A Biology-Based Approach for Ag Biologicals and Biostimulants

Almost every young biologicals company I meet has the same story. They ran a field trial, the result came back, and it did not answer the question they actually had. Or even worse, it came back as no difference.

That is rarely anyone's fault. It is a design problem, and it starts with a playbook that was never built for biology. This post goes deeper on how I think about yield field trials, and what I want you to consider before you design your next one.

Where the chemistry playbook comes from

Field trial designs did not appear out of nowhere. Over time, chemistry and plant genetics work settled on what was good enough to detect real differences between treatments. In those systems, a lot of the variation comes from the land itself, so the whole industry learned to pack as much as possible into as little ground as possible to cancel that variation out.

In fruit and vegetable crops, a replicate is a group of plants, and you pick three to five of them to get the mean for that replicate. Say you have three plants per replicate and three replicates. That means you are measuring nine plants. In row crops, the unit is a strip of field sized for the combine, and one replicate is usually one combine pass.

Biology breaks that logic. With a biological you are working with biology on top of biology, so plant-to-plant variation from the product itself is higher. Some plants are simply more receptive to a microbial intervention than others. And the problems biologicals tackle, like uniformity and stress response, are not always a clean yield bump. Nine plants cannot show you uniformity.

What I push for instead

In fruit and vegetables, I would rather add plants per replicate than add replicates. More replicates cost more and take more land. More plants let you characterize the variation, and then a statistical model can treat normal plant-to-plant variation as background so you are looking at the variation from the product.

In row crops, you cannot add plants to a replicate. The contract research organization (CRO) often sets the minimum number of treatments because of how they block the planter. So I will sometimes push for fewer treatments and a strip about twice as wide, and for untreated blanks between treatments so there is no drift between plots.

The right trial for the right stage

Not every trial belongs at every stage. Here is how I think about the path.

Very early, before you have a solid mode of action, you need small, controlled work in a garden or greenhouse to get at the mechanism. Play around, get creative, and run robust greenhouse trials when you can. Get a real good idea of when your product produces a response (any response) in a plant, and when it doesn’t. Then start formulating what problem your product will solve.

Sidenote here: I firmly believe that companies should start with the ag problem. But unfortunately that rarely occurs. So instead, figure out the problem you’re solving early. SUPER EARLY. Then build your field trial program around that problem and mode of action.

Once the mode of action is solid, and you’ve found an ag problem, you should move forward with super controlled contract research organization (CRO) trials. If your evidence points to acute drought, you can partner with a CRO and test five to seven crops under that one stress. The ground at a busy CRO may be less responsive because it has seen so many products, but at this stage you are looking for any evidence that crops respond, not a winner.

You also learn about the market. For example, if you are targeting acute drought - short periods of little to no water - you might learn that most fruit is irrigated. So you might not want to pursue markets that are irrigated even though you tested a few of them in greenhouse work. That’s OK. Just because you started in tomatoes doesn’t mean you continue in tomatoes. Pivoting is necessary and not a sign of incompetence.

The next stage is a combination of a few things.

  1. Go back to a few of the same CROs and look at rate and application timing to narrow your application methods. Get more commercially focused on application, handling, and formulation. This will also get you year-on-year results at the same location and facility.

  2. Add larger trials and plots to the mix. Get on more land (not more sites) to look at what happens when we increase the variability over larger land plots.

  3. Narrow down the crops. Choose 1-3 crops to continue with. Drop any you don’t have good feelings about from the data and meta-data you collected the year before.

How I read the results

A non-conclusive (not significant) result does not mean your product does not work. It may mean the system was not the right one. A significant positive result, in my experience, usually means the response is real, and a statistically significant negative is a kill switch: do not proceed. Whatever the direction, I trust the significant ones. The neutral ones are where the work is.

More often than not you will get no difference. That is exactly why your mode of action and your yield trial have to be designed together. You need something to explain a no-difference result: precipitation, timing of fertilizer and everything else about the trial, plus tissue analyses, root digs, sequencing to confirm your microbe is on the roots, or greenness over time. A crop that turns nitrogen deficient halfway through the season and recovers when the farmer adds nitrogen can tell you something yield alone never will.

.

“More often than not, you will get no difference when running a field trial.”
— Whitney

Why you need the translator

Here is the piece most young companies are missing. They do not have agronomy expertise in-house, so they trust the institution, and the trial ends up not answering the questions they need answered. A CRO is not a business consultant. They do not care if the trial aligns with the trajectory of your company. They only care about producing reliable field trials. They will not spend the time diving into your data and advising on the correct types of trials.

This is why I suggest two layers of translation when you’re developing your field trial program. Field trials are the single most expensive part of product development. And companies can waste millions chasing trials that can’t be strung together (I’ve seen it). Spending more money upfront can have lasting positive financial impacts.

Here’s how this works. I uncover the core mode of action, what it does and how, and then look at your current plant data to see where it works best. I work with the agronomist, whether in-house or fractional, to iterate on the questions that need answering. The agronomist then contacts CROs and other trial partners to design trials around the biology, since they have the trial connections, not me. They also take on the day-to-day conversations with those partners. And because some data points will fall outside a CRO’s normal offerings, your agronomist can go collect those samples too.

“I’m working with a client right now where the CEO is having to email back and forth with a trial partner, and it’s about an email or two a day. And here’s the thing. Trials are time sensitive. It’s not something that can just be gotten to when the CEO has time. ”

Biology-based products need biology-based field trials

We have to recognize that not all trials are designed to pick up small differences with high-levels of variability. We need to approach our bio-based products understanding exactly what we’re looking for and designing trials that do just that. We must have enough supporting information that helps us explain why we see no differences which ultimately will help our sales folks sell the product.

Good biology gets heard when the trial is built to hear it.

If you want help thinking through your next trial, reach out and let's talk.

Next
Next

Three Pillars: Sales and Marketing Translation