CALPEST

California Pesticide Electronic Submission Tracking

The problem: Registration ran on paper, and paper moves one desk at a time.

My role: Product owner during development, January 2022 to June 2023. I owned the product from discovery through customer acceptance for a statewide system.

The system: California launched it in September 2024. Final phase, late 2025.

Scale: About 5,000 submissions a year, 700+ registered users, 13,500 registered products.

To sell a pesticide in California, you register it with the Department of Pesticide Regulation. You submit an application, pay a fee, send the product label, and include the data that supports it. Then up to ten scientific evaluation stations review it: chemistry, ecotoxicology, microbiology, groundwater, surface water, air quality, worker health.

That ran on paper. Paper gets lost, and paper gets misrouted. A submission would leave one desk and never reach the next one, and nobody found out until someone asked.

The fees came in on paper too. DPR took checks in the mail, which the department said could hold up payment processing for as long as fifteen days.

Farm workers handle these products. Groundwater carries them. Food is grown with them. Those delays sat between a scientist and a decision that mattered.

Only one person can hold a piece of paper.

So the application went to chemistry, and when chemistry finished, it went to ecotoxicology, and then to groundwater, one station after another. The waiting in line was the slow part, not the reviews.

A digital submission has no such limit. Every station that needs the data can open it at once. The state's auditor said it plainly: CalPEST was built to allow concurrent evaluation.

That is the product, and everything else serves it.

State labor rules kept me from interviewing the end users directly. So I interviewed DPR staff who worked next to them, treated those conversations as proxies, and stayed honest with myself about the distance between the two.

Then we built what those users would live in: the workflows, the procedures, the artifacts. If I could not watch someone work, I could at least make sure the thing we handed them matched the job they do.

 I designed for concurrent evaluation without ever talking to the people who'd use it, working through DPR staff as proxies. The state's auditor confirmed the design was right.

Twenty-four legacy systems had grown up around this process. Each one solved a piece. None of them talked to each other. We consolidated them into a single platform on Microsoft Dynamics 365.

Dynamics gives you a lot out of the box. The hard part is that Dynamics documents its elements, not your process. No page tells you which combination of components adds up to how California registers a pesticide. You build that map yourself, and every gap in it becomes a decision about whether to customize.

We customized where DPR's process had no off-the-shelf equivalent. I weighed how much of the business process a customization would cover, what it cost in engineering time, what it depended on, and what it would break later.

Customize everything and you have built a bespoke system wearing a Dynamics logo. Customize nothing and the software does not match how the state works.

DPR's team had recently finished Scrum training, which meant they were doing something genuinely hard: learning a new way to work while delivering a system the state was counting on. 

When you're new to a method, you want to get the method right. Is this a spike or a user story? Did we run that event correctly? Those questions come from caring, and I never wanted anyone to feel small for asking them.

We could spend an entire session getting Scrum right and move the outcome exactly zero. So I kept gently walking us back to why we were in the room. I facilitated the events first, so the team had a working example in front of them instead of a rulebook, and then the DPR Product Owner took them over. We iterated through all of them together until they were theirs.

The method is a means. The outcome is a scientist getting the application they've been waiting for.

I chased consensus when I already had alignment.

Alignment means we agree on where we are going. Consensus means everyone feels good about how we get there. I kept reaching for the second when I already had the first, and it cost us weeks. A decision was made, we had what we needed to move, and I would reopen it because one person still looked uneasy.

Some slowdowns came from elsewhere. We lost time to disagreements over business rules, over process, and over what a term even meant. That happens in a system that has to satisfy scientists, attorneys, and a regulator at once.

The consensus habit was mine, though. I could have named the decision, written it down, and kept us walking. I say that to teams now. Alignment is enough.

A company registers, associates its people with the business, and submits: new product registrations, amendments, data, renewals, ownership transfers. Two roles carry the work, Company Administrator and Company Submitter, and the administrator approves who gets to act on the company's behalf. CalPEST checks federal registration status against the U.S. EPA's own listing, verifies addresses through USPS, and takes payment by card, transfer, or check.

Dashboards answer the original problem. Action Required. Pending. Completed. Drafts. Invoices. When the state needs something from you, it tells you. When your submission has a deficiency, you see it, with a due date, and you upload what is missing against the same record.

Nothing disappears anymore.

By the end of my tenure, processing time was down 60% and the product backlog down 50% over six months, every feature accepted.

I was the product owner during development. California launched CalPEST in September 2024, after my time on the project ended, and DPR publishes what happened next:

• New products with already-registered active ingredients: 279 days on average, down from 512 in the legacy system.

• New active ingredients, the hardest category: 0.8 years, down from 3.4.

• DPR cleared the backlogs in its major scientific evaluation programs in 2025.

• Registrants track submissions in real time and pay online.

DPR credits the new system and added staffing together. Both are true.

CalPEST is live. Read DPR's account of the results.