No margin for you.
Prime Future 152: the newsletter for innovators in livestock, meat, and dairy
One of the challenges with agtech adoption is that we <the collective we for all the people working to bring tech solutions to ag> are generally selling into cyclical, volatile, relatively low-margin businesses.
Those are definitional characteristics of commodity markets which make them definitional characteristics of businesses that are in the business of commodity production.
Someone was recently telling me about a diversified producer who grossed $12M in 2022…and netted $300k. Ouch. This person is considered a top-notch producer and the discussion was about the concern this is more of a trend than a one-off bad year. Granted he's building equity in the land, but still....those are excruciatingly tough numbers. And you don’t have to look very far to find similar examples.
If you were setting out to buy a business, without knowing anything else about this business except the financial profile, you'd likely pass.
I call this out because 1) I think a lot of folks in agtech miss the economic context their customers operate in, and 2) this context of cyclical, volatile, low-margin business plays out in a few ways as it relates to agtech adoption.
While I see this as the next layer to the discussion about Agtech 1.0 and why farm management software did not deliver what most of us generally thought it would / could deliver, the following considerations are not specific to software - they apply to the entire universe of new products a producer might purchase, from equipment to feed additives.
(1) You can only stack so much value creation.
If a farmer bought every feed additive/widget/other product that sales reps sling while promising 3 points of feed conversion, then cattle in a feedyard would convert feed as efficiently as a 5 lb broiler and pigs would convert like fish.
News flash, they don’t.
Most agtech products are ultimately about increasing efficiency and reducing cost. And you can only drive so much cost out of the system, particularly when the system is already generally pretty efficient.
The real question is not how much efficiency is created by this new product; it’s how much efficiency is created by this new product and under what conditions. The existing combination of management practices, genetic decisions, nutrition SOP’s, etc all impact the maximum potential ROI for a given producer by adopting any new product. But the conditions for max ROI are not normally articulated well, if at all.
(2) You can only detect so much difference in a system with a lot of noise.
There's natural variation in production systems caused by weather, disease, genetics, management, labor, feed formulation, feed ingredient consistency, markets, and a million other factors. There's variation whether we're comparing feedyard pen to feedyard pen, dairy pen to dairy pen, or poultry house to poultry house...let alone when you get to the individual animal level.
All of that natural noise can make it difficult for producers to attribute performance changes to any one thing, particularly when multiple things are changing at once. If a producer has a hard time ascribing value to a product, they will (naturally) be willing to pay less. So this one can be super tricky.
(3) You can only capture so much of the value you create.
The general rule in ag is 3:1 ROI.
This means the producer has to be able to *detect* that 3:1, see point #2.
And this assumes the value proposition can be quantified, which isn't always true.
Sometimes value proposition is more about the alignment of management philosophy, e.g. a packer who fundamentally believes in the power of analytics is more likely to invest in analytics even when ROI is really hard to nail down precisely.
Going back to our farm management software discussion, it's really hard to nail down the value in a quantified way; it's more of a "does it make my job easier/better" thing which is subjective. Subjective value propositions are fine, if the perceived value is high enough. But clearly there are some segments of livestock/ag where the perceived value is not high enough since adoption remains low.
A lot of non-ag SaaS companies have been built around the idea of replacing processes otherwise done in Excel. Which is all good and fine when you're dealing with a huge market, or a non-cyclical market, or a mid-margin market.
The above dynamics may be a few reasons that idea hasn’t translated directly to agriculture and that agtech plays don’t always follow a predictable 'up and to the right' sales story.
But I think the biggest challenge is that no matter how high impact or easy to use or <insert other positive tech adjectives> your product is, if you're selling into a cyclical, volatile, low-margin market then...it's just a bit harder than it would be otherwise.
But, of course, it's not impossible.
Frank Wooten, founder of Vence, spoke at the Animal Agtech summit about his first step as a founder, which was to look at financial statements with cattle producers in order to understand their economic context. He and his team then backed into their product design decisions based on that understanding of their customer's financial reality, a complete flip in the order of operations many companies take.
IMO, flipping the order of operations in that equation is one example of the kind of thinking needed to be successful in markets that are cyclical, volatile, and relatively low margin.
Because it's not that there's no margin for an ROI-creating product; it's that there's minimal room for error in these types of businesses, which elevates the bar right out of the gate and demands a different level of creativity & intentionality from tech providers.
Perfect 3 points. Really we should use these as invetment criteria. An obscure point on noise I have been pondering. In manufacturing we talk about the need to have the tollerance band within 3 stdev of the mean. If so the process is "Under control". What that also means is you can attribute data to cause. If Signal to Noise is low, you can not. Six sigma takes this further to say you need to be at +/- 6 stdev. I sometimes feel like I am the only person that gets this (except all the other people that have been in manufacturing). I like it as an investment filter. And as a goal. In many ways I would like Vence, Earth Optics, Tillable, Sentara and Airable (and even Benson Hill) to optimize the variance. Frankly I would like all the crop input companies to minimize the variance. And all the measurement/data companies to observe the variance and their correlation.