11 – The Options by Context Principle

[00:00:00] Lucie: Hi and welcome to the IDEMS Principle. I’m Lucie Hazelgrove Planel, a social impact scientist with IDEMS, and I’m here with David Stern, one of the founding directors.

Hi David.

[00:00:16] David: Hi Lucie. I’m looking forward to today’s discussion.

[00:00:19] Lucie: I thought it’d be really interesting to discuss the principle of Options by Context. So this is one of the bigger principles, which has three smaller ones embedded within it.

So I think a lot of my work, I think a lot of all of our work fits into Options by Context, but I know my work fits in because of the history of Options by Context. Should we start with history? Should we talk about, first of all, what Options by Context is actually, what does this actually mean in a very practical sense?

[00:00:46] David: The options by context principle has, as you said, been elevated because it contains then three other principles, which are Open by Default, Local Innovation and Continually Evolving. The wording that we have for the Option by Context principle is the following. This principle is part of core company thinking. It requires context to be considered when proposing options or interventions, in particular it precludes the search for single or blanket solutions.

And this is really important. This is central to our way of thinking. It affects how we do science. It’s come out of different areas and it actually, it’s become so important in my daily thinking, I don’t even notice when I’m using it anymore. It is definitely not how most researchers think. It differentiates us from many researchers in terms of how they think and also just how the world seems to work at the moment.

You’ve come across this, as you said, this comes out of our agroecology work. And so you’ve come across this in the context of the agroecology work. Is that right?

[00:02:04] Lucie: Exactly. I was really interested to hear that years ago, agricultural research tried to find one solution for many. It tried to find the ideal solution which just seems obvious that it wouldn’t. It just seems obvious that lots of people, that ideal solution would not be ideal for them. And you can tell, say more about how the options by context approach came out.

[00:02:28] David: This is a colleague of ours, Ric Coe, basically took what happens in breeding trials, where they moved from actually looking at breeding trials for these single source to be able to move towards breeding trials which they called multi environment trials which had interactions, I could delve into that, of to where, what this origin was. Because in breeding it was then seen that actually we need to consider different environments and which genotypes work in different environments, which sort of varieties are more productive in different environments and different conditions, different climatological conditions, different soil conditions and so on.

And so that element of breeding, crop breeding, then was recognised as being something which came into the research much more generally and was now a standard part of the research process when you consider new varieties.

But the observation that that doesn’t just apply to breeding, but that applies to management practice, and all these other things, and actually is a sensible, better way to do research. Ric Coe should be given credit for actually drawing this out. To my knowledge, this has its origins in Ric Coe’s thinking, as a research method support specialist.

And he brought that out, and made it explicit. And since then, it’s been adopted in so many different areas. There’s a group at Reading University have made this central to their, to PICSA, the Participatory Integrated Climate Services for Agriculture. And then, we use it everywhere. We discuss it in our thinking, in research, in so many different ways, we bring it into our work where it’s how we build technology.

We build technology to be adapted to different contexts in certain ways. That’s part of what differentiates us in terms of our thinking about how we build technology. It comes into everything we do now. Actually thinking about the fact that this contextual adaptation needs to be baked into research processes, it needs to be baked in to products as they get designed, and all aspects of our life.

Now, implicitly, people had been doing that. But I think this explicit formulation of this as a principle is surprisingly rare, as I see it. And it’s something which we find central, it’s one of our core principles. It captures a lot of other things. We mentioned Open by Default, and we can dig into that in another podcast.

But it’s, again, it’s this fact that there was some context having a default and then considering the context to see when the default is not appropriate and when you should use something else. That’s part of it. Instead of saying that everything should be open, there are going to be contexts where you should be open, there are going to be contexts where you shouldn’t be open, and so you should be thinking about what your default is. So the important thing there is by default.

Local Innovation. When you’re actually going in and trying to empower people locally to do things, you’re wanting to give them the power to make things and contextualize them to their context. This is something which is supported. But you’re wanting to do that as part of a broader framework. How do you empower that local innovation within a bigger framework of something which is global? It’s a really interesting and difficult question. And our local innovation is really about enabling and empowering that.

And this idea of Continually Evolving, again, contexts change, but more than that, if you actually build evolution in to what you’re doing, then you enable different contexts to adapt differently and then to potentially evolve in different ways over time. You’re baking into what you’re creating this fact that over time, your options can differentiate themselves more in their different contexts based on how they serve the needs of that context.

And so all of these things are tied into this fundamental principle of Options by Context. We need to think about the context within which things are embedded as part of what we’re thinking of in terms of interventions.

[00:06:55] Lucie: Exactly. So can we start off with an example then, perhaps, let’s say if we have a project, I’ll let you choose a product if you want. Can you explain how you allow that product to be, to fit in with the principle of Options by Context?

[00:07:10] David: It’s difficult to understand exactly what the best way to think about this because it’s so encompassing in different sort of contexts.

[00:07:19] Lucie: And I’m aware that I started at the product phase whereas really it should start before you’ve got the product.

[00:07:25] David: Yes and no. Maybe let’s think about this in terms of when we developed the app for the parenting. So we have a parent app. This is a big part of our work with Oxford University on the Parenting for Lifelong Health. We developed an app for them, for a specific context. But while we were developing it for that context, we were very consciously thinking about how this would be adapted to other contexts. And so we made compromises in terms of the efficiency and the nature of how we developed that app for that specific context, so that when it goes to another context, it would be able to be adapted more easily and more effectively. And that, thinking about those trade offs, thinking about how in our development processes, we took decisions which enabled us to prioritise reducing the cost of adaptation of future versions to future contexts over getting the version, the best possible version for the context we were working in.

[00:08:35] Lucie: Yes, Options by Context has a cost. It has benefits but it also has a cost.

[00:08:41] David: Absolutely. And understanding how to manage those trade offs. And I think this is important that it is harder to do and to think in terms of Options by Context, but it is much more rewarding. Maybe I’ll take another example of PICSA where the cost is not there and the benefit is much clearer. In PICSA, they establish options by context as an integral part of how they give out the advice to farmers. So PICSA, let me remind you, is Participatory Integrated Climate Services for Agriculture.

And so the point is that this is helping extension agents, people who work with farmers, to integrate climate information. And one of the principles that they have is Options by Context. And one of the things that means is in their training of the extension agents, they do not offer advice on what a farmer should do.

[00:09:46] Lucie: Yep.

[00:09:47] David: That is very explicit in how they train the extension agent to do this. They do offer information about different crops, the climate in their context, about different options. They recognize that the farmers who are receiving this are going to need to make their own decisions and each of them will have their own context.

Historically, extension agents were often trying to tell farmers what they should do. Whereas, in this particular context, the Options by Context principle manifests itself as changing that relationship so that the extension agents become a source of information, not a source of advice.

[00:10:36] Lucie: An enabler.

[00:10:37] David: And the point there is that in some sense there are advantages and disadvantages of this. And it’s not all context where that’s what they’re, what people are wanting to do, but it has shown to be extremely effective. Because what’s happened is different people have taken that advice and received it differently and used it in many different ways.

And they’ve got something like a 90 percent success rate in terms of people using that information to change their decision making in ways that they perceive as having positive impact on their farming practices. And that’s been consistently, they’ve been in 20 something countries, and most of the evaluations I’ve seen have been well over 90 percent success rates.

But a lot of it, I think, does come down to it, and they attribute a lot of this to this Options by Context approach where they are trying to help the extension agents to move from this advice role to this information provider role, which enables the decision making to lie with the farmers much more.

[00:11:41] Lucie: Could I just quickly say too that I’ve seen some of PICSA’s evaluation methods so saying that’s a 90 percent success rate at everything, you were very aware of what words you were using there. And I think it’s important to say it’s not just a sort of, tick box that people were saying yes, they were happy with the PICSA approach. The researchers really evaluated the project with getting locally self evaluated things as well, bringing that in as well.

[00:12:06] David: And there were many layers, you’ve seen the case studies, you’ve seen the sort of the broader components. This has been a really quite serious evaluation process that they’ve devised. And when you talk about this over 90 percent success rate, it’s it really is astounding in this that, those 90 percent actually say these are the changes I made and this is what…

[00:12:24] Lucie: Exactly it’s backed up.

[00:12:26] David: You know it is a very sound approach that they’ve built to actually evaluating this but the evaluation has been really positive.

Again, to come back, this is an illustration of Options by Context. Options by Context, I use them as an example of Options by Context because how it plays out in their context is really this change of role for the extension agent and actually thinking about how that role changes. And just like when we think about how we create the technology, our role as technology creators is to think beyond the individual task we have now, to the broader task we don’t yet have, but which we know will be needed in the future.

[00:13:16] Lucie: Which is where the Open by Default comes in, as you were saying earlier.

[00:13:20] David: Exactly. This relates to Open by Default and thinking about that and how it might not be us doing that. If it’s open, then other people could take that on. We can’t be precious about being the ones to do that contextualization. So we have to enable and allow others to take it on and be the ones to drive it and develop it.

[00:13:39] Lucie: And a lot of Options by Context seems to be also then like you just said about allowing other people to do what they want with something. Like you said with the PICSA example, the extension agents are providers of information and then it’s the farmers who make the decision, as opposed to before it was farmers receiving information and just perhaps deciding whether or not to follow what they were told.

[00:14:01] David: Yeah, exactly. Farmers receiving advice rather than information. Then the extension agents often are trained in saying this is how you should grow your crops. And what’s been found is there’s all sorts of reasons why that is maybe true on average, but no individual farmer is your average farmer. And so that might not be true for any individual. So that’s part of this as well, it’s this recognising variability. That is central to understanding.

[00:14:30] Lucie: Can I ask also other examples, like perhaps using the PICSA approach, other examples where people have too many options. I know if I go to a supermarket or something, there’s too many options.

[00:14:40] David: There’s actually some very interesting research around this, and forget about Options by Context. There is research which shows that being given options, being given choice decreases happiness. So don’t get me wrong. Again, as any good principle, the opposite is maybe good as well. So Options by Context is not blanket better. This is where we take this approach and we believe in this approach as a guiding principle for how we develop things. But that doesn’t mean that it’s definitely better. There are many contexts within which not following Options by Context could be desirable.

And I love this piece of research which actually looked at people’s satisfaction with a piece of art. They were given a, I’ve got to make sure I remember this right.

[00:15:27] Lucie: I’ve not heard of this example, yeah.

[00:15:29] David: It was beautiful research. Basically, the researchers wanted to understand whether choice increased or decreased happiness. And there was a process which happened, and some people got just given a piece of art and other people had choice related to the process of the piece of art that they got. What they found was that on average people who were just given it who had less choice were happier with the art six months later, I think it was, than those who had choice, who had buyer’s regret.

Choice is not always a good thing, exactly as you put it. And having more options doesn’t mean that you’re necessarily going to be happier, going to make better decisions. There is an element, though of, therefore, balance between these two things. And this is where I was careful to explain that actually our approach to being able to think about creating Options by Context when we’re developing the technology comes with a cost.

We try to minimize that cost in certain ways, but there is a cost associated to that. And I think we could potentially develop better one off technologies with the same resource package if we weren’t taking this approach. But it wouldn’t be in our opinion, meeting the needs that we’re trying to serve as well.

And one of the things I think is central is that actually, in our needs, we are looking at prioritising impact. And in prioritising impact, each marginalized context is different from one another, and so you’re going to have to go through this process. In commercial contexts, where you’re looking at maximizing market, then this may not be the right approach because you get efficiencies of scale.

If you can get a large segment of the population that want something, then the fact that you’re excluding people who need something else, you still get efficiencies of scale by targeting them. So there’s good reasons why some people would want to take other approaches. You can maximize and optimize different things.

[00:17:44] Lucie: Yeah, I like that phrase, efficiencies of scale. I’m sure it’s well known, but I hadn’t heard of it before.

[00:17:51] David: It’s a really important one. Almost by definition, you, by having something where you’re offering multiple contexts, you lose some of that efficiency of scale. And so there’s always these trade off analyses.

A good principle, in general, is not good for everyone. It is a genuine decision. It’s a position you’re putting yourself on, where the alternative is also good. Yeah? But it’s good at serving other things. And you’re saying that, okay, I’m making this choice to be serving this side of the equation rather than that side of the equation.

If there isn’t a genuine choice, then you don’t have a good principle. Options by Context is a, I think it’s central to who we are, it’s central to what we do, and it’s made more powerful by the fact that it shouldn’t be central to everyone. Other people should make different choices for different reasons.

If we were a purely commercial company, I don’t think Options by Context would be a good principle for us because I think we’d lose efficiencies of scale, the cost of it, in terms of maximizing profit, this would be a very difficult principle to live by and it would be in contradiction to certain things.

So I think there are elements of this where we should always be having our principles, which are not obvious. And this is why they are guiding principles. We can do things which are in contradiction to our principles, as long as we understand why and we think about why we’re not following this guiding principle, in this particular case.

Options by Context. This is a, this is part of a principled approach! Principles are not rules. There should be cases where actually we follow a different approach. So it’s a really interesting and difficult one.

Options by Context is central to who we are. It is not obvious for everyone. In research, I would argue this is not the norm at this point in time. Most statistical methods aren’t about Options by Context, it’s all about finding the best. No trying to find out differences which determine that you are, this is the best or whatever it is.

And so actually, a lot of scientific practices are not aligned with Options by Context. That doesn’t mean that they’re wrong or that they’re bad. That just means that our approach, by and large, means that we are innovating on the research methods, on how to do research, because this isn’t central to all approaches.

[00:20:28] Lucie: Fascinating. Thank you very much, David. I’ve learned a lot about Options by Context and especially the interestingly, I’ve especially learned about what the converse of that is. So thank you.

[00:20:39] David: I’m really grateful for that, because to me, what’s really important is with all of these principles, the converse is at least as important. Thinking about that, thinking about what it would be not to be Options by Context, and why that could be good is at least as important to understanding why, because otherwise you don’t understand that there is a decision point. This is helping to guide decisions. If it’s only positive, it’s not a good principle because it doesn’t help to guide decisions.

Anyway, this is maybe a broader principles discussion, but this has been nice. Thank you.

[00:21:12] Lucie: Thanks.