Project Everest

Adopted Experiment

[Experiment Adopted]: FarmEd Timor - Extended Soil Sensor Case Studies - July 2018

Reuben Dowie
Reuben Dowie | Jul 23, 2018 | in FarmEd - Timor-Leste

The July FarmEd team has ideated an MVP: a Facebook page farmers can message for specific agricultural advice. A messaging platform will be able to simulate the chat bot which would be integrated into a potential FarmEd smartphone application. 

This service is to be paired with another, completely different MVP: farmers using soil sensors to easily determine important soil metrics such as pH and moisture levels. These will work well together, as farmers can use them to test their soil then get expert advice. Face to face interactions not being required for this process is key for eventual scalability.

We have decided not to launch these MVPs this month because we need to have a stronger assurance of their value to farmers. We do not want to get ahead of ourselves and spend the final week trying to push something that isn’t ready to go. Instead, we will establish extended case studies with selected farmer candidates. These will serve as thorough user experience surveys, running from July to December.

The case studies will involve loaning soil sensors to selected farmers, teaching them how to use them, then asking them to regularly test their soil in the months between July and December. They would message their soil data to our Facebook page ‘FarmEd Timor-Leste'.

When the December team steps on, they would re-approach case study participants and talk to them about the information learned from the sensors, and any changes the Farmers could make to make to improve the quality of their soil, specifically for the crops they are growing. This is the main incentive farmers have for participating in a case study and we have written simple MOUs outlining this agreement.

So, how exactly will the case studies run?
This week, we will travel to previously visited farms in Díli, Dare and Hera. Candidates for the study have been selected based on all the empathising we have done this month, which is summarised in a potential customer contacts spreadsheet. When we go to each farm, we will assess whether or not the farmer is genuinely interested in the case study by explaining what they would need to do, what we would do for them, and how we could both benefit from it (more on this later).
Farmer responsibilities:
- Use the sensors once per week to test their soil, for each crop they are growing.
- Record the data and message it to FarmEd Timor-Leste Facebook page each week.
- Do not cause any damage to the sensor and return it in December.
- Notify us when the battery goes flat as well as any other issues.
Our responsibilities:
- Provide farmers with soil sensors and spare batteries.
- Provide adequate written instructions on how to use the sensors to determine soil data.
- Ensure farmers are connected with the Facebook page
- Analyse the soil data that the farmers send to us.
- Provide the collated data to the farmers in December.
- Advise of any changes that farmers could make to improve the health of their soil.

These responsibilities are detailed in the MOU. If the farmers agree to these and sign the MOU, we will give them a sensor and instructions on how to use it, plus instructions on how to send data to our Facebook page. Care will be taken to ensure the farmers are certain on how to use them before leaving the farm. To be certain they can follow the instructions until December, translators will call the farmers daily until the project finishes and ask them to test the soil while on the phone. It is vital to get this part right so as to not waste time and resources. Before stepping off from July, we need to be 100% sure that the farmers understand what they need to do and know how to do it.

From the case studies, we hope to get all the sensor UX data we need for the MVP to be further developed in December. This will be gathered by posting monthly UX surveys on the Facebook page that farmers in the case studies can complete. More qualitative UX surveys would also be taken in December, by empathising with the farmers. The value farmers see in soil sensors can also be gauged from the case studies. It is quite possible some farmers will not continue to test their soil and send us the results until December, because they do not see the value in it. Or, if they follow the MOU faithfully, it will indicate that they see real value in the sensors.

Incentives for the farmers to participate and to continue in the case studies need to be carefully planned. If incentives are too strong, there will be a false value in them participating. For example, if we tried to get farmers on board by offering to pay them $1 per week they send us their soil data, they might be inclined to do so for money, not because they believe sensors can help them. But if we simply ask them to test their soil and send data without explaining how it may benifit, it is highly unlikely they would agree. So, we need to give realistic incentive that comes from the intrinsic value of agricultural sensors. This is it: for farmers who hold to their side of the MOU, the December team will analyse the soil data for that farm, and, if necessary, provide advice to the farmers on how to improve the health of their soil in a cost effective manner, that will boost the yield of their farm while supporting sustainability. This will most likely be simple advice, such as the addition of lime to raise soil pH.

We would love to know your thoughts on the value and methodology of these case studies, and will consider any insights you share!

edited on 6th September 2018, 01:09 by Justin Hakeem

Zoe Cahill Jul 25, 2018

Status labels added: Proposed Experiment, Experiment adopted

Reply 0

Zoe Cahill Jul 25, 2018

Hey Reuben,

I have just a few questions for you:
- What kind of information are you providing to farmers once they engage with the simulated chatbot service? How are you going to ensure that the information provided is of reasonable quality?
- How are you going to ensure that the UX data that is collected and posted by farmers is valid, quality controlled and quantifiable?
- What is the reason for collecting all of this data? Is it merely for testing how farmers interact with technology over an extended period of time, or are you trying to collect farm data as well that can then feed into the advice and information that is provided back to farmers? if it is the latter, how would this data assist in the overarching goal of FarmEd?
- What MVP are you developing further in December? Would you have enough of a validation to roll out an MVP of the actual app?
- At this stage, would FarmEd have the ability to do a comprehensive enough farm analysis in December, and thus is that something that can be guaranteed as an incentive for farmers to engage with this service?

Hope everything is going well, looking forward to hearing from you soon.


Reply 2

Reuben Dowie Jul 26, 2018

Thank you for your questions Zoe, they are very important questions. I will answer them in order.

We don't currently know exactly how the simulated chatbot should run and what kind of information it would provide. Earlier in the month, we were thinking of running case studies of the actual MVP and providing advice through the Facebook page. We scrapped this plan because we were unsure of the quality and impact of the information we would provide. Moving forward, we need to be cautious about what we provide. There needs to be more agricultural expertise available to the team before beginning the MVP.

To ensure that UX data is properly captured, we have created a survey that we will send to study participants once every month over the break (we will get this translated tomorrow!). From empathising this month, we learnt that it is cultural for Timorese to say yes without always fully meaning it, so we made sure to word the survey in order to encourage honest responses. A more detailed UX survey via direct empathy will also be conducted in December.

We are running the case studies to collect UX data and to feed into the advice that we give back to farmers. We set up five case studies, so that is 20 months of total user experience that can be analysed. Whether the farmers engage well over the 20 months or not, will still be valuable data on the viability of a sensor service and messaging based platform. We also will use the actual soil data to potentially provide (basic) advice to the farmers, because there is also UX data that can be gained from this. The farmers are not paying for this (they just put down a refundable deposit for the sensor), so we see it as a a way to gain empathy data and simultaneously increase the overall traction of FarmEd.

At this stage, we have validated that there is technology accessibility in Timor, and that there is a want for agricultural advice. The point of the case studies is to validate farmer's engagement in using technology to receive that advice. So, if the case studies have a positive outcome in December, I think the MVP could be rolled out, but the December team needs to assess that themselves. We should only move forward when we feel prepared and we shouldn't make promises we can't keep.

We carefully worded the written agreement so that the December follow it. The farmers are not paying for anything, so what we provide is reflectively basic (yet still tangibly useful). For more comprehensive analysis, this could be provided as part of an MVP in the future, and we would need to ideate the best ways to comprehensively analyse farms for this.

Thanks again,

Reply 1

Wade Tink Jul 27, 2018

Confused as to whether this is proposed or you have implemented it? Could you give us an understanding of timeframe?

Users tagged:

Reply 0

Reuben Dowie Jul 27, 2018

We have implemented (five) case studies. They will run until the December team steps on. The December team will give farmers their $6 deposit back, take the sensors back, empathise with the farmers regarding their user experience, and if appropriate, provide soil advice to them.

Reply 0

Justin Hakeem Aug 27, 2018

Status label removed: Proposed Experiment

Reply 0