Avoiding idea traps

Experience design lessons in AgTech

Thousands of pounds, and easily more than twice my height, I’m climbing up the ladder and into the cab of this huge modern John Deere tractor. This equipment was not a natural fit for our prototype, but we were so confident in what was made we didn’t see the harm in testing this in a new tractor.?

After we connected our prototype to the windshield, we were off to the field for our first test.

?I was keeping my professionalism while trying not to grin too broadly.

Imagine a small boy getting a toy tractor for Christmas.?

Now make that toy thousands of times larger. It's summertime. And that little kid is now sitting beside a German farmer, who has just let him start off-roading into a patch of field.?

Five minutes into the test and my excitement deflated. The good news, the core idea worked! The bad news, what we were building wasn't going to be successful. The mistakes were bad, and blindingly obvious. Yay, they let me drive a tractor but shit. Imagine being shown fine print, fine print on a screen outside in the sun. Then asked to read it while off-roading extremely heavy machinery. Yes, it was that bad.?

Driving into a new zone in the field our prototype would flash and beep telling the farmer to manually increase or decrease the flow of fertiliser. The most important details were far too small to read on a bright day, we communicated information that wasn’t helpful, and the beeping was almost inaudible.

There was no way a farmer would trust what we had just made.?

How did we get here?

Flashback: Two weeks earlier. I leaned back into the office chair, pulling my feet up onto the seat. Hendrik pushed me around the office, while I called out commands with my eyes fixed on a screen in my hands. “Slower, slower,” “Okay now faster!”. What looked like two 28 year old adults playing with office furniture — I assure you, was professionals at work.??

We were testing the motion tracking of a prototype, and it was working!? We were feeling pretty good about ourselves.? First to understand this excitement you'll need to have some farming context. Picture your kitchen table strewn with cups of various sizes. As you puncture a large bag of rice, it trickles out steadily. You systematically move to coating every inch of the table at a consistent pace. After this meticulous process, you'll observe some cups brimming over while others remain partially empty. This mirrors the application of fertilisers: Some plants need more, some less.? Similarly, applying a uniform amount of fertiliser across the field can lead to economic and environmental issues for the farm.?

To tackle these issues, systems were developed to work out exactly where in the field a crop needed more or less nutrients.

Using methods like soil sampling, the reflectance of light in satellite imagery, and some clever calculations. Fertiliser needs could be worked out for each part of the field.1 Then these systems could generate files to control spreading machinery, automatically adjusting the flow of fertiliser as they moved through the field. A little less here, more there, voilà. ? Most of the farmers we remotely interviewed didn’t have access to the latest cutting edge equipment, and were reluctant to buy expensive spreaders when still sceptical. Here is why we were excited. Our prototype would allow farmers to “manually” mimic the benefits of modern variable rate equipment, while still using old fertiliser spreaders. All they would need is a smartphone.??

I arrived on that farm confident that we had something special. After all, we did the work!? Our previous remote interviews with farmers taught us a lot. We followed a structured double diamond design process of understanding & diagnose, concept, solution, and delivery. To understand farmer behaviour we co created this with internal agronomists(crop specialist). We knew the theory and science was sound. The concept was rooted in science from a Research and Development team.

Finally, we'd interviewed different farmers, albeit remotely, and had ended up with a viable solution to a real problem.?

I didn’t see it then but we had fallen into an “Idea trap". We found an insight and put on our blinders. We had begun refining the work, fixing problems in the smallest detail,? without actually putting in its real environment. Ten minutes in it’s actual test, would have told us a very different story than what I had in my head.?

When we are designing something new we can get too caught up in our own ideas, or zoom into small details far too early.? It happens, we get excited, believe our own hype and dive into a rabbit hole. All the while we refine, adapt, and edit endlessly, touting our process as lean and iterative. “We are being lean!”. “We are iterative!”. Ugh. Let me paint another picture. Perhaps you can relate. You're in a meeting or workshop, working with colleagues in a closed room. Time ticks by, ideas flow when one of you then leaves the room for a couple of? minutes. When they come back, they are hit with it, a pungent odour of people cooped up.? The human aroma that says “Oh wow, please open a window!”.

Something similar can happen in Experience design.?

If we're closing ourselves in. Feeding our work with our own internal bubble, the work can veer off course. Growing and developing away from something actually useful to our customers. If this goes on too long, what started as a fresh idea, moulds and starts to grow on its own.

We need to open up the metaphorical doors and windows, airing out our work.? The sunlight of customer feedback helps disinfect bad ideas. Real customer input, helps ensure the work addresses the actual needs of who your building for.

We had gotten caught up in our own bubble of feedback for too long. Testing in simulations rather than evaluating ideas with the people we were meant to be helping.?

Before being in the tractor, I was adamant I knew where the issues might be. It was clear to me where we could improve.? Afterwards, I felt foolish, only wishing we visited this farm sooner.

I am a huge advocate for remote interviews, and remote testing. It’s a perfect tool in many cases. Though it can be taken too far.

Since the Pandemic it’s even easier to do, in many cases remote testing and relying on feedback from remote testing services have become the norm.?

This can give you false confidence. A simulation is not the same thing as evaluating ideas in context.?

Takeaways

When you kick off your next project, set artificial time constraints. Try to book a number of customer interviews, and stakeholder check-ins from day one. These checkpoints with customers will help prevent you from going too far off course.

• Schedule tests and check ins from day one?
• Accept the artificial constraints. You can alway improve it later.?

Bring in customers before you feel comfortable, don’t wait until you feel it's ready.? It is almost never too early to get input, even if you’re still outlining the work, once you have something tangible, test it.?

• Test your thinking before you feel comfortable. Can you do this on day one?
• Remember that often the best improvements come from conversation with real customers.?
• Accept you’re going to get it wrong, you can always eat candy on your way back from the test.

Resist the convenient allure of remote testing, and all of its time saving benefits. Evaluate your solutions in the real environment, and it should be encouraged to leave your comfort zone. That’s where new things are learned.?

• Run at least one person test, not only simulations.?
• Accept that in-person tests will take more time. Getting it wrong will cost you even more in the long run.
• Exhale, leave the office, and put yourself in your customers' context. You’ll remember that more anyways.?

Lastly, Just be sure you are making something people actually want. If not you may find yourself like me that day.? Alone, standing in a field thinking? “How am I going to fix this?”.

A very big thank you to everyone in Write of Passage Cohort 12 who helped spare with me to work out this essay.

if you are interest in more writing please let me know, consider following for future work.

1 Pawase, Pranav & Walunj, Avdhoot. (2023). Variable rate fertilizer application technology for nutrient management: A review. International Journal of Agricultural and Biological Engineering. 16. 11-19.?