[Introduction Music 0:00:25.1]
Welcome everyone, my name’s Amanda Scott, I’m the Program Manager of the Farming Together program, and Simone here is the Workshop Facilitator for this evening and a fabulous team member from the program as well. We’re really excited to have the first of a series of seven webinars for you starting tonight on collaborative solutions for farmers, fishers and foresters, and we want to provide the opportunity for you to hear the stories about farmers from farmers themselves. For those of you who don’t know about the Farming Together program, we’ve been running since 2016, and in that time we’ve helped 30,000 farmers, fishers and foresters from across 23 different industries to implement their collaborative solutions.
Our guests tonight Birchip Cropping and Chris Sounness, we have worked with for a number of years now, and Chris was actually one of the IAG members on the Farming Together program. Alright so tonight’s topic is dare to share your data, and I think it’s a really important and hot topic at the moment, we’ve had a lot of interest in this webinar series which has been fantastic. Particularly we want to hear about the data sharing concept from farmer and farmer groups who’ve actually tried to implement that concept themselves, talking particularly I think about finding shared value in data sharing, and all the tribulations and triumphs with going through trying to establish that data cooperative.
As I said we’ve got two amazing guests tonight I would like to introduce, the first one is Philip Guthrie who’s Birchip Cropping Group, innovation portfolio, developing and managing projects that help Birchip members and Australian farmers increase their productivity, profitability and sustainability, so welcome Phil. And our second speaker is Chris Sounness, formally with Birchip Cropping Group. Chris is the Executive Director at Wimmera Development Association, and during his time with Birchip Chris facilitated 40 farmers to connect their weather station data so that it was uploaded to a single shared source. So welcome to you both, we’re really looking forward to your presentation and I’ll hand over to you now.
No worries. Thank you for the introduction and thanks for inviting me back, and I think I’m going first and Philip is actually going to talk about the hard work, but I’m going to talk about the early experience around data sharing. I’ll take you on the journey, I might rush through it a bit but that’s because I think Philip will actually give you far more truth than I present, and mine’s going to be a very Panglossian view of the world. But let’s get underway, can you see the screen? Everyone can see the screen, all good?
Great Chris, yep.
Alright. So I started looking at data Coop when I first joined Birchip Cropping Group and it was about 2014, and I could see there was a lot of interesting farmer data, but whenever you tried to actually find any decent sets of it there wasn’t anything anywhere, it was an un-curated source. So I suppose someone that likes a lot of order and is really very much about process, and those two statements if you know me at all will be absolute lies, it’s not the way I think at all, but I could see there was a real effort–there was a real reward if there were curated data sets. By doing the boring stuff, the librarian sort of tasks, because actually whenever you try and find information it was all over the shop, and that was probably where I’ve started my journey around this data Coop space.
And along the way I picked up a few people who were likeminded and Philip’s one of them, and yeah we’ve had an opportunity to work together. So that’s why Philip gets to tell a story, because I moved on from Birchip Cropping Group to Wimmera Development Association, but the reality is this should all be labelled BCG. So I suppose the piece we looked at was a data AG Coop, and the question we kept on coming up to was it a solution or was it just amplifying problems. So what’s the opportunity, Philip and I did some work with a couple of other partners and we came up and got an economist involved, but there was potentially up to $120 a hectare available in farm data, so that sounds pretty exciting and all of it was available to the farmer.
In an ideal perfect world when you think of a value chain, that $120 dollars a hectare and most regions are millions of hectares in size, you start getting huge scary numbers, and you think it’s great and there’s lots of money available to the farmers. But when you start having conversations about the farmers with this you sort of go to this world here [Movie Clip: ‘Show me the money’] and that’s basically the response every farmer gives you then and still today around data, they think it’s great but they don’t believe any of the things. So what are some of the costs involved or the rewards available, we know there’s some work we did around the cost of poor mobile phone coverage, we’ve identified some challenges there so that’s part of the opportunity, so that was a connectivity piece.
We identified it was a capability piece also, and this is the way I frame it, is if any farmer came to me and said I have $100,000 to invest because they want to get the benefits of digital agriculture, and this was a question I was asking in 2016 or 2017, I think this question’s still prevailing today. Who would you recommend they talk to so they can get the best bang for the $100,000 dollars, so it isn’t just buying the bike shiny thing, but it sets them up foundationally for a 10 or 20 year digital agricultural strategy. And I looked around and I actually found there was very limited capability in this space of who could actually guide, to help farmers set up a digital strategy, which I think is a real shame for our industry that we haven’t developed this capability and be able to help farmers realise the opportunity.
So I think it’s not just a farmer challenge, it’s a whole industry challenge, and I was always thinking well what role do grower groups have in this place. And then we got the trust issues, and back in 2016/2017 when I was really diving into this space around trust there were a lot of discussions about it. I suppose since then probably things have got even more challenging, there’s probably even greater mistrust between various players. But in the end we’ve got to be able to–for these type of institutions that are data Coop to work, we identified you need ethical regulatory issues; you need to understand around ownership. And that ownership of data is an interesting statement, because the reality is you can’t actually own data, but it’s a concept people can understand except when it comes to legality when data doesn’t actually have any legal standing.
You’ve got to be able to offer security and trust, and I think there’s a lot more awareness around that in the industry now. And you’ve got to have a business model that’s viable, and that’s possibly the biggest challenge I faced while I was at BCG, trying to strike that business model, and that was the support we had from Farming Together, and we’ll touch on that shortly. So what are the barriers, there was a lack of standards that are still here today, lack of leadership, and Philip I’m sure will touch on that during his presentation, if he doesn’t he’s sillier than he looks. The technology always promises more than it delivers, and the business case or the value proposition isn’t necessarily clear, and that was the bit that we started with, and that’s the bit that Philip’s been working on and I think he’s getting close to cracking it.
Because it’s very hard when you’ve got a blank sheet of paper to convince someone that’s a business coach you can’t actually–you haven’t got any benchmarks to go off. And that’s the bit I think Philip’s been working really hard on and he’s trying to crack the business case, and it’s very hard to create a business case when there is no other model you can compare it to. So what was the data Coop, we put the concept to the Farming Together program, we were trying to develop a business and understand what legal protections are needed and some of the technical challenges. And technical challenges can always be overcome, and I think it got to the point where we know we can overcome them, it’s just the business model that’s going to find the way to overcome it, and I’m sure Philip will be touching on that shortly.
We wanted to try and develop a data Coop, have 50 growers, and we want to be able to collect machine data sets from harvesters and tractors. And this machine data set coming from harvesters and tractors is really old school, it’s not new technology, but as we did with one trial with CSIRO trying to get those data sets off the farmers was absolutely a huge challenge, and very hard to actually get curated datasets. So what’s a data Coop, a data Coop enables farmers to collectively manage how value from their data is created and shared, a mechanism to actively engage with other participants and build value across the agricultural industry. I suppose we also identified what you need from a market, what the farmers need, what the service providers need, what the users need, what the market operators need and what everyone clearly needs jointly, and all this work was work that the data Coop helped support.
So while I was at BCG we did a lot of the theory, I’m going to flick through a couple of slides quickly because I don’t want to talk for too long, or I might want to talk for too long but I realise Philip will start to realise I’m eating into his time. But we did a lot of theory, but I suppose through the program we didn’t actually get to the point of the viable business model, and the last thing we wanted to launch was a business that was just a whim and didn’t have a sustainable way forward, because that would actually put the whole thing backwards. So we were always wanting to make sure it was a sustainable business if we were going to make it live. So we invested a lot of time in understanding from a whole range of different points of view around what the value proposition is, we also went through what the mission is or why.
We did a lot of work with CSIRO trying to actually get the theory right, and as I say when you’ve got a blank sheet of paper having an understanding of that theory becomes very important. Once again it’s a bit like a data Coop in the first place, it’s not very sexy stuff, it’s a lot of social science work which isn’t everyone’s cup of tea, but in the end it’s actually going to be very valuable in going forward, because it can actually work out where the weaknesses and challenges and opportunities are and what peoples behaviours will be, and I think some of this information has been very valuable going forward. And what the benefits are, where it can be delivered and how it’s stored. So as I say we did a lot of the theory and a lot of the hard work, created all the potential players, mapped out the whole thing.
So really the business case is sort of coming together, but once again the business model was the challenge, we saw even these barriers where it kept on going back to that, what are the motivations, what is the ability, reasonably big barriers, and I might just about leave it there. But the value proposition is as I say it enabled valuable insights from farmer data, and the implications as I say you can read them there now, but enabling farmers to share their data but they have ability to control how it’s shared, and they get valuable insight also, but the benefits can be wider than just the farmers. But the challenge is once you share data once how do you make sure that value isn’t lost, and that’s the challenge.
That’s why Google and everything like that are quite happy to give you Google Maps, because as I say you share your position once but they get a lot of insight from that which they continually resell. So yeah that’s probably where I’ll leave it there, I’ll just put this last slide up. As I say a lot of this is all a theory, but it’s how you actually turn it into a hard piece of work, and I think that’s the piece that Philip’s going to take you through now is making it real. So I suppose as is always my whim I was all about the theory and what could potentially happen, but other people actually got to have the fun job of trying to make it happen. And Philip with that handball I’ll leave it to you, or as a Rugby pass in your language, leave it to you to actually make it happen.
Because as I say I had my five years at Birchip, I can see the value in it, all the people in the industry could see the value in it, even the farmers could see the value in it. But there was just no way to actually provide an incentive enough to change behaviours from everyone concerned to actually make it worthwhile financially or behaviourally or whatever. And around changing behaviours we all know incentives have got to outweigh the pain. So thank you, Philip over to you to turn it into the real world.
Thanks Chris, that was great. If anyone does have any brief questions now Chris are you happy to answer a couple now if anyone has any? I haven’t got any in the chat, otherwise we can come back after Philip’s presented, he might answer some of the questions that you’re thinking of now. Okay, wonderful, thanks again Chris that was excellent, and no doubt we’ll have some questions at the end as well. And are you there Philip?
Oh Olivia’s just put in there what type of data was being collected?
Well as I say it’s generally farms involved in broad acre agriculture, so the data sets were from what the tractors are collecting automatically, what the harvesters are collecting automatically, such as yield and quality and weather data from. We encouraged a number of farmers to buy weather stations, and we worked closely through this weather station concept of trying to educate and encourage farmers to think about would they share the data, why were they sharing it with others. Because if they bought the equipment outright what was their view, were they happy for their neighbour to see it, were they happy for other members of the group who were sharing to see it, were they happy for everyone to see it. And we just tried to use weather data which in theory everyone values, but actually probably has very little market value because there’s so many different ways of getting it, to get them to think through some of the challenges around data sharing, so we use weather data as an educational discussion point.
Thanks Chris. Did you set up groups of like cropping farmers to establish benchmarks?
I reckon that’s best that Philip touches [on it] because as I say I went through the theory, and we talked a lot with other grower groups, but once again we just couldn’t–in the time while I was at Birchip Cropping Group and through the other project, we just couldn’t get the business model over the line, because I think mainly it was the technical challenges, and I know Philip will be talking about that, so I think Philip will be best to be able to answer that.
I think just to preface the presentation, the benchmarking side of things is absolutely critical, but what we’re trying to do at the moment is establish what we’re benchmarking, and what we’re benchmarking from both the farmers perspective and from the industry perspective. Because we’re caught in this loop at the moment, where unless we can get industry investment we’re unable to demonstrate value to farmers through benchmarking, but without the data from farmers we’re unable to create the benchmarks in the first place. So that’s the loop we’re trying to shortcut, and that’s the loop we’ve kind of been stuck in for the last four years and getting close to resolving.
But all of the discussions that we’re having at the moment with the various CRC’s and the RDC’s, it’s benchmarking farmers on a variety of aspects production wise, efficiency wise, and obviously with what’s coming down the pathway at the moment and the funding available, their ability to manage drought and recover from drought is an absolute crucial one. I might just pull up my [presentation] if I can find it, for some reason it’s not popping up in my share list [Short Pause], there we go. Okay, so first just as an introduction for those that don’t know BCG or Birchip Cropping Group, we’re a farming systems group based in north western Victoria centred around Birchip, so the southern Mallee covers the Wimmera Southern Mallee region.
So cropping country kind of extends from the South Australian border up to the Murray River and down to about Ballarat, Bendigo, a fairly large chuck of territory. We have 400 farms as members and 700 individual members within those farms, so multiple generations within family farms, and our farms cover about a million hectares of the north west. Our farmers are predominantly mixed farming, so grains only or grains and livestock on their properties. I think we’ve all heard the hype over the last couple of years about the potential for Ag-Tech, the value of Ag-Data, how much it’s worth, the difference it can make to the industry and what we can do with it, 20.3 [billion dollars] has been banged around by AFI since about 2017 through Precision to Decision.
There’s discussions again around benchmarking and the value of benchmarking, so not only being able to learn from your neighbours and those in similar regions to you, but being able to learn from those in similar conditions across the country, so similar agroecological zones whether they’re in New South Wales, Southern Queensland, Victoria, South Australia or WA. Lift the capability of researchers, so being able to provide access to large quantities of high quality data to boost the amount of research, and I don’t think we can underestimate the amount of duplication in data space and the cost that researchers are going to to collect data for their research projects.
And we’re working with one major university at the moment that last year sent $12,000 dollars to install three soil moisture products, to get three sets of soil moisture data from three paddocks, and what we’re talking about is the opportunity to open up tens of thousands of sets of soil and moisture data. Now that university has its dataset, other universities have their datasets, and they tend to treat those datasets like their own intellectual property, and don’t necessarily share them with each other, so the capability of what we can do by bringing that data up in the RD&E space is phenomenal. And again reducing that silo is absolutely critical, and the discussion at the moment is that we have the potential to lead the world, billions of dollars of value across the sector.
And if, and that’s a big if, we can gain access to farming data, and it’s this notion that it’s a Facebook for agriculture, Facebook, Google, Amazon and TESLA, which isn’t really a car company, TESLA is a data and machine learning company that just happens to use that data to drive vehicles autonomously in some cases. These are industries that didn’t exist years ago, and this is data that’s been created that a lot of us didn’t even anticipate we could use 20 years ago. And just to reiterate, AFI Precision to Decision Australian Farm Institute is the big document at the moment, this notion that unconstrained digital agriculture will deliver dollars for producers, and dollars for the Australian economy.
And the expectations of people we deal with within the industry, you know farmers are just going to throw the money at them. You go to a evokeAG and there’s new companies talking about the technology, how fantastic it is and why farmers should be getting into it and what they can do with it. So obviously farmers are purchasing this stuff left right and centre and installing these technologies on their farms and the flow of data is started. Unfortunately not, that’s not the case, and a good analogue of why that’s not the case is the use of Precision Agriculture in Australia. So this is a study that was done by Terry Griffin, who at BCG talked to Australia in 2017, and that’s just part of the Vesti Fellowship, and he set up a number of surveys and some Twitter work around this.
And what he found was this was the adoption Precision Agriculture across Australia, so obviously very promising that 85% of farmers are using yield monitors, huge adoption of GPS, auto steer section control, and a very high adoption of various tools like yield prophet or sensor spraying, weed mapping, protein monitors and auto drafting for livestock. But the problem is that’s where your data comes from, that we use and is going to potentially unlock that 20.3 billion dollars of value. So soil sampling, imagery, NDVI, imagery in NDVI, weather stations for soil probes etcetera, and even your VR are there, and the challenge with the is that at the moment we have 85% penetration.
So when we surveyed our growers we had 80 farmers turn up to a series of PA workshops and asked them about their uses of yield monitors, only one out of nine farmers was calibrating the yield monitor. So out of that 85% of farmers that are using it less than 10% are calibrating, and what you end up with is a margin of error of about plus or minus 40%, which a farmer can work with, they just need to know whether that yielded well or not, where it yielded well and some rough estimate. But you take data with a margin of error plus or minus 40%, throw it into research, particularly machine learning, and you’re going to have some really interesting stuff coming out the other end, which probably isn’t going to reflect reality or do much use to anyone at all.
As part of those workshops we also surveyed our farmers about where both Precision Agriculture and Ag-Tech, Ag-Data sat on the Gartner Hype Cycle, so the hype cycle as this adoption curve basically of how products are developed and how they’re perceived. We started off with the technology being developed, that drove us to a peak of hype, or what’s called the peak of expectations, where everyone’s very excited about what it can do, and then there’s a staggering realisation that that’s been overhyped, oversold, and can’t actually do much of what people said it would anyway. And you end up in trough of disillusionment, and that’s where you get your amalgamation of companies, products and services being bought up, brought together.
And the real work’s starting, and you start to climb up through what they call the slope of enlightenment to the plateau of productivity, which is where it’s actually working, delivering value and people starting to use it. And as we surveyed our growers we found two things, so looking at Precision Agriculture we were sitting about there, it was real value being demonstrated from PA-Technologies, a lot of growers were starting to use it, and we hoped to see adoption of those technologies that are delivering data increase over the next couple of years, and certainly the Gartner Hype Cycle would indicate that that’s happening, and we’re about to see that happen. When we started to talk about agricultural data and agricultural technology that’s where our farmers put it.
So yes the PA was demonstrated in value, but the data they were collecting from their PA machinery they didn’t really know how to use, their service providers–a lot of cases didn’t know how to use it, the people that sold them the machinery didn’t know how to use it, and as a concept I’d say weren’t getting value out of it. And what then exists there was this massive mismatch between what they were seeing in the marketplace, what they were hearing and what they were actually experiencing. So the good news is we will hopefully see some value out of it at some stage, the bad news is it seems like we’ve got a bit more pain to go through before farmers are really clear on that value and preparing then to adopt.
But the other thing to notice when we talk about data, as I mentioned before Facebook, Amazon, Microsoft, TESLA, all making absolute use of data, making millions, billions, trillions of dollars out of data, but it pays to remind ourselves that farming is a very different industry. As Kennedy once said ‘a farmer is the only person in the economy who buys at retail, sells at wholesale, and pays the freight at both ways’. That means two things for a farmer, one, they are very conscious of cost, they are very conscious of the need for technologies to pay their way, they are very risk adverse when it comes to adapting new technologies that don’t have a clear value case.
And two, it reminds us very clearly why like Facebook and a lot of these organisations, but the data we’re trying to access is privately held in a similar way, but those companies all have very clear value propositions for us to share our data with them. We had some issues with Facebook around privacy and trust, but a lot of us continue to use it because the convenience of staying in touch with people is there, the ability to shop through Facebook is there, there’s a really clear value proposition for us. There are some really significant barriers as Chris alluded to, to realising the potential of data and the perceived risk versus reward isn’t really balanced at the moment, and it doesn’t encourage farmer data sharing.
So on the value front farmers perceive little value to be gained from sharing data, that’s something that continues in all the work we’re doing, and the discussions we’re having in the workshops we’re running are coming up again and again and again, and there’s a couple of reasons for that. So the value for farm data use is poorly articulated, ‘give us our data’, ‘why?’, ‘well we can do something with it that’s going to make you money’, ‘how?’ One of the things we see a lot is Ag-Tech providers coming along and giving us presentations on what they can do and how much money they can make farmers, and how much decision support they can offer. Out of the probably 30 or 40 companies I’ve dealt with in the last 12 months they’ve had some pretty good stuff.
What they haven’t had is when I’ve said okay give me a farmer, give me a contact, show me a case study and how you’ve done this and the farmers who have used it and let me talk to them, I’ve not had one come back to me to date, so that’s an indication of the hype and the promise versus the reality. Farmers take all the risks in this space, they invest in the technologies, they invest in the data collection, they share that data, and for the most part at the moment they are poorly compensated for the cost associated with the data collection and the risk taking, so again the value’s not there. Poor RDC engagement with levy payers, I don’t have to go out on a limb and say this because there was a report that came out and stated this at the end of last year.
It means that the utility of research developed from that data may not necessarily be apparent to farmers. So they’re providing the data that’s going into the RDC that’s going through a five, ten, 15 year research cycle and it’s coming out the end and they can’t apply it anyway so what was the point. The value created along the value chain quite often doesn’t flow back to farmers; more often than not it doesn’t seem to flow back to farmers at all. We have this discussion at the moment about traceability and how it’s going to create market access and protect market access, which is great until you realise that the amount of value being created by the value chain is massive but that’s not going back to the farmer at all.
And despite significant cost, products and services fail to deliver against needs and expectations; our tick at the moment seems to be a solution in search of a problem. Someone’s had a great idea and they’ve developed a product, they’ve launched the product, a farmer looks at it and goes I don’t need that, or I can already do that at no cost, I don’t need to pay $14 a hectare to get a slightly better version of what I’ve already got. And the conversation I had with a lot of Ag-Tech companies is farmers make decisions based on practical wisdom, you know 10, 20, 30, 40 years of working in the same paddock over and over and over again, they know how it operates.
They can walk out onto that paddock, they can show you the soil zones and where they change, they can show you the high yielding parts, they can show you the low yielding parts, they can talk you through why it’s low yielding, and they can talk to you where the frost damage is coming from. And in some cases going to a farmer and saying ‘look you’re making poor decisions you need our product’, it gets their back up because it’s an absolute insult to their business and their knowledge of their business and how they operate. On the flipside farmers perceive significant risk in sharing their data, so there’s apprehension in regards to intellectual property, concerns about loss of data, sovereignty, and the ability to access their own data and use it.
So we’ve already had cases of agronomists who have collected data from farmers who have then sold their company or their practice to a larger organisation, and the farmers gone back to them and said ‘well actually I don’t want to go across to company X, I’d like all my data back thank you very much’. Only to find out that they signed that contract that says the agronomist owns their data and it’s not actually their property anymore, that’s happened in Birchip to a local farmer and it’s happened in a number of other places. There’s a very real fear of data misuse, so a company gathering data for one purpose and then either on-selling it or using it for another.
Again very real cases, companies coming to farming systems groups and asking for a forecast of the yield to be able to improve the logistics to then use those yield forecasts to cut prices, contract prices, and the prices being offered to farmers in advance of harvest, they do that once and then they don’t get a look in again. So I think the point here is there’s also potentially temptation to take a regulatory approach to force farmers to hand their data over for biosecurity or traceability reasons. The NLIS, National Livestock Identification System offers a real lesson of how that operates. Now the NLIS is a fantastic thing, it does protect market access; it provides a lot of biosecurity protections to Australia.
The NLIS has been around since the mid-60s, and it’s only now that it’s starting to gain traction, and the data that’s flying from it is starting to be used in a way that’s meaningful on farm. And a lot of that is because farmers were forced to hand that data over, and they resented it and it took a lot of time to overcome that concern and mistrust. This notion that farming is a risky business and farmers are very tightly guarding their purse strings and very cautious of the investments they make including Ag-Tech. So this is an ABARES graph out of the most recent series of ABARES reports on the effects of drought and climate variability on Australian farms.
And what you’re seeing charted there is farm profit versus pricing and the international markets and rainfall, and you can see that it’s obviously very variable, half the year’s there are negative returns or drops in profitability for farmers. What’s also interesting is just because you’ve got good rainfall doesn’t necessarily mean you’re going to be profitable, and just because you’ve got good market prices doesn’t mean you’re going to be profitable. So again that value proposition needs to be very clear, it needs in some ways to start to mitigate some of these risks rather than add to them. This other notion of mistrust, this is Canadian data, but the Australian situation is very similar.
So what we’re looking at here is the GVP of the Canadian agricultural sector since 1926 through to 2016, the blue line is the gross value of production, the green line is the net farm income, that big blue thing in the middle is farmer expenses, for how much of what a farmer makes goes to someone else. So you can see that that gap has been growing and the net farm income in real terms remains reasonably stagnant or actually even dropped, and there’s a number of years there where it’s actually gone negative. There are two things about this that really concern me, the first is the two big revolutions in farming, the first is the Mechanical Revolution, so tractors, common harvesters, not using horse power, literally horse power to move machinery anymore.
The second is the Green Revolution, so the introduction of farm chemicals, fertilisers, herbicides, pesticides, those are huge for the industry, they have massively increased productivity on the farm, they massively lowered the cost of food for all of us and the availability of food for all of us. But clearly what they didn’t do was make farmers more profitable, what they actually tended to do was by the looks of things, open up the gap between what the industry was worth and what farmers were getting. And what frightens me and a lot of farmers, is we’re currently staring down the barrel of the third revolution in farming which is the Digital Revolution, and we cannot afford to let that gap grow any further.
And we need to ensure that farmers are sharing–not necessarily taking all of the profit, but even if those gaps stay the same that the net farm income is holding at the same rate. The second thing around that is farmer’s share of the food dollar has fallen as well, so the share of retail food items, in this case a loaf of bread has decreased and continues to decrease, so again someone’s making money out of these products, someone’s making money along the way but it’s not the farmer. From the AFI Precision to Decision publication report, we switch across and we start to look at farmer trust of the organisations they’re working with.
So 56% of farmers have no to very little trust in the organisations to maintain or care for their data, 25% have some trust, and less than 20% have a high level of total trust in organisations to use their data appropriately. So when we take those things, that absolute lack of value, seeing someone else take your product or your data and making money out of it, and the fact that you don’t trust them to use that data appropriately in the first place, we’ve got a massive problem that we need to resolve. The way we think we should be doing it and then outlined as you’re a farmer, owned and operated agricultural data marketplace. So providing a secure trusted mechanism to share agricultural data that’s owned by farmers and operated on their behalf but it provides them with a clear value proposition that provides a secure repository to manage and share their data, or not share their data if they choose not to.
And ensure farmers sovereignty over their data so they can share it, they know who they’re sharing it with, and why they’re sharing it, and what they’re getting for it, and that they should have some sense of confidence that it’s not going to be misused or on sold. We’ve been working on four project streams, this is where it starts to evolve from when Chris was talking, and the first of these is around developing a value case for data collection and use. So we are looking at what we call digital agronomists, which is a combination between a data scientist and agronomists, and an agronomist understands where the data’s coming from and how it can be used. It can help the farms collect, collate, clean and interpret their data, to help them develop PA plans to exploit areas of value identified from the data, to identify and develop clear value cases and economic modelling to support data use.
And also to act as a bridge between farmers, researchers, product developers, start-ups working on Research Development Extension Adoption & Commercialisation projects, which is how the Federal Government is now determining what we do. Really that’s just shorthand for ensuring that the researchers understand the needs that farmers are trying to solve, and the research at the end of the day actually delivers what it says it’s going to do, so that we close that loop and show farmers that sharing the data has value to you in an on farm sense. Because it helps you make money and improve your profitability, more likely reduce your cost of operation, but it also has some indirect benefits through better problem solving, better research and better product development.
We’re looking at recruiting three digital agronomists as part of this, so one in New South Wales, one in Victoria and one in WA, and we’re working closely with a number of partners and some State Government partners to get those digital agronomists up and running. The second’s around implementing a farm data repository, so if we’re collecting data we need to store it somewhere, we need to have a mechanism so where farmers are storing it can be accessed, utilised and shared. The main way we’re looking at doing this–you know the repository concept as well, there are data lakes all over the place. We have Agriculture Victoria with a rather sizeable one for its IoT [On-Farm Internet of Things] project, we’ve got Telstra with a fairly sizeable data project as well, so how do we adapt those products, how do we customise those products so we can protect our farmers data.
And what we’re looking at doing as part of this is developing an integrated sandbox, which is a bolt-on that sits on the back of the data repository, and the analytics occur within that controlled environment, so there is no opportunity to export that raw data, you export the results of the analytics or you gain the results of the analytics. And it also provides us with a clear audit trail of whose accessed the data, how it was used, what it was used for to ensure–and again any alarm bells, if something’s not been used correctly. The third thing we’re investigating as part of the federated learning, so the patterns not the data as shared. So that’s a rather vague statement, but in essence what it means is the data is retained locally, so it sits on a non-farm server, or the farms Cloud, or their farm management software, and the algorithms actually reach out through processing locally on farm bringing the results of that back rather than the data itself.
So again it’s just another way of ensuring our members and potential Co-operative members that their data cannot be misused or cloned. A large part of last year was working with a range of partners to develop that, which is information architecture for the data Co-operative and structures to how it’s going to work. We developed this with some assistance from Agriculture Victoria and then validated it with DELL, IBM, Microsoft, Amazon Web Services. And on one interesting occasion took it into Telstra to show them and say ‘hey we think we’ve got something here, we’d be interested in talking to you about building it’, to have most of the people in the room rotate their heads 90 degrees to the left and either laugh or look really horrified.
Because Telstra just spent about 12 million dollars building exactly that, and if you rotate our information architecture 90 degrees to the left it looks pretty damn similar to theirs, so we were very happy about that because we know we don’t have to fill in 12 million dollars to build that, we just need to customise what Telstra’s already built to be able to move forward. The third, collecting the data, storing somewhere, we need rules around how we use it; we need rules around how the Co-operative operates. Recent research with BCG members around their ability and interest in taking part in the Co-operative is positive, 75% of our members are keen to be involved with the Co-operative, 25% vehemently opposed to it.
The vehement opposition comes from the fact that they’ve been burned from Co-operatives before, and the reason they’ve been burned is you have one or two things potentially happen. The large farmers dominate the Co-operative and the small farmers feel left out, or the small farmers dominate the Co-operative, and the large farmers who are in this that’s paying for it feel like they’re not getting value for money. So we need to do some significant work around governance to ensure that that doesn’t occur, that everyone is represented and identify a model that works for all. One of the partners in this is The Business Council of Co-operatives and Mutuals, who are providing some assistance to look at the governance structures.
BCCM is the peak body for Co-ops in Australia, so they have a significant amount of experience in this space, and through the CEO, Melina Morison has been on board right from the start, so a big thanks to BBCM. Also looking at the legal frameworks, contracts, intellectual property protections, data sharing agreements, again the boring stuff that holds everything together, and the development of business models. So data sales, direct value transfer in terms of potential shareholdings and start-ups, or royalties from researcher products and services, and indirect value transfer, which are products and services that farmers can use themselves. So legal frameworks, Griffith University, Leanne Wiseman, who would be one of the most acknowledged people in the agricultural IP space is partnering with us on this, and the business models, the Department of Primary Industries in WA and GGA, the Grower Group Alliance, which is an aggregation of about 70 WA grower groups.
The fourth project and I acknowledge this is complex and I’ll anticipate that question, there’s a lot of moving parts. This is a complex problem; yes it’s an incredibly complex problem that requires a fairly complex solution. The thing for us is to mitigate those moving parts and identify the right people with the right skills to be able to work on each of them. The fourth part of this is research product service development, so the digital agronomist working with the farmers and the researchers again to identify the problems. RD&E projects being spun up and funded through traditional CRC and RDC pathways, we’re talking to Food Agility CRC and the proposed Smarter Region CRC, RDR, GRDC, all very keen to get their hands on the data that comes out of this, and are very keen to be setting up research projects.
And looking at data market too, potential private sector partners, so your input providers, your banks, your insurance companies, on the caveat again that those have clear benefits to farmers and a clear use for farmers. So one of the options being discussed at the moment for future drought fund potentially is benchmarking, looking at identifying elements of risk and then how farmers address that risk, and then working with insurance companies and banks and others to reward those farmers who are managing that risk better, or have practices in place that allow them to mitigate some of that risk. My initial thoughts around this were that farmers would be quite opposed to that, what we’re finding is that the farmers who have the data–and also the farmers who are feeling a bit screwed over, because they’re the ones who aren’t able to get crop insurance at a great rate, because they’re being penalised by what they see as farmers with risky practices.
So we’ve been getting a lot more traction with that than we anticipated, and I think it’s important to note again that we started in grains because that’s where we operate, and that’s where Chris was when he had the initial concept around this. But there is significant potential for this to move into other Ag sectors, livestock is a key one for us, and there are opportunities across horticulture, dairy, and some interest coming out of viticulture as well. Partners and collaborators, as I said we’re leading at the moment we have confirmed partnership and investment from Telstra, Department of Primary Industries at WA, the GGA, FarmLink which is a New South Wales grower group. The BCCM have strong interest in accessing the data, Smarter Regions CRC, the Australian Institute of Shared Learning at the University of Adelaide, the Melbourne Data Analytics Project at the University of Melbourne, the Centre for Digital Agriculture at Curtin University, Griffith University, Leanne Wiseman, and the Graham Centre for Agricultural Innovation at Charles Sturt University.
And one of the service providers Proagrica, so Proagrica basically has a series of API’s that allow us to import and export data in different forms and overlay that data so it can be used. Progress to date, quite a bit; I must leave time for questions. But we’ve developed a range of white papers into the value case, we validated the information architecture, the process of working with the Food Agility CRC on the project proposal, we’re working with Smarter Regions CRC on project proposals, we were going to put a project proposal into round ten of the CRC in September of this year, except unfortunately they’ve decided that for the first time they were going to target CRCP’s and it was going to be around recycling, which ruled us out straight away. So we’re now hoping that round 11 and round 12 will potentially come back to a more open sense.
Some of the challenges to completion we’re still facing, in essence we’re building research infrastructure, and none of the CRC’s or RDC’s actually have building research infrastructure among their guidelines. They are there to commercialise research or to develop cooperative research, but building infrastructure fits somewhere else, so we’re working in a number of areas to get that research infrastructure completed. As I said before the CRC’s and RDC’s are very keen to get their hands on the data. Investor misunderstanding of the sector and misperceptions regarding ease of data collection, I think a lot of companies out there just think that farmers are going to throw them their data on the promise that they’ll do something handy with it in the future.
And also a bit of a misperception that they think a lot of farmers are collecting data as it is and that data’s in a usable format, that’s unfortunately not the case. We as farming systems groups, so BCG, GGA and FarmLink don’t have the resources to fund anything on our own entirety, and for the private sector even though there’s value cases around, the data collection and using the data, actually setting up the infrastructure to move that data around has no direct pathway to return of investment for them. So we’ve tried a number of pathways, we are getting closer to the funding side of things, we had a meeting with Minister [0:50.03.0] the other day to talk through some of this and some of the options available to us.
We have a meeting today with Agriculture Victoria and some of the options available to us, and we’re talking with Leanne Wiseman about the ARC pathway. So I mean if anyone’s out there who’s interested we’re ready to go, we’ve got the whole thing together, we’ve got the infrastructure, the mechanisms to be able to build this, yeah we just seem to be operating in a sector that the RDC’s and CRC’s aren’t necessarily geared up to do what we’re trying to do. That’s me, thank you.
It feels strange without an applause, but thank you very much, that was really really fascinating to listen to that. We did want to open up to the floor now for questions, if you have any questions pop them in the chat and we can direct those to Chris and Philip.
Well while we’re waiting for questions I’ve got one, have you started rolling out the digital agronomists can you say Phil?
Yes and no. So yes we’re kind of doing that unofficially in the background, and then we have a number of digital agronomists identified and who have been testing some of the concepts through various projects that I can’t mention, because some of it’s happening a little bit unofficially. I think that the logic behind the digital agronomists for us is just the serious disconnect between the various partners in this space, with the various organisations in this space, and it’s what they’re trying to do and trying to achieve. So there’s huge value to the commercial sector in accessing that data, but in a lot of cases they don’t necessarily want to pay for it, they expect that the farmers will hand it to them on the understanding that there may be some value for it in the future.
And I said there’s just the mismatch between what they’re producing and what farmers need which is fairly significant. We’ve had interest in funding one element of this project, so we’ve had one of the CRC’s come back and say ‘hey we’d love to fund the digital agronomists and would love to get that up and running’. My response to that is that’s great thank you very much, but just collecting data with no understanding of how we continue to collect that data on a project by project basis doesn’t move us forward, what it does is it just creates another silo of data sitting within the CRC. But they have a clear understanding of how we collect that data on a large basis and share it, so I think there will come a point when we just say look we have to do that, just to get this done we need to break it up, we need to do it project by project, but we’ve still got a couple of irons in the fire before we get to that point.
Great, thanks Phil. So we’ve got a comment, really impressed. And we will be posting a recording on our website to be followed up so we can let you know about that. So I’ve got a question, what sort of funding and runway do you have right now?
At the moment we’ve got about a million and a half dollars in co-investment sitting there waiting. Most of that kind of investment is being provided to us on the understanding that that’s going to be matched from a CRC and RDC or another source, so we’re working very closely with a couple of RDC’s and CRC’s to identify that. I’m reasonably confident that we can get over that barrier of falling between the cracks, but we just have to work right and potentially send the project in a different direction to get there. One of the approaches to Food Agility CRC is potentially we just pitch ourselves as a company that’s aggregating and selling farmer data, which there are a number of out there at the moment and they seem to be worth a fairly substantial chuck of money. That’s not the intent, but if we have to go down that path and do that in the first instance to be able to get this up and running that’s one approach.
And just to follow up on–I’m just looking at Nathan’s question, what makes a good digital agronomist and are you looking at medical or other domains. A good digital agronomist, first part of that question, obviously a very clear understanding of agronomy, how farms operate on the ground, what you need to do to grow a good crop and grow a profitable crop, not necessarily talking about just a good yielding crop, but a profitable crop, because there is a difference. And they need an understanding of the data, how to collect the data, how to use that data, and at the moment those people are few and far between.
We do have one Adrian Roles sitting on here at the moment unless he’s–yeah no he’s anticipated the question and dropped off so I couldn’t drop him anything, fantastic. But the second side of things medical and other domains, so those that understand data dimensions, absolutely, being able to allow patients or allow people to share their medical records for medical research, to have control over how that’s utilised, have control over who accesses that and under what terms. And that’s certainly something we’re considering moving into other domains, that’s one of the key value cases outside of agriculture. In a previous life I worked in the pharmaceutical space, and probably had I–had Chris wandered around I would have stumbled in the medical space rather than the Ag space in the first instance.
We’ve got one from Pamela from before, now do you set up groups of like cropping farmers to establish benchmarks?
Yes, I think I addressed that before the presentation. But the way we see this benchmarking is absolutely key, that the first side of things is around the data use, but the benchmarking comes in when as a farmer you’ve got your basic agronomy and you’re looking forward to what’s next, what are my constraints, what are other farmers potentially doing that’s working for them, what are the other farming systems I could be looking into that fit in with my agroecological zone, with the reality on the ground with my climate, that’s when benchmarking becomes absolutely critical. Better understanding your farm and operation by looking at what others are doing, and then potentially looking at what they’re doing differently and the results that’s having.
Benchmarking on its own doesn’t seem to be enough of a value proposition for farmers because again the effort to go to to collect that data, to aggregate that data, to clean that data is immense. The project Chris referred to earlier where CSIRO came and did some work for us, it took them five days per farmer to do that. So if you’re looking for our project where we’re working with 90 farmers five days each with three digital agronomists, you’re still looking at half a year just to clean that data and make it useful and available. And then farmers aren’t going to pay that cost without a clear value case, so everything comes back to value case because value case is why they collect the data in the first place, once they’re collecting it we can discuss what we do with it.
Actually Philip it’s interesting, one of our other webinars we have coming up in a few weeks’ time is from a few benchmarking groups who have been benchmarking for a number of decades, so it would actually be interesting to ask them that question about how long it takes them to prepare the data for their benchmarking.
Look it depends on whether they’re using data for benchmarking, what degree of data they’re using for their benchmarking, because as I said earlier a lot of farmers are doing this without collecting a hell of a lot of data in the first place, it’s all sitting up here. They’re out at harvest and they’re looking at their yields and they’re thinking about what they did during the year, and what’s happened in previous years, what the potential causes for that are, and then they get together and discuss that without ever a spreadsheet changing hands.
I’ve got Tania, what low hanging fruit whose cases can demonstrate value to grower and the concept to the RDC’s and the CRC’s?
I’m beginning to wish I hadn’t invited Tania to this webinar [Laughs]. Look the really low hanging value cases are the free data that farmers are already collecting. So a really good example of potential use of free data and low hanging fruit is instead of paying thousands of dollars per hectare to get soil testing done, Google Maps at the right time of year will give you a very clear indication of where your soil zones change and potentially where you need to be testing. And your internal data off your tractor, so as you’re pulling the seeding rig through the ground, as it’s going from sandy soils to clay soils the strain on the engine goes up and down. So you print out your engine load map, you overlay that with your Google Earth reference and potentially throw DVI over the top of it, and then you’ve got a fairly good data set to start making decisions from.
Adrian’s just sent me a text just pointing out the benchmarking data is not spatial data, thank you Adrian. So I think in this case it’s making use of the data that they’re already collecting, that they may not have been shown how to use or the value of it as the low hanging fruit, and it’s really helping them inform the PA decision making. And again one of the big learnings for me over the last 18 months working in this space and getting my head around PA and how it operates, is Precision Agriculture value from it is not necessarily around growing more profitable crops, it’s around using your resources more efficiently, and it’s around the potential savings in terms of your logistics. And that’s around actually reducing the cost of producing that crop, rather than increasing the amount of crop you’re producing. So that to me is low hanging fruit, and Tania and I can argue over that a little bit later on.
Thanks Phil. We’ve got Ruth; on your last slide can you give a couple of examples of what you’re looking for from a new potential partner?
I think that the co-investment around actually building and customising that data repository is our core area of challenge at the moment. So output around finding partners to undertake the research or the product development not a challenge at all, GRDC, MLA, Integrity Systems, Food Agility CRC, Smarter Regions CRC, all have expressed a great deal of excitement for what they think they can do with the data once we’ve got it. It’s around actually getting that data, amalgamating that data, and having it in the system that we can share that data repeatedly and widely that are the challenges at the moment. And apart from things like [1:01:58.6] which is specifically there to build a research infrastructure, that’s the area we’re struggling to find at the moment. We’ve got a million plus dollars sitting there from Telstra on the table, we really need someone to match that, and once we’ve got that we’re off and running.
Thanks Phil. Will you be developing data collection standards to farmers?
We’ve been working with farmers to help them collect their data, I don’t think–I mean there’s already organisations out there like Fed uni [Federation University Australia] that are working on data standards and collection standards and usability standards, I don’t think we want to add another to what’s already a fairly significant and confusing pile. It’s about sifting through those and identifying the right ones, the ones that work for us, the ones that work for the industry, the ones that make that data easy to move around and don’t lock it down. One of the challenges I have with data standards is it’s the minute you decide that you’re going to use a specific standard to something you tend to strip the richness out of it. And the example I use is Esperanto, I don’t know if anyone remembers Esperanto, but it was a broad attempt in the 60’s, 70’s and 80s to develop an international language so we would all communicate together.
And that worked, you had this mismatch of different languages that were thrown together, but the problem is there are things that you can say in French that you certainly can’t say in Esperanto, there are things you can say in Italian and English that you can’t say in any other language. And one of my studies at university were weird things like philosophy and ethics, and language is a really key part of why French philosophy is French philosophy, because you think in a particular way because it’s the language you’re using, the minute you have that universal metadata standard you lose the richness in the language you’re working with. And what tends to happen is you’ve developed something like Dublin Core which is a universal metadata standard rate, and everyone starts to agree with that and says ‘oh there’s things I can’t do with Dublin Core so I’m just going to adapt it slightly’. And they move it and they change and they add to it, and within ten years you’re at the point where you’ve got various versions of Dublin Core and you’re back in the same situation in the first place. So yes not looking to recreate new standards, looking to adapt what’s there and use something that’s already there.
Right. And Pamela’s asked, in terms of drought resilience what data measures are you looking at? Soil carbon, soil humus and water retention, and are these measurable with current data technologies?
Yeah a really good question. It depends on where you’re farming, so the data you’re going to be looking at, the foothills of the Snowy Mountains is going to be slightly different to what you’re looking at in very dry low rainfall zones. Obviously water retention is a huge one in there, data around soil types, data around soil compaction, climate data, a lot of stuff around practice data. The fact that you have farmers as next door neighbours and one farmer is much more profitable than the other, and one farmer is much more resilient than the other, kind of indicates that it’s not just your physical characteristics that impact on your ability to be resilient to drought and to recover from drought. And I think that’s a challenge when we’re doing this stuff at the moment, is we’ve spent a lot of time looking at the biophysical characteristics and not necessarily got our heads around the farming practice side of things, and what those farmers are doing differently with what they have.
Fantastic, and a big thanks from Muriel Hobbs to both you Philip and Chris for your presentations tonight.
If I can I might just throw briefly that soil carbon question to Chris, because he’s worked in that space a hell of a lot longer than I have.
I think Pamela’s raised a good question. I think part of the challenge is around measurement, and as Philip highlighted I think in different farming systems what’s measured–range is important. So the drought resilience I think this is an interesting space, because I think the soil stuff is very important in the drought resilience, but it is important for that particular place of what you measure. So I do know that if you were in 800mL rainfall and you miss out on 50mL of rain compared to a 300mL rain and you miss out on 50mL of rain, there’s a big difference there percentage wise. But actually resilience wise it might be that the 50mL less in the 300mL system may well be not as impactful, because of the low input system they’ve designed and they’re used to variability. So yeah the drought resilience in the soil is a very interesting question and I think it’s not straightforward, but I think the most important bit and I think Pamela’s raised, is the measurability of the data. Before we can even measure those things you’ve actually got to get the tools that can actually meaningfully measure, particularly soil carbon, because it’s so stable and generally weathered soils that are millions of years old the soil carbon is very hard to change in most low rainfall, medium low rainfall environments.
A really good example of how that can operate, one of the projects we’ve been working on for a little while is just trying to understand farming practices, and understand the decisions that farmers make throughout the year. So at harvest or not at harvest, you have a series of critical decisions and a series of inputs, data that come and inform those decisions. You have the same throughout the growing season in terms of putting fertiliser on, how much do I put on, am I comfortable that I’m getting enough rainfall, am I close enough to a rainfall that’s going to wash into the soil, etcetera, etcetera, etcetera. What really surprises me is that I don’t see any research at the moment where people have actually mapped those decisions farmers are making and the inputs, the data they’re inputting into those decisions.
The types of senses that we need to collect that data, and how we need to feed that back to farmers to improve their decision making. What we tend to have is a whole bunch of NDVI and algorithms running over those NDVI images that may or may not be ground truth, spitting out results of really varying quality. So again understanding those decisions, understanding what farmers are doing and why they’re doing them will then tell us what data we need to be collecting, whether it’s soil carbon, or as to Pamela’s question whether it’s humus and water retention or whatever. I just think at the moment we’re collecting a lot of data without really clearly understanding why we’re collecting it, and that’s leading to a lot of waste and a lot of duplicated effort. Look it may be leading to a lot of research publications and people getting tenner within universities but it’s the practical outputs on the ground that we’re keen to see more of.
And I think that’s probably a great way to finish off the discussion. There’s a few more questions, there’s a lot more questions hanging over where to from here. I just would like to say I find it really interesting that the two fundamental issues that I know Chris spoke of in the really early days about value and trust are still some of the biggest challenges, issues that are faced in moving forward in this data cooperative space. But it sounds like there’s so much amazing work that’s being done Phil, so congratulations to you with these big leaps forward, and we really look forward to staying in touch and hearing about the progress.
If there is anyone out there that also sees opportunities to connect we can provide those details for Phil and Chris, and we’d also be happy to help facilitate any of those introductions and connections.
Lovely. Thank you it’s been a pleasure.
Yeah. So thank you, and thanks for the great questions from our audience too for our first webinar, I think you’ve been a fantastic interactive audience, so thank you very much. But next week same time same place, we have another collaborative solutions webinar topic which is about Out of the Box solutions. So these are about two amazing women in agriculture who’ve come up with two different ways to connect community directly with the growers and producers in their region, through Out of the Box food delivery and through innovative market concepts. So we’d love you to join us, if you enjoyed this week’s we’d love you to join us next week, same time same place. Once again thanks to Phil and Chris for such a fascinating discussion and we hope to all see you again soon.
[End of Webinar]