Be careful of platform plays in agrifood
“The chief products of the tech industry are (in B2C) developing new habits among consumers and (in B2B) taking a business process which exists in many places and markedly decreasing the total cost of people required to implement it.”
Beware of the platform!
Recently, Ag Insights platform Gro Intelligence went out of business.
According to WeeTracker a pan-African media company (highlights by me),
“The root causes of Gro Intelligence’s downfall appear multifaceted. A challenging funding environment, coupled with a perceived mismatch between its product offerings and market demands, contributed to its financial woes. While the company touted its AI-powered insights and vast data resources, it struggled to secure consistent revenues, relying heavily on a few key clients for the bulk of its income.”
According to AgFunder News (highlights by me),
“It had positioned itself as a food security platform to a small Asian country and a country in the Middle East exporting oil, without success. It had also attempted to engage the US government under a variety of guises but was only picking up bits and pieces of business here and there.”
“They were chasing deals for projects that resembled bespoke consultancy work as opposed to something that would generate replicable revenue streams.
The website Semafor had some real numbers in their analysis,
“But Gro’s crash, preceded by a valuation plunge from $850 million two years ago to under $25 million, is sparking a sober and clear-eyed review of the startup’s history.”
Let us see if we can take a “sober and clear-eyed review” and break it down further. The only goal here is to learn.
Bulk of revenue from one customer
I have been in situations like this before, where the bulk of the revenue comes from just one customer. It didn’t end as badly as Gro for us, but it was not a great outcome for anyone except the company which acquired us for pennies on the dollar and a few execs who sold the company.
It is important to understand how one ends up in a situation like this, and can anything be done to prevent it from happening.
If you are a new startup in a B2B space with a big ambitious and different vision, depending on the strength of your story telling, you might be able to snag a large customer who buys into your big story. The reality often is that this customer is a set of one, and there are hardly any other organizations with such an approach.
You make the mistake of believing there are a large number of other potential customers out there, who will need similar products and services which this one large customer is buying from you.
If you are not able to expand beyond this one large customer, it starts a vicious cycle.
Given your survival is beholden to this one customer, every time they ask you to jump, you ask “how high”?
You start bending backwards to satisfy every whim and request from your one customer, as they are the ones who are keeping the lights on for you. Relying for a majority of your revenue on a handful of customers is also extremely risky.
You become an extended engineering for hire group for this large company, rather than trying to build products which can be used by a wider customer base. It is a vicious cycle, as you prioritize immediate revenue for longer term growth.
Chasing deals for projects
An overreliance on one customer forces you to chase all kinds of deals to bring in additional revenue. It can also happen in the early days of a startup, when one is trying out different approaches to see what works and what doesn’t.
According to the Agfunder report, Gro was trying to sell to a small Asian country, and a country in the Middle East exporting oil. I don’t have the details on the types of projects that were being attempted to be sold, but I guess they were related to some version of remote sensing models done at an aggregate level.
For example, as late as January 2024, Gro had said,
“For the two major soybean exporting countries, the US and Brazil, Gro’s predictive yield models were 99.6% and 98.1% accurate for 2023” and
“US soybean yield forecast model had been “on average within 99% of the USDA’s final number by October for the last six years” (3 months before the US government published these numbers).
The accuracy numbers seem quite impressive, but the question becomes what is the value of this accuracy, and for whom? Will commodity traders pay for this information, and if so how much?
Will other countries who do not have access to this type of information pay for this information and how much? The information in itself is not useful, unless it leads down a path of feasible value creating execution.
It is not very clear how this information might be useful for downstream companies, even after more than a decade of many of these companies being around.
When you don’t have a cohesive strategy and a story on how you want to leverage your strengths for a repeatable and consistent revenue generating mechanism, you start looking for nails with the hammer you have. You start chasing deals, and your chances of building a repeatable product and a viable company decrease significantly.
You start taking “bespoke” projects to get some immediate revenue, with the desire to convert the “bespoke” project into a repeatable technology product, which gives you leverage in the long term.
Remember, Indigo saying they were five startups within one startup?
If you are unwilling or unsuccessful to convert these bespoke projects into a repeatable product, then you are no longer a product company, and do not deserve the big multiples on your valuation.
Everything for everyone is nothing to anyone
As you start chasing deals, you lose track of your consistent storyline.
You have this particular problem, Ms. Potential Customer? We have a solution, if we put these 3 things together.
You have a different problem, Mr. Potential Customer? We can put these 2 things together, and create a third thing to solve your problem.
Let us see if we can pick some signals from past statements.
In an interview given to Columbia University for the Winter 2023-24 issue,
“Gro has generated some 2.9 million new data series on such topics as the yields of and consumer demand for specific crops, as well as climate risks like fires and floods. Its platform includes user-friendly applications, such as a drought index and an agricultural price-inflation tool, and predictive models. Gro is now working to merge its data with that of other industries beyond agriculture, such as real estate and mining, to establish a more comprehensive picture of climate risk across sectors.”
This is often the case, when you call yourself a “platform” or end up calling yourself a “platform” because you are not able to find an actionable use case, which you can service effectively.
But it is important to remember, it is almost always “product before platform”, irrespective of the industry.
Within the agrifood industry, I am quite guilty of overusing the word “platform.”
HarvestMark (fresh produce traceability platform)
ShopWell (nutrition personalization platform)
Climate Corp (digital agriculture platform)
Mineral (backend AI platform)
Hopefully over the years, I have learnt some lessons along the way.
If you hear the word “platform” within agrifood, you should really look hard at what is being done
- Can the “platform” address a specific use case in a unique way to solve a particular problem?
- Does the “platform” know who their customers are?
- What unique capabilities do you bring to the table to help you scale the “platform”?
- Does the “platform” solve their customer problems in a meaningful and differentiated way?
- What other options does the customer have to solve the problem, if they didn’t use their platform?
- How do you evaluate those options against your offering?
Going back to the AgFunder news article,
According to one industry source who spoke to us earlier this month, Gro “has a great product and a good group of intelligent people behind it, but I’m not sure they know how to sell it… I think it’s a solid business and something that the market needs, but it also needs the right people to operate it. Like most startups, it tries to be everything to everyone and needs to find its niche.”
Gro labeled itself as an “Ag Insights” and the world’s largest agriculture data platform. They tried to be everything to everyone and so ended up being nothing to anyone.
If you believe people are going to use the data and the tools you provide to create additional value, and you can support a variety of use cases, then you could potentially justify the investment in becoming a large data platform.
The logic often used is that if it is a winner-take-all market, and if we can build the largest dataset with significant value creation potential, then we can force everyone to put their pencils down.
It would no longer make sense for anyone to invest in it. For example, once Google built its Google Maps capability, it didn’t make sense for anyone, other than Apple Maps, to build a mapping capability.
In the case of Gro, both the winner takes all, and the value creation hypothesis turned out to be incorrect, leading to misallocation of R&D dollars, while not advancing product-market fit.
Startups in a challenging funding environment
So what should a startup which is operating in the current challenging funding environment do?
How do you survive to fight another day?
There is a clear temptation to chase deals in an effort to get revenue to elongate your runway.
The best approach is to be surgical about who your customer is and position your product to solve meaningful customer problems, be extremely disciplined about your cash burn, prioritize like hell, and stay focused on your core strengths and strategy.
Don’t try to be everything to everyone.
Beware of the platform!