Our first meeting with Sven and Ravi was very productive -- alot of good feedback, guidance, and actions to follow up. Their thoughts, along with the discussion with teaching staff and amongst ourselves led us to pursue the weeding business.
Quote of the day: "Those poor migrant workers! You guys are evil!"
Key questions:
-Can we find a market big enough to build a business?
-How segmented is the market? Different weeds for different crops? Each crop need a different system?
-What is the farmer's value chain? Who do they trust? Who do we align with? How do they acquire equipment? How does the channel work?
Actions
-Talk to more customers!
-Determine market segmentation, more detailed breakdown by crop. Find market concentration.
-Talk to agriculture department. Research weed types
-Gather data - overhead video, well labeled "weed" vs. "plant". Give to CS guys to test if image recognition is possible
Advice:
-1st customer is key. Must be vocal and also have skin in the game (Bolthouse Farms could be that customer... more to come with them)
-Good that this would only require small investment to determine effacacy of sensor system
-Labeled dataset for input to image classifier is the treasure trove - key IP!
Meeting # 2 is scheduled for February 2....
Other thoughts that came up during the meeting:
ReplyDeleteWe should draft a letter of interest outlining our product and estimated cost and determine if there are customers who are willing to sign that they are interested.
We should also figure out what the minimal experiment would be to validate the viability of the technology.
We should think carefully about our sales channel. The average sales rep needs to bring in about $1M in revenue a year. How many deals would this require?
Finally, our group needs more granularity on the market breakdown. How many different crops will one version of our technology be able to address?