The big greenfield cannabis cloud

The marijuana business is making the leap from paper ledgers to the cloud and big data analytics, offering a provocative example for us all

The big greenfield cannabis cloud

Imagine an agribusiness that never had access to a modern commodities market and had never used any real information technology. Suddenly, it had the opportunity to leap into high tech and potentially enjoy whole new economies of scale -- a pure greenfield proposition. What would happen?

With new international and state legalization, as well as a current moratorium on federal enforcement, marijuana is that agribusiness. Untouched by the traditions of an internal IT practice, it's creating a road map for the future, charted almost entirely in the cloud.

Out of the filing cabinet and into the cloud

To be clear, trepidation remains about digitizing the business side of cannabis. Ask Eddie Miller, CEO of eCann, which runs a site called

Because of the gray area in the laws, I find that a lot of the actual vendors, the retailers, the cultivators are reluctant to go into the cloud, put their data into the cloud, to share this type of information and to have digital records of it.

Nonetheless, with multistate legalization and optimism that federal legalization will follow, some organizations have started moving to the cloud and integrating analytical technologies.

eCann uses Salesforce along with an ecosystem of marketing tools to find interested buyers and investors. The founders are hosting an event during the Dreamforce conference for cannabis investors.

This has raised interesting legal questions. Since the cannabis business is illegal federally and in most of the United States, what is the legal exposure to third-party vendors like Salesforce? I asked Salesforce for a comment on whether it was encouraging cannabis growers and dealers to use its platform. Despite multiple attempts, the company failed to get back to me, saying it was unlikely to meet my deadline even after I extended it.

Baby steps

Legal ambiguity has kept most of the business side of cannabis in file folders or personal cellphones.

There are, however, exceptions. I spoke to the folks from Grow Buddy, and they described their back end as a Cold Fusion, SQL Server, and .Net on some cloud infrastructure. (Ironically, the last time someone told me they were planning to use Cold Fusion, I asked them if they were smoking something.)

Grow Buddy is an app designed to let home growers record the parameters of their “grow,” such as nutrients, watering, and size. At the moment, the app doesn't support cross-user analytics -- many customers regard this as proprietary information. The team wouldn’t reveal where its back end was hosted, other than to say it was a cloud provider and not colocated, regarding that as both a security and privacy issue.

I also interviewed New Frontier Financial, which told me all about its analytics and a lot about its methodology, but on the technology side would reveal only that the company operated a NoSQL database in the cloud -- and it used data mining tools written in R -- not which database or ingestion tools were in operation. The company actually regarded this as a kind of competitive edge. I suggested that if the NoSQL database and ETL tools in use was really New Frontier's competitive edge, it was in trouble.

Data difficulties

Frontier is working to solve an entrenched problem in the cannabis industry: data. Data comes in at different quality levels from multiple sources, and financial data has been hard to come by.

In fact, up until now, even branding hasn’t been protected. In the cannabis industry, marijuana growers in different regions or states frequently call the same strain by different names -- or different strains by the same name. This is a hard data problem to solve.

While Frontier works to solve this at a meta level, another company, PotBotics, has a product called NanoPot that detects gene sequences in cannabis to determine which seeds belong to which strain. Meanwhile, the company has a recommendation engine called PotBot -- and a similar product aimed toward medical marijuana called BrainBot -- that uses an EEG to pick the right personalized strain based on your brainwaves.

PotBotics uses a cloud-hosted combination of NoSQL databases, Hadoop, and Elasticsearch to power its products. The recommendation and discovery engines are based on SPARQL.

A serverless world

No one I spoke to talked about buying servers or building a hybrid cloud. Though none of us know what President Trump might do next year, everyone began with the premise that they should start with a hosted solution -- SaaS if possible, custom if necessary, and without a pool of system administrators.

This isn’t even a conscious choice. I brought it up when I asked if they'd ever thought of buying a server. Grow Buddy was an outlier with SQL Server, but everyone else started with a NoSQL operational database and various other technologies in the cloud, including Hadoop for back-end analytics.

It strikes me that a lot of companies with a lot less legal or financial liabilities claim they need their own data center and avoid the cloud for security reasons. Yet even the “legalized” marijuana business carries the potential for a visit to federal prison.

If that's not an indication of our future in the cloud, I don't know what is. The people I interviewed for this story all struck me as perfectly capable of making reasoned decisions. Despite very real risks, no one was standing up their own servers in a basement somewhere. The cloud, including big data systems to optimize production, was an assumption.

Copyright © 2015 IDG Communications, Inc.

How to choose a low-code development platform