Big data. Machine learning. Artificial intelligence. All commercial cannabis cultivators should be paying attention to these concepts, as the application of data analytics is beginning to gain a foothold in the cannabis industry.

As the industry matures, operators are raising the bar for optimizing efficiency, increasing consistency and, most importantly, profitability. Data is the vehicle for accomplishing these goals.

Methods of Collecting Data

It all begins with data collection. Before turning on the switch for a flood of data, it’s important to understand what data points need to be tracked and why. Actionable data is the key here.

Data collected will differ across the types of cultivation businesses and their objectives. Different methods of cultivation will have different data points to track. What’s important for an outdoor cannabis farm might not be important for an indoor grow. Take, for example, soil testing, which is important for an outdoor farm, but not relevant to a hydroponic grower.

– Personal observations: Data collection can be as simple as doing a daily walkthrough to take stock of your plants. Visually inspecting your plants and noting anything out of the norm is the most common way to begin collecting data from your grow.

First-hand observations are a great source of data, especially if they’re stored in an easily accessible database.

Many long-time cultivators have used this method. These growers mastered their craft over time through a combination of dutiful observation and intuitive connection to the plants. The trick is to stay as objective as possible. Without a steady stream of data to back those guesses, it’s hard to know with confidence.

Historically, growers have tracked data in notebooks and on whiteboards. Usually, they keep track of feeding schedules, environmental conditions and how many days are left in the vegetative or flowering cycle. While collecting data by hand has been the status quo for decades, it’s not practical for gathering insights on a large scale.

The simplest and cheapest way to collect and centralize data is in a spreadsheet. There might be a number of tabs to support tracking different metrics, locations or strains. While this method is the easiest, analyzing this data is the hard part. One might see a pattern emerge after collecting data for some time, but the more complex a scenario, the more difficult it becomes to interpret and use for experimentation.

– Equipment: Modern grow equipment can provide diverse data for cultivators. Environmental sensors are the most common tools for tracking climate conditions in various zones of a cultivation facility. For indoor cultivators, a controlled and consistent environment is critical to a successful harvest. Data points like temperature, humidity and CO2 are important to track.

Larger grows depend on automation with advanced control systems. Cultivators can monitor and control lights and HVAC equipment with precision, remotely. Sophisticated control systems also offer data analytics and reporting that can optimize usage.

Having access to real-time data provides instant value when a system alerts you that something isn’t right, like an air conditioner not turning on. But there are greater insights to be had from reviewing data across time periods, locations or strains.

– Software: Making sense of all the data is the biggest challenge. Those notebooks filled with notes and numbers typically get forgotten after the next round or notebook is started. How would one even go about synthesizing the trends, outliers and variables from three months of notes — especially if there’s no data scientist on staff?

Cannabis software that’s built with for cultivators in mind can help synthesize data into actionable insights and provide access to it in a way that empowers strategic decision making.

The Many Variables

When there are many variables, it can be hard to pinpoint the exact change to the process that produced the highest yield or quality.

Tracking all the variables manually and in isolated systems has previously been the go-to method. But these days, commercial grows are adopting digital systems that provide instant access to data for all stakeholders. Studying the data can reveal areas of opportunity for streamlining and optimizing results.

The following list of variables offers some ideas for data points to map out and track in a commercial operation, in order to accomplish your goals.

– Physical data: lighting level and spectrum; soil type; pH fluctuations; nutritional requirements at various stages of growth; insect/pest presence; crop condition; yield estimates; and labor times on tasks, batches or strains.

– Environmental data: water quality (including run-off, leaching and residues); CO2 levels; temperature fluctuations; humidity levels; and energy balances for carbon footprint measurement.

– Commodity data: actual yields; qualitative attributes; trends in market demand for product types and genetics; and price analysis.

The Scientific Method for Growers

There are seven steps to the scientific method that should be followed. By following it consistently, you’ll be able to confidently say what change to the process resulted in the desired outcome. Then the test can be scaled and replicated across facilities and teams.

– Define the question you want to answer or the information you want to know.

– Conduct research and gather information. This could mean reviewing data from previous rounds or speaking to employees to get their perspective.

– Establish the hypothesis, which is like an educated guess. It’s the answer to the question that you’ll test with the experiment.

– Experiment to test the hypothesis. Remember to only alter one variable at a time.

– Collect and record the data from your experiment.

– Analyze the results to form a conclusion. Do the results validate your hypothesis?

– Communicate the results and replicate the study to begin the experiment again or test a different variable.

Strategic Business Decisions

But sensors and data tracking systems are just the beginning. Organizing the data to make sense of it all is what brings real value to businesses. Cannabis software and other tools that consolidate data from different sources are intrinsic to an operation that seeks to optimize its processes and results.

The general goal of data analysis is to discover the right combination of variables that produce ideal results — and to make them repeatable and ultimately scalable. Finding that sweet spot is like a gold mine. Once there, keeping things consistent across your whole operation is the next challenge.

It’s not necessary to track all the data in the world, but it may benefit you in the long run to have historical data to access when questions arise. However, if you’re interested in a specific cause-and-effect hypothesis, set up an intentional experiment with precise metrics to track.

Consider an operation that wants to establish the perfect curing process.

After harvest, growers are excited to finish the cycle, sell their product and bring their cash flow back to life. But because terpenes are volatile and extremely important to the smell, flavor and effects of the final product, it’s crucial to dry and cure carefully and deliberately.

Consider the many variables that go into the curing process: How close are the racks and hanging flowers? Is there enough airflow between the branches? How strong are the fans and wind speed? What temperature and humidity results in a slow and steady cure? How long of a curing time achieves the highest terpene levels? What about for the highest THC percentage?

Experiment with these variables by using the scientific method. To get insights with confidence, test only one variable at a time while keeping other variables as constant as possible. Review the lab results after each round and compare the THC and terpene levels.

By using data to tweak the conditions or setup over time, you’ll find the optimal combination that delivers consistent, quality results each time.

Competing Goals

When setting goals for tracking and experimentation, it’s important to first identify the priority. Different business models will prioritize one goal over the other.

Take, for instance, yield versus potency. In a commercial cannabis grow, where less personal attention is possible for each plant, total yield might be the dominating metric. However, for a boutique genetics cultivator, high THC and terpene levels are desired for their higher selling price.

Whether higher yields or higher quality are the drivers for process optimization, one should use the same method for analyzing the success of the experiment.

Calculate total input costs, including labor, to determine whether profit margins and net income are higher for smaller yields but higher quality, or the other way around.

Once you introduce different cultivars into the mix, inputs and variables can change. Keep things consistent when experimenting and you’ll have the insights that empower strategic decisions.