 Well, thank you very much and thanks for having me and thanks to all the people I can't see for listening. The, I'll say just a little bit, but now I can see some of you. A little bit about the for 11 years from 1990 to 2001. I directed a program called implementing policy change that advise ministers of government around implementing large scale policy change. And over that, that period about 40 countries, or perhaps two or 300 policies that we're helping people on the change management of. And the reason the rationale behind that was the recognition, not unlike the conversation and scaling that a lot of things appeared on paper, usually as written approved policies or legislation. Didn't translate themselves all the way through into performance the way people had hoped and expected. At the end of that period, McArthur foundation. Asked me asked us if we could flip that experience inside out and instead of looking at change from the top down. Look at it from the innovation up, but with the same question in mind, namely, how does large system change happen. So for the last years. I'm preoccupied with this issue or if you count the 11 years before that you could say for the last 27 years. So it's an issue I think is important and I guess I'm a slow learner, but I do feel like I've got a sense now as to the dynamics of what's going on in this kind of change process and a healthy respect for the complexity of it. So I'm going to say just a little bit by way of definition then I'm going to try and take you through the logic of scaling and try to with some reflections on what it means for people in research enterprises like yours. So I'll begin with it with the definition of I might the definition that we use came from my colleague and friend Johannes Lynn and aren't a heart man. And it reads like this scaling is the process. Lottie, are you are you sharing your. No, I just wanted to check that we weren't missing something. This is the last thing before the slides. Scaling up is the process of expanding, adapting and sustaining successful policies and practices. Geographic space and over time to reach a greater number of people. I think that's pretty close to the intuitive notion of what scaling is, but the. I would have you take away from that is it sometimes he's. I'll see sometimes their their practices, but the issue includes not getting them to large numbers, but getting them in sustainable ways. The large numbers of people and in this case means that they stand the test of time, not test of space. So the for obvious reasons that gets you deeply where those incentives lie and also into the issue of systems. And it's those 2 links. I'm going to try to flesh out for you just a little bit. Now I am going to go to the slides. I'll begin with what you might call the true test on this, which is outcomes that match the size of the sustain. And I'll tell you a kind of a Rorschach test way of thinking about this. When I've used this same graphic with government officials or sometimes with large corporations that focus on. White space to the of the bar, the businesses see that as. Market and the government government see that as the unmet need. When I show this to people in the development business, they usually focus on the little orange bars on the left, which is the private. And I also a legitimate way about change. And when 1 of those little bars moves a bit to the right, we're quite proud of ourselves. And again, I think appropriately, but not if you're focusing on what I would call the denominator, which is the problem. And so this approach begins every conversation about any change with an attention of what's the scale of the need. And what's the plan at least over time or hypothetically for trying to reach that need. I said, think about that from the point of view of businesses and governments. It's kind of part of their deal. Businesses think of that as addressed market governments think of it as relevant population. But if instead you think about it from the point of view of the world that many of us live in a donor project world that works. From 3 year plan to 3 year plan or 5 year plan to 5 year plan or in the worst case plan to 1 year plan. And the reward for the project is often and most typically another project. That's the world. I know many of you live in. I can tell you what I lived in for 35 years. And that world was constructed with, I'll say this with a maya culpa back in the 1960s when we did the most efficient way to deliver foreign assistance in most cases. The most efficient way to deliver foreign assistance in projects and the most appropriate duration for a project. And we conformed everything to that. So if you work. Pick your donor of preference work backwards. Their strategic plan. Probably of not much longer duration than that. And they're probably composed largely or at least. That means that the way the way we train now to think about change is what is the project achieve and how does the project get picked up by a subsequent. The that's. Contrast to the scaling mentality, which looks at the system. What was the system prior to the project? What's the system going to be after the project? How is the project or the intervention of the research material alteration of the after as compared to the before? Give you a couple more things that compound this complexity. The 1st, both says double double half half in the last. The number of official donors has just about doubled. And if you add in the many foundations that are involved, it's a larger multiple than that. As a result, the number of projects is just about doubled. The average size of project is just the average duration is just about half. So over that period of time, we have more and more smaller and smaller shorter and shorter. Financing bundles chasing the same big problems. We've been chasing all the law. It's maybe not a surprise was a disappointing surprise to me. When I went through a quite a large sample of projects that considered themselves successful at the R and D or the pilot phase that you'd have to stretch to say that 5% of those ever went to scale. 5% in my opinion is a generous estimate. So if you look at it that way, that means 19 out of 20, not of the things we do and claim success. Ever reach scale. That's not a good track record. Even venture capital is still for 10%. So as a minimum, we ought to be able to double our success rate. And I want to suggest at least a couple of guidelines for how people have tried to do that. Third metric that I think adds complexity to this is my own analysis suggests that the typical amount of time without the do scale successfully to get to scale is a 15 years. And there are exceptions to that. And particularly with the information revolution, you sometimes get remarkable exceptions to that. And even there, when you know the full backstory 15 years is not an unreasonable estimate of how long it takes an innovation to get to scale under good circumstances. So that means the timeline for imagining a pathway to see if it's realistic is probably not much shorter than that. So one to seven or one to 20. This is the point I'm trying to illustrate here is that the relevance of forms in general and by that I include the kind of money that funds your centers. That money had become a lesser and lesser percentage money going into these sorts of. If you look at it as foreign assistance to foreign direct investment. It's now 1 to 7 if you look at it for an assistance to domestic resource mobilization, meaning local taxes in the countries where we work. It's now 1 to 20. I'll give you an extreme case of that it would. In 1977. Indonesia's ODA was official development assistant was 43 and a half percent central government spending in 2016. It was 0.08%. So think of it as. From $1, $2.5 to $1 out of $1,250 trying to make is that the things we're doing while important are no longer the dominant story. If things scale and particularly if they're going to be sustained at sale. It's only going to happen if the commercial private sector and or the governments, the countries where we work take these things up and integrate them into their new normal. If we're taking our kind of research and prototypes and projects and trying to imagine how do they. In the world that I'm describing, I want to say that there are 4 features that really help to think about it. The 1st is to really understand how what we're doing. It's into our understanding of what could potentially make the. For those of you who've studied diffusion of innovation, a lot of which began with work in the agricultural sector. There's this magic that happens where it goes slow, slow and then also. It goes fast, fast, fast. That's the verticality. But there's a bit of a mystery about what actually initiates that verticality. If you read popular books like tipping point, what they said was it essentially it's a contagion factor. There's a critical, you hit a tipping point. And if the thing that everyone else is doing is what else wants to do. That I think is representation of what happens in things like consumer goods and also in some agricultural practices. But for the most part, when I've seen things go vertical, it's because something has changed not on the demand side, but on the supply side. Something happened in government policy pricing in availability that made a material change in people's willingness and ability to use that. That's often because the buyer and the the payer and the user are not 1 in the same person or at least the role of subsidies plays an important role in the process. The 1st thing. If we're using projects is to say what what's the constraint that could potentially if unlocked, allow this verticality to happen. And is we can do in a short term 3 to 5 year intervention. Materially affect that constraint may college const. Maybe it's a systemic constraint. Maybe it's a subsidy constraint. The 2nd related is to say what's the link between in your case, the research. It's a more general a short term intention and systems change. So there's the project, the prototype, the analogy, the innovation over on the left hand side. As I said, all too typically what happens is that those things become gateways to other interventions and research. If they're going to move to sustainable change, they need to do something for the systems that are applied by governments and markets and in a very light type philanthropy. What does it mean to affect the system? If I was simplifying that, I'd say it's either trying to change policy and incentives or capacity. Is there something we're doing. Change those things. And if so, what beyond our research is going to be necessary in order to make those changes real. Excuse me. Here's the 3rd point is that in every piece of large system change that I've studied. There's an element that I think gets less attention deserves. We have a gear on the left hand side that I'm calling innovation. That gear I would submit to you is actually spinning quite well and I know from your perspectives and others more different and better could help. But really the innovation system globally is doing quite a remarkable job of generating new ideas and new methodologies. On the right hand side, the world now has quite an array of delivery systems. As I said, mostly governments and there are very few places you can go where they're not commodities for sale. And there are very few places you can go where government hasn't penetrated in some way or another thing that's missing and missing in fairly dramatic form. The gateway, the thing that was innovations into ongoing platforms of delivery. If you think about this for commercial activities, particularly very profitable ones that intermediation is primarily performed by venture capital investment banking, things of that sort. But if you're in the pro-poor world that intermediate is usually uncompensated function, the innovators can't pay for it and the deliverers can't pay for it. So unless owners or someone else targets that as an area of activity, it tends not to happen and therefore you get innovations going on scale and going unimproved. When I say intermediation, what do I mean? I mean things like this. I mean things like system strengthening, change management, fundraising, packaging, advocacy, convening. The list could go on beyond that. But everything that it takes to like successful innovation by successful, I mean effective, concept effective innovation, widespread practice. And any of you who've ever been directly involved in policy change or in just making systems reform know that this kind of thing is very difficult, whether you're talking about something in Australia, in the UK, in the United States, in Germany, or in Kenya. Basically, it involves large bureaucratic systems, be they corporate or individual, or a large number of small entrepreneurs having to adopt a set of practices and be able to deliver them to large numbers of people. This does not happen easily whether you're in the education sector, the health sector, or in this case the agriculture sector. It takes a lot of momentum to make that happen. And a lot of work, at least equivalent to the work necessary to generate the innovation in the first place. Finally, the fourth that projects can do is they can think about how we can use research and an experimentation to move closer to basically fill in some of the gap between innovation. And the first thing I would suggest to you is to think this way, you have to get beyond the design proof of concept rollout paradigm. That would suggest all we need to do is figure it out, test it, prove its efficacy, and then it's simply a question of innovation and rollout. Having now worked with 400 scaling innovations, I can tell you I haven't seen one, not one. What happens in the section is that in the course of rolling things out, the contextual out, and you're having to adapt them to the fact that people don't behave perfectly, and that the people who are implementing the people you greater control over, and that the incentive structures are different, and that farmers use it partially but not totally. And all of that requires a much more adaptive engagement than the design proof of concept rollout paradigm tends to assume. So if you say, well, how do we use projects to get closer on that? I would say there are at least four things. One is you can use the research phase, early phases, not just to develop the thing, but to do everything you can. You build up the evidence base that a real decision maker, meaning a minister of finance or the head of a company, would need in order to make that decision. That means envisioning the audience for information that's not a scientific audience. It's basically a very hands-on audience and not dumbing it down, but trying to figure out how will we present information that would be compelling to them, or they would at least answer the questions that they have. Directly related to that is this idea of simplification, and I often say that scaling is a game of subtraction, not a game of addition, because when we're in a research phase, just as Ian said before, if the addition of something more would help make something successful, we typically do it. Every one of those things we add for purposes of efficacy complicates things for purposes of scaling. And so the process of scaling is to experiment with, could less, could simpler, could we do it with half implemented, and so on. Second thing is, if we're going to hear the scaling, it's quicker than we normally would to transition to the direct engagement of the people who would be doing the delivery at scale, be those commercial seed companies, pharma cooperatives, government agencies, or large corporations. So there's a lot of weight and assumption that will develop it, and once it's really good and ready, then we'll crosswalk it, engage earlier than you think you might need to with the people who would have to adopt it at the end of the day. Third is focusing every possible ounce of attention you can at cost. The, a penny or a quarter of a penny can make a big difference and whether things scale or not. And also on the implications for current providers, if what you're doing has the inadvertent or the effect of displacing some people who things right now, they will work hard to resist the change. Typically, the, the status quo forces are stronger than the forces of change. They're trying to figure some way that this is less threatening to existing providers actors is usually an important part of trying to move it at scale. And finally, obsessing about the weakest link looking where could this possibly go wrong and what more research or evidence or engagement could we have to deal with that. So think about something like, like, or maze and ask yourself, if you were thinking about it this way, where all the steps that could prevent this as effective as the actual seed itself is from reaching the scale that you're hoping. And I would suggest you that most of those are not agronomic. Some of them are, but most of them are not. I'm trying to figure out how far the people who work on drought tolerant maze cannon should go. That other range of issues are in your case, I could have put up a cow on this and said how far do you go to anticipate and engage some of the issues either as a researcher, or as a betas between research and other people that would really make a difference. Let me give you another example, then frequently talk about bags. It's a huge success now lots of countries, lots of obvious data on the benefits. But it sat on the shelf for 20 years as an innovation until somebody figured out how to get local manufacturers and producing these things and doing it in a way that was really cost effective and got them into the hands that wanted needed them. So a great technology when I'm used more or less for two decade with the try to address these issues. Billy and I produced this source book it's got nine chapters that's certainly available online. And if some of you want hard copies I think we can easily make that happen as well. We distributed these at a hard launch at the the agrf last week. I'm just going to hit points from that is that it includes a framework this happens to be that the framework I know best and the most associated with for trying to think about how to plan and manage the pathways to scale. It breaks it down into 10 discrete tasks that are in three steps. The first step has to do with how you plan with scale in mind. The second step has to do with how you we call preconditions but basically get the decision by governments and corporations. That would be necessary in order to scale thing. And the third step has to do with actually managing the scaling process. Another thing the source book lays out is some new tools for assessing scale. The variety of them. The one that I know best is a 32 item checklist that we then elaborated for the USA IDs Bureau for food security looks at another actors directly affect the scale ability of different intervention. There are several other simit has done an interesting one. For other side in the in the source book, all of them basically address the same four dimensions. Characteristics of certain interventions that make them more or less scalable. If so, what can you do about it. Are there characteristics of the organizations that do the scaling. Make a difference. If so, what can you do about that. The enabling environment that we know are going to affect it. What are those and how do you do something about those and finally contextual factors like the degree. And how does that affect the process. The piece of this source book that I want your attention is it Spain's quite a bit of thinking about metrics monitoring and evaluation. And it says that when we look at innovation, we usually look at what here is called tier one, which is proof of concept validating the model or the intervention pilot testing it impact evaluation prototype and so on. But we don't do near job with tier two and tier three. Your to stuff necessary for refinement, streamlining and assessing scalability. Clues things like robustness and what we call second stage pilots second stage pilots means if it works under these conditions will it work under those conditions if it works at this level of funding will it work at that level of funding if it works here. There. And of course cost efficiency and alternatives and the change. The one that I think is attended to the least. Looks at what kind of metrics do we need to monitor in fidelity during scale up. It also looks by the way at the relationship between fidelity and adaptation. Validation of scale, continuous improvement and so on. In this series of discussions about the effects of markets financing and enabling it. I won't try to run through just to let you know that we try to give some examples that we think are good examples where people really did use for inclusive scaling. We also the typical devices for the creative financing of scaling and some factors in the need to make a difference, particularly as they relate to partner policy behavioral change institutional reform. Two more slides. I want to give you seven conclusions and recommendations that came out of the source book. The first is we were asked by the organizers to talk about commercial pathways to scale. One of the things that we concluded is there is no such thing as a fully commercial pathway to scale. No agricultural innovation that scaled was in regulation don't make a big difference. So even if government isn't directly provide the role is critical in this and their engagement and with the whole process as I know is a central feature, but we've tended to differentiate in many quarters between the commercial pathway pathway. Our finding is that the two are very intertwined. The second is that the concept of implementing partners that we have as often to. And particularly people and organizations that are involved in things like equipment leasing input provision product aggregation. These need to be there. They're obviously part of the value chain, but they need to be. Not just the people who receive whatever it is that we produced. And you source book several good example exactly that are the 3rd as I mentioned before. Is it the most vexing bottle scaling innovations are usually non technological non agronomic. Nature 4th, which I haven't mentioned so far. Poor farmers time horizons tend to be extremely short for you and me. If we were to make a mistake, it probably costs us a few percentage points on our return. If they make a mistake, it might cost them a child. And so they prioritize minimizing risk over math. And that means unless our solutions include some way of indemnifying them against some portion of the rift. Unlikely to go to scale quickly. The 5th again, when I've not mentioned so far is. In order to get things to effectively. We need to allow or even encourage monopolies. Of something being licensed to do something. Is in our experience often inevitable to try to get the efficiency into the supply chain. That was scaling to happen, but it almost always presents a challenge later. So at some point, there needs to be a strategy. I call it having your eye on the exit ramp. So you've got a plan for how to extricate yourself. Broaden yourself from the monopoly that you yourself. Six, the projects and the innovators for the most part saw themselves as policy takers, but the most effective ones were also policy influencer. And they really were actively involved in trying to use their research or their projects to influence and affect the policy environment because that's such a scaling multiplier. Before it's not a straight line. It's never. And it's important for donors as well as implementers. Really understand how to build some greater level of after management into the things that they do that. That's much more easily said than done, particularly on the donor side. But I think there's a growing recognition of the importance of doing it. Finally, just a couple of things that I would address to you directly. And to your colleagues in the CG system. unrealistic and perhaps even unproductive to expect researchers people whose heart is really in the research side of this to take responsibility for the entirety of the scaling problem. I don't think it's unrealists broaden the concept of research to an additional range of things that affect the obstacles to scale and the factors that are involved in scale. Those things need research just like just like the the agronomic and the and the veterinary things need research and they ought to be part and parcel of what we're doing in our research protocols. Secondly, even though we tend to put things together as packages, it turns out when they scale the packages often come apart or people bundle them in different ways. So the more our research calls either call for or allow for a greater level of bundling and unbundling of the components of our interventions. I could give examples if you the more likely they are to be successful in this along the scaling dynamic. We tend in our programs to think about attribution and attribution tends to mean direct beneficiaries direct directs directly serve people. But scaling terribly needs to look at things like like indirect beneficiaries and sustained adoption and things that go on beyond and outside our direct engagement. So trying to find better research methodologies for making it what I would call possible association between work and those larger metrics is important without looking like we're being self serving. It's not that we're taking credit for those things. It's that we're trying to make it clear to interested how the work we did at least influenced or had some role producing that. Fourth is a baton passes between controlled and unfold or between research and application settings in our observation are not nearly as good as they should be. Neither is is the commercial platform as it should be. And finally, I've made once before but that I feel very strongly about I think that the the CG centers and the don't themselves could do much more to help strengthen these intermediation functions feature of scaling very much depends on someone doing that. Those of you just point away to just to first is this source book and where it can be downloaded if you don't already have it. And the second is for those of you interested in staying involved with this, the community and alluded to at the beginning is quite active. It does have a working group on an agriculture and rural development currently chaired by someone from summit. And if any of you are in becoming part of that. Me an email, I'll be glad to make sure that happens. Thanks for. That was great. And I'm a great introduction to our to our workshop. Let me throw it open to questions and comments. Let me begin here in Nairobi. One of the questions in Nairobi and then I'll check if there's any comments or questions online. Who would like to kick off Jimmy. Introduce yourself. I'm Jimmy Smith. I really enjoyed your presentation. That's very good. My question is about targeting. What does your source books say about targeting the thing we're trying to scale is usually quite and therefore addresses the need of a specific typology of farmer for example, landscape, we have a very heterogeneous set of farmers. Yes. So we likely introducing our product. To the farmers who can scale it, but also many of them who can't scale it. So the shotgun approach. How do we get around that problem and be more target. Do you say anything in your source book about targeting. Not much to be perfectly honest, but I'll say just a little bit about it now if I could. The specification of target audience, I think is. People are very imprecise about it's what I was calling the denominator for in that very first graphic that I. The orange bars on the left and the big white to the right. Before you even draw that you have to say, well, what's the total audience that we're aspiring to on this and some intervention. Really specified and you say that audience is a very large audience, but you know in your heart of hearts that 70% of that audience is not going to adopt the technology. That you're pervading on this thing, you'd be much better off to draw the things smaller in the 1st place and say that we're really reaching 50% of the addressable market for this. And if somebody says yes, but we want you to work on a different technology that works for that would address a number of people will take that as a separate question. Just in mixing the 2. If you basically say, well, the audience is that large group. Our technology, you know, in our heart of hearts is not really designed for that large. You're going to keep finding yourself with the little orange bars on the left hand side, which is a very. Place to be that the thing that I. I think is unfortunate now. Is that people are being pushed because I feel about this, but it's telling conversation that's pushed to claim very large numbers as their goals. Very large numbers, but often it's a false claim. There's no chance that the things that they're doing currently are going to reach those kinds of numbers. Better, I think to have a discussion about what we're working on and how many people it has plausible reach to and have that conversation earlier rather than why we're not reaching the large. You think we're supposed to. Yeah. Hi Larry, this is vision and I'm based here in Nairobi as well. I co-lead the animal and human health program. I'm working on the biosciences side. So, one of the things that struck me in terms of some of the definitions or descriptions that you were using was when we've been thinking now about developing animal health products and specifically vaccines for example. We're being encouraged to differentiate between the terms proof of principle and proof of concept. The proof of principle is basically showing that the design that you have made works, but it doesn't necessarily take you to proof of concept. Proof of concept is taking that proof of principle beyond that stage to show that it is at a stage where you can now get sale private sector company or somebody else to be interested in taking it into further large scale piloting. I'm wondering whether the similar kind of logic could be used in some of the things you were describing. I saw that your first tier was proof of concept and wondered whether if that that should be proof of principle, rather than proof of concept and proof of concept then gets to the stage where your pilots don't fail. Yeah, because one of the things Ian you were saying was that pilots are failing. It's because there's the distinction between the two. And for example, and what we've been writing more recently and the discussions we've been having with private sector they're making a very clear distinction between those two. And unfortunately those two terms still get used synonymously. It's true and I think I'm guilty about myself. The, let me speak to one piece of that. That's the distinction between proof of principle and proof of concept is not one that I've, I've been using I'm going to think about that following our conversation now. But, but I will say one thing that I think is directly relevant to that there's a discussion that happens in a in a lot of development space about potential versus versus I'm sorry about capacity or potential versus performance. And, and for me, it's a it's not quite the same but it's similar to what you're you're alluded to let me see if I can draw this out for you. So you might. He got the capacity to run a six minute mile. And I would say well, is he ever run one. You would say well, you know, he's got the following lung capacity and he's got his body weight masses is such and such and. And the, and so he's got everything he needs to run the six minute. I feel much better about this if I'd seen him run a six minute mile. So, for me, the analogy on that is the net between capacity or potential or scalability. In actual scale and or performance on this so for what you were calling proof of concept. The closer it gets later the happier I would be, you know, it was not just showing you could work, but then it does work. And so, every time I was early career I have a little bit of hesitation and saying this in some quarters early in my career as one of the people who developed into log frame. And the, and the log frame had a very clear notion. Instituted evidence and predictive evidence was very different from descriptive evidence. So, to me, on the issue of scale, I'm not very persuaded by a proof of concept that still remains a kind of a in the potential arena. And so if proof of concept goes all the way to we've shown that there are, for example, seed companies bidding for it. Or that there are large swaths of territory where people are doing to me is great. Then I would say that a level that's materially different than proof of concept proof of principle. But if it more like we did here and we did an extended pilot there. I would say it's calling second stage pilots. It is not yet what I'm calling this other category of evidence. Okay, can I check if anyone online wants to come in. I just above on the speaking. Yes, need, please go ahead. Okay. Thank you very much Larry for the very elucidating presentation. So a few things were running through my mind as you spoke by the way I've read the scale up. Program in Hillary. One of the things that I'm struggling to resolve bothers on some of the issues you've raised those intervening factors you mentioned that can. Project to scale there is a factor. There is a cost element that is done up. So I think there's need as you rightly observed that. Be a good systematic on stand across. The landscape for donors for implementations. Even for the CG system. To be able to handle this business of scaling. We find ourselves in situation where donors are telling us we've had enough research. Any more research. On and bring these technologies, but in order to be able to do that second phase of piloting that you mentioned some element of research is involved because what works in one situation may not necessarily work in another situation. So how do we reconcile all these seeming. Fronting issues when handling the subject of. My second question is one of tools tools to assess scalability of our technology. You did mention that there are now tools how widely applicable are these tools. Especially for people like me who are interested in scaling and for others as well. Thank you. Okay, a little bit to each of those questions I think the. I'll begin in a kind of an abstract level. Try to get practical really quickly. I think. My mother used to say, you can't ask someone to be a put tall. It's it's not realistic to ask people or institutions to do things that they weren't designed to do. I don't think that research institutions are the best institutions, scaling technologies don't. And I think if they put too much pressure on you to do that. You'll be unhappy and they'll be unhappy. What I think is. Appropriate and useful is for you to take a significant step in that direction because I think. A bad point of contact between the. Responsible for scaling or the people who are responsible for delivery at scale. And the people who would search and I think. Each of you would take a step toward the middle. It would work a lot better. For me. That translates to in practice is trying to understand as much as you can, what's going to be involved in getting the technology to. And I say. That involves a lot of experimentation along the way, not just before you roll out throughout the process. I think there's a research dimension, some of its social science research, some of its agricultural. And veterinary research, but there's a lot of ongoing research throughout the process, but trying to figure out where you as an institution relate to that I think is a case by case. But it's dependent. Very solid crosswalk. Between yourselves and the people who are responsible for the actual deliverance. If there's no intermediary making the marriage between the two. It basically mean you and the deliverer need to be in close communication because you're there's no marriage broker. You have to simply engage with each other directly. In that, in a way that is to take the research, but even more importantly, the roll out of the search and continue and continue during the scaling process to keep an active research and learning dimension as part of that. And I think all of those are areas where the system and you guys in particular. Could tremendous service to the process of application, but you need to see it. Not as trying to get people to adopt your innovation, but as trying to help them improve their practice using your innovations. And that means being willing to let go of components and make compromise with elements and trade off things, all the things that are a little bit uncomfortable. If you're coming at it from a more narrowly research perspective. Switch now to the issue of tools to the, the tool that I know best is we were hired for about, I think about 3 years. US AIDS Bureau of security to 1st do scaling assessments of 5 cases. We looked at that irrigated rice and Senegal at hybrid maze in Zambia. At corp in Uganda. At picks bags in Kenya. And it to will tractors in Bangladesh. Look at how they had scaled or not and to understand the scaling dynamic as much as possible. Those 5 cases and a kind of a summer. On them are written up. I think quite thoroughly. Then we were asked kind of on the basis of that to take a. What we call the scalability assessment checklist. We'd use other sectors and adapt it for use in the agricultural sector. So we did that and it was applied. I maybe a half a dozen might by now, maybe 10 or 12. Different technologies. It's not. Simple. It's not hard, but it's not that simple. I wanted it to be easier. But enough things got added in a little bit complicated. The cases I know were. Generate some genuine insights, but the ones I know best the applications I know have involved someone from our team who was sort of coaching the people who are doing the skip who are applying the tool. I think it doesn't need that. And if if you decide to use it, I'd be very interested in your feedback as to how user friendly it is and whether there are other things that that might be done. There are much tools out there that are. I think they're less, they go less deep. Leonard who you'll be talking to later. I helped to develop one of those. I'm sure you should ask him about it. It's got 10 dimensions that are involved in scaling and it quite straightforward to apply. Sure, it does not need any extra help. Then there was 1 that I think mark and his colleagues were involved. I think feeling readiness. I know less about about that 1, but when you talk to the 2 of them, you should ask them. To to give you some details on them as for the 1 that I know best, which is the most complicated, but also the end up wanting to do anything more with that. Let me know and I'll hook you up with people who know more about it. Thanks, anyone else online wants to come in. And this is Helen. Yes, please go ahead. My name is Helen. I work on the CGI research program on livestock. And I was in what you said about making allowance for bundling and unbundling of packages. Because in our project, in our ground currently, we're trying to do some work in a few countries to show impact of putting different technologies together in order to support livestock. So the siloing of research and showing, you know, feed intervention together with genetics and together with other technical interventions can work in a particular livestock value chain. So I'm just interested in, you know, in hearing a little bit more your thoughts on on how we can go about this because we have a. We want what we're doing to be available and to be able to show impact, but we also to be able to get some that we can report about how this works in different communities and working it. For better or for worse, I have a strongly held view about this. I think that for those who are on the development provider side. We tend to think in packages because we put together a complex of factors that together produce an outcome and that's the way we approach things. But when things go to scale, there's simply no way to enforce that cohesion and people will pick and choose as they. Now, maybe the logic of the package will be strong enough that people will take it intact and apply it the way you hope they will. But I've seen many more cases where that's not what they do, where they find some piece or the other and they, they either hide it off that way anyway, or it doesn't. And they sometimes do it in ways you would never predicted. There's 1 case that has been reported to me. I don't know this personally and it could be apocryphal and 1 of you may know the true story. About this, but I'll tell you as it was told to me, because even if it's not exactly true, I think it makes a point. It has to do with the roll out of the heat tolerant maze. And apparently, and at least setting after the research had proven the viability of tried to roll it out. What farmers said was, look, we had droughts 1 year and 6. 1 year and 6 when we have a drought, this seed would really save our crop over 5 years and 6 we don't have droughts. And in those 5 years, it's an unnecessary insurance policy. We're poor and we can't afford unnecessary insurance. So no, thank you. We'll take. There's a 5 and 6 chance that we won't have any trouble this year. So it didn't scale. Someone said, okay, maybe we can. The same seed disease tolerance. And it turned out that basically saved you. Someone would 3 years out of 6. That was enough. According to the person who told me the story to persuade the farmers into a different problem. The problem. Trying to get the seed to the. Because it was basically made is not a high value crop and it's heavy. And we were talking about last mile kinds of farmers. And they only saw that when they packaged it or when it got bundled together with veterinary. Particularly vaccines because those were value low volume. We're going to many of the same places and it didn't hurt to necessarily take a long commodity when they were when they were going. And only at that point did the curve begin to go vertical. Maybe that's correct. Maybe it's not. But the point being only that it took a lot of. Tinkering before you found a formula that worked. And at the same way, I think that's what happens at the house. The farmer level that they basically. Giving them a package, but you have to be prepared for the fact that they're going to, they're like. Pick and choose. You probably won't even know how they're going to pick and choose so you watch and see what they do. So I think I'm arguing for trying to be more sort of. Having less of a catechism. About what the end result is going to look like. More willingness when we have things. Continue following the chain for a while. As for the scale up and recognize that. Involve something we think is less. What we originally had in mind. I give you. One more example. This is a slightly. I'll tell you the short version of what could be a story, but the. Extremely affect. Neonatal survival set of interventions. Cut child mortality or neonatal mortality by 50% in India. We're helping it scale to national. The only way to scale it was to. Two of the central components of this for different reasons. This was not a decision that was made by the household. These were different constraints. We ran into about trying to scale the intervention. So we took what had been very well proven. Been written up seven times in Lancet in randomized control trials. And with the, with the agreement of the developers stripped out. Two of its essential components. With at least some. It's not as much evidence that it would work pretty well without. Those two components. And in fact. They got about 60% of the. And it's now being. Delivered by 740,000 rural health workers in India. But if you'd ask the people who developed the package. They would have. Originally. Horrified at the prospect of pulling out those two components. Okay. Yeah. Yeah. Sorry. Who's that? Oh, okay. Thanks. Thanks, Helen. Right. Let me ask for anyone else online before I come back here to. In Nairobi. Anyone else online want to intervene? No. Okay. Let me come back to the room here. Oh, sorry. Michael. Yes. So, hi. I'm Michael. I'm the chief operating officer. Here at Hillary. It sounds like a little bit of what you're talking about just now was this concept that they use in, in kind of tech startups like lean startups where. You constantly try to get products out to. To the customers as soon as possible without making them the perfect, perfect product. So that you can actually test it. See what the reaction is. Take it back. Kind of re rethink about it and then get it back out to the beneficiaries in our case. Again. It isn't, it isn't my. I mean, in some things, it is that in some kinds of technologies, I think is very close to the lean startup thing, but, but there are certain cases where it's really not that where it, but. That it shares with the lean concept notion is the need to recognize the probable, the probability of. Modificational. And two things modify the scaling strategy. And the second is the intervention itself. Both of those things end up being modified in the process of implementation. A lot of things, in my opinion, don't lend themselves. To the sort of minimum viable prototype and rolling out. They just don't, they aren't that kind of. So some really do need. Exactly the kind of basic research fundamental pilot. Project. First, it's just that even when you're at that stage and you think you've got something that works. The need for. Continue. Program of engagement as you roll out. And a willingness to modify. I think is a characteristic that it shares with the lean startup stuff. Okay. Any final comments or questions before we wrap up this, this session. And. Yes. It's me again. Okay. This is probably an, an odd question that I'm posting to Larry. I don't know. Because it's probably not been done before and probably for different technologies. It would differ. In terms of quantification. If you were to quantify. The different components. Of a scaling process. The technology itself. The. Enable. Enable. Enable. Enable. Enable. Enable itself. Enabling environments. Private sector. And the delivery system. what would you put. On this different components. I said, I know it's probably an absorption. Just. Something to also get those things. The kind of emphasis we need to place. on these things when we are attempting to scale. Yeah. Well, the premise of you, namely it's likely to be variable is right, but at risk of being provocative, I'll say I would put about 10% on the technology. And the reason I say that is that if you look at the products that have scale almost without exception, there was something better up there. They had the formula for trying to move the thing through the system when Cargill or Syngenta decided to move something to scale. It may be the best thing, but even if it weren't, they have so many things that are right for it. There's so many things they've judged correctly about the environment of the committee for doing this that their chances of success go up by an order of magnitude. I think I wouldn't draw the distinction in your case, though, entirely between the product and the strategy for reaching scale. Because I think, as I said earlier, I think the research continues into the strategy stuff, not just the product. And there are a ton of research issues. Too often they're done without the evidence they ought to have to move forward. So if I could wave a magic wand, I would extend the handholding between the institutions and the implementers for a much longer period than often it continues. So you're kind of their evidence buddies as this thing begins to move out. And then it becomes more difficult to answer your question concretely because the evidence is kind of an ongoing function. But if what we mean is the initial analogy, I'd say 10% as a row file. Okay, on that note, let's draw this to a close. So Larry, thanks very much for getting our workshop off to a great start, as I knew you would. I think you've provided us with lots to think about, a great platform to move forward in the next couple of days in particular, but also given us a lot of things to think about as we evolve as an institute, as a center. And I hope we can continue to engage with you this workshop is a sort of start of a process that we would like to continue. So I hope we can continue to engage with you in the coming months and years as part of moving the community forward to be much more effective in terms of achieving the impact that we want to achieve. Well, I would like that very much. And I would say in return myself as somebody just trying to push this idea of a scale perspective in this community of practice that I've met before in the agriculture working group are very much committed to doing anything we can to support this kind of thing. It's a very, the community of practice is very loosely formed in the sense that it does whatever it wants to do with whomever wants to do it. But I know that there are at least a number of us who are very committed to advance this. And if there's anything we can do to help you we're glad to do that.