this is not an efficient way of assessing properly. So we started working with them and through them, we got to explore the whole industry from arable crops through to plants, vegetables and ground veg, through to horticulture. We kept on getting dragged back to horticulture, particularly to perennial crops because the value is much higher, because the ability to make interventions is much higher, and because the need was so much more. So over time, we found ourselves very focused on apples and now we're trying other fruits as well.

Rob Ward:

It's the oldest one in the book, but you can't solve the problems in the world of which they've been created. And it sounds like coming from out of the industry completely to look back in, you've really got something exciting here. You know, that's a key message that we have at Foward Food Tech is that, think about technology across all platforms with different sectors. And FinTech is a great inspiration for a lot of food and ag tech trading platforms. There's already a lot of stuff out there that's been done. So and they're very good at it. I live in Central London. And they're pretty good at facial recognition software that's got an iris recognition now with the eye and how this still then interchanges, as a huge technological community and interchanging with other people who have got ideas is the best way forward. It's great what you've done there.

Jim McDougall:

I don't think anyone was thinking when they were designing new technologies that this is how we'd be using it though. The facial recognition technology, I don't think they were thinking we'll be using it to count apples, the drone technology to survey orchards, cloud computing to upload the data to the air, it's an amazing calculus of new and exciting technologies, for sure.

Rob Ward:

They're using the same technology to monitor poultry and welfare issues. So, way ahead of any human. And that's the world that we're talking about now, so long as it makes money on the farm. And it's delivered in a way that's it's simplifying a problem, not just replacing a complex problem with an even more complex answer, then we're into good business here for everybody. Yeah, that's what we believe in. And it sounds like you're on that too.

Jim McDougall:

Any agri tech startup who's in this space is trying to get into this that exactly as you say, you can have an impact, you can change the world, you can improve food security, you can reduce environmental impacts. But first and foremost, it's got to be a business. If it's not affecting someone's bottom line, if it's not making someone money, you can never have the scale or the impact that you need. So those two things I think go hand in hand, it's not impact or businesses, it's impact through business.

Rob Ward:

Good to hear, I couldn't agree with you more. And I can tell you investors are thinking that too and this sort of, one day, it'll be profitable idea that if you know, and it's just not the way forward. And we're talking about the impact for the customer here. Obviously, some some ag tech businesses will take a while to go through the machine learning data to to get critical mass. And that's that's obviously a challenge. So what has been your biggest challenge you think you've overcome?

Jim McDougall:

For us as a business, one thing that we continue having to keep a very close eye on is the diversity of different solutions that are possible in this space, and making sure that we're using ones that, exactly as we're just saying, are affecting the gross bottom line, and providing value to the grower, and that are scalable as well that enough people around the world can use. If I ever speak to any growers about this, they get really excited about, you know, the technology we're providing. And then the next thing that happens is say, oh, could you also do tree row volumes? Could you also track the spraying system in the field? Could you also tell me more about almond crops, there's a million ideas that come out. And some of them are really good, but the growers aren't Outfield, and they don't know how hard some of the things we'll be to do. It might fix a problem for Charles down the road, but it might not fix the problem for growers all around the world. So one of our biggest challenges has been making sure that we stay super focused on what it is that we're delivering, and making sure that we're delivering value to the growers, and not just exciting ideas.

Rob Ward:

That's is so true, well done. Great, because there's so many, it's so easy to get distracted. Single solution for a single problem and then focus on that, brilliant. A bit mean, really, to ask you this. But if you could rewind the tape, the old school videotape of your journey from starting the business at the very beginning to where you are today, and you went back to the beginning, what would be the one thing that you do differently?

Jim McDougall:

That's a really interesting question. If I'd known then what I know now, I would have focused on horticulture a lot sooner. There are a few rabbit holes that we went down exploring products that would have made a perfectly successful business but not a big enough business, if that makes sense. So I think we could have, we could have done that earlier. And I think also just having faith in the speed at which technology comes through. There are some ideas that we've come back to, you know, a yea later realising that technolog from other sectors has gon faster than we thought, the dro e technology being a gre t example. If you go back thr e and a half years, we're buildi g our own drones, crashing our o n drones, all across Kent and East Anglia. Now I can send a rone, I've got a video of my fiv year old god-daughter flyi g a drone automatically at the ouch of a button. That s how the technology has co

Rob Ward:

It's a difficult one to call, isn't it, because you had lots of options and I guess, we push for is get that minimum viable product out there as quickly as possible and fail as quickly as possible but affordably. And then pick the winner that's going to take it forward. So fail as quickly and as affordably as possible. The biggest lesson for any startup. So you've had a fantastic journey and you've got an amazing opportunity and it's really exciting, really look forward to seeing your progress and hopefully having you back here one day to hear how this has moved on. So on your journey, you must have met some interesting people and some really influential people. Who would you like to thank to help you get to where you've got to?

Jim McDougall:

I'm really glad you asked, because I mean, so many people, and there's no way we could have gone down a journey that we have without the support of lots of people around us. I'm going to miss plenty, but I guess the ones that spring to mind immediately are NIAB, particularly Charles Whitfield. And you know, the team down in EMR who have given us tonnes of opportunities and tonnes of insights into plant science and are passionate about the research and the technology. They kind of gave us a start. Belinda Clarke, here in Cambridge in the UK. She's head of Agritech East, or AgriTech-E, they're called. And they're a fantastic network that have connected us to so many people. The Judge Business School, here in Cambridge, we've been on their accelerator programme, got business support there, and they help you concentrate on, sort of, business acumen for years. Lots of growers, but one that really pops into my mind is Tom Hulme, down at A.C. Hulme in Kent, just South of Canterbury, who's really passionate and really enthusiastic and sceptical about everything that we do, which is really powerful. Because he's, you know, he's always keen to see where we can go and also asking challenging questions. So he's been great. We've done a lot of work with him. And yeah, thanks for Tom.

Rob Ward:

That's brilliant. We thank you, because you've been very honest and open about your business. You've got solid metrics that make a big difference to a farming business, the world needs a lot more fruit and it sounds like you're going to help them do it more profitably. It has been a great pleasure to meet you, Jim. We can't wait to hear your next steps. And good luck with hitting your targets for this year. Look forward to speak to you soon. Thank you, Jim.

Jim McDougall:

Thank you, Rob. Real pleasure. Take care!

Rob Ward:

Thank you!

Iliyana Dimitrova:

Thanks very much for listening, and we hope you stay healthy in the midst of this global pandemic. If you enjoyed this episode, please subscribe, give us a five star review and share the podcast with your friends and colleagues. For more information and takeaways from this episode, please visit forwardfood.tech. See you next time!