Imagine this: a greenhouse equipped with the latest HID lighting, automated irrigation & nutrient distribution systems, as well as sensors that can detect the humidity, the height of the plants and even cultivation readiness based on image recognition of the flower.
All of these technologies exist today in different phases of maturity. Some of the larger cannabis licenced producers are already equipping their greenhouses and production spaces with them. So, have these companies already fully embraced Industry 4.0? Not yet! This is why:
Industry 4.0 refers to the fourth industrial revolution where machines are equipped with wireless connectivity and sensors, in a fully integrated eco-system of applications that can visualise and control the entire production process and make decisions on their own.
“From the first industrial revolution (mechanization through water and steam power) to the mass production and assembly lines using electricity in the second, the fourth industrial revolution will take what was started in the third with the adoption of computers and automation and enhance it with smart and autonomous systems, fueled by data and machine learning,” describes Forbes contributor Bernard Marr.
When it comes to cannabis, there is still a lot to be done to adopt this fourth revolution. As Tahira Rehmatullah, one of the top 5 women in cannabis according to Fortune, put it at the last Collision conference in Toronto: “We’re seeing cannabis 2.0 right now. There’s still 3.0, 4.0, all the .0’s to come.”
In today’s “smart” greenhouses, every device is connected to the Internet with its own Internet Protocol (IP) address, that allows it to be controlled at a distance in a fast and effective way. This is what is referred to as IoT: Internet of Things. The Master Grower controls the greenhouse according to his or her own experience and known recipes. This already represents a huge step forward because large spaces can be controlled, and different conditions can be ensured for each species by room, zone and stage in the growing process.
However, even those that are the furthest along the road to 4.0 with these smart greenhouses still have one more step to take: leveraging the data and machine learning.
Machine learning and Cannabis 4.0
Having thousands of sensors connected to the Internet sending information at every second produces a huge quantity of data – too much for a classic data analytics platform to handle. This is when we will begin hearing about Big Data.
Big Data consists on using specific technologies to conduct computational analysis of vast amounts of disparate data to reveal patterns, trends, and associations. Upon the analysis of those patterns, trends, and associations it is possible to transform these results into actionable insights:
“Big Data is the perfect way to obtain actionable insights based on the analysis of trends in data silos such as production, sales, HR and any other data sources relevant to your cannabis business. Big Data will unlock the wealth of information you already have at your disposal. By using that information to its fullest, you gain considerable knowledge in order to improve your whole process in a measurable manner” mentions Sébastien Daupleix of Montreal-based custom and IOT solution provider, Uzinakod.
Finally, after all that information is collected, machine learning is applied to the data.
Basically, machine learning (ML) is an application of artificial intelligence (AI) that allows systems to learn and improve without being directly programmed. For example, if a machine is shown enough pictures of a cat, it will learn to identify cats in other pictures. When the machine makes a mistake (these are detected by a human), it will take that mistake into account to modify its algorithm, making the algorithm better every time.
This same logic can be applied to crop management. By knowing the “recipes” of light type, nutrients, atmospheric conditions, and energy consumption for each stage of the growing cycle, combined with the output (in ounces/grams or better still, the cost per ounce/gram) the machine will be able to tweak those recipes. “Instructions” to produce the best yields in the future will be suggested by the machine.
Since no human can process that amount of information, the Master Grower will continue to make important decisions, but will be greatly aided by the results of machine learning. The Master Grower will become the innovator – by trying new things and letting the machine learn from the outcome, it will be a lot easier to experiment on a multitude of factors: producing the lowest cost flower, the highest cannabinoid content or the plant that needs the least water.
Getting to a granular cost with Cannabis 4.0
Let’s talk about cost per ounce/gram: in order to be able to optimize this metric, the data related to yield generally found in the seed-to-sale or crop management software needs to be tightly coupled with costs generally found in the ERP (Enterprise Resource Planning). Moreover, costs in the ERP need to be broken down to the smallest level that are feasible to manage. The more the costs can be associated to a specific strain, batch or even individual plant, the more useful the data on yield will be for machine learning.
Say labour is spread over all strains and batches based only on weight. If a particular strain is more labour-intensive, then the costs for that strain would be understated and overstated for others. Add to that some costs aren’t even included – they are only found in overhead. For example, the labour to change lights is likely found in building maintenance – generally considered overhead.
Now imagine that the labour cost to go up and change the light bulb outweighs the increase in productivity due to the increased luminosity. The analysis shows that 10% of the lights need to be burned out before it’s cost-effective to send someone up. Better still, when sending someone up to change lights, they are told to change others that are about to burn out based on predictive maintenance schedules. To make these kinds of decisions, the data, machine learning and detailed costing in the ERP all need to be in place – that’s the power of 4.0.
Don’t get me wrong: there are trade-offs. Managing costs at that kind of level in an ERP can quickly become a data management nightmare – it’s for that same reason that a lot of companies opt for standard costing vs moving average. Constantly fluctuating inventory values can be hard to manage – just think if you are managing those values at the individual plant. But that’s the direction the industry is headed. Companies that achieve that level of analysis first will be the best positioned to produce the cannabis with the highest return.
Cannabis and the road to 4.0
The road to cannabis 4.0 requires planning. As smart greenhouses are built, thought needs to be put into where to store the data and how it will be analyzed, on an accessible but secure IoT platform.
Secondly, any application containing data that will need to be compared or analysed should be chosen carefully. Here are a few questions that should be asked of potential vendors before acquiring new software or technology:
- Do they have open APIs allowing other applications to access and exchange data?
- Can they support tracking, costing and inventory valuation at the level needed?
- Does their application have an artificial intelligence (A.I.) roadmap?
And lastly, start planning early for data-driven decision making – this will come considerably before machine learning. Implement best practices on data from day one to avoid the infamous garbage-in, garbage-out. We’ve all heard stories about major clean-up of master data (item, supplier, …) when a company implements a new system. Avoid this pain by defining master data early and consistently across all systems, and be vigilant about maintaining its quality. This will facilitate integrations and decision analysis later on.
While the road to cannabis 4.0 may not be mapped out with a guaranteed-success recipe, the technologies to get there exist and are maturing fast – the steps to get there are clear. Those that start down that road earlier rather than later will surely have a competitive advantage in an industry full of new entrants and constantly evolving regulation.
Rachel Bachmann is President and Founder of Akiri Consultants (www.akiriconsultants.com), an IT consulting firm. Her firm helps clients in Manufacturing, Distribution and Retail choose and acquire the right technologies for their business, with a growing number of clients in the cannabis vertical.