You’ve undoubtedly heard about AI or Artificial Intelligence, whether it be IBM’s Watson beating the best chess players in the world or Amazon Alexa’s role within the modern home. Business owners and operators that aren’t among the elite tech companies probably don’t think AI is relevant for their businesses. Indeed, the popular examples do come from Silicon Valley today. However, beneath the jarring headlines predicting AI’s takeover of the world, real, approachable companies are building AI strategies that will set them up for strong leadership in their industries for the next generation. How can you make sure your business doesn’t fall behind?
Start with understanding data
To understand AI, you first need to understand data. Data is another term that’s been used in business publications more often in recent years, but one might wonder what it really means to utilize data in the way they talk about it. All of our companies need to track revenues, expenses, assets and liabilities for accounting purposes. At the end of the day, that is business data in its most basic form. It’s a way to observe your business cleanly and consistently using numerical data points displayed through the lens of a specific time. However, accounting has been around for a while and it isn’t what the business pundits are making a fuss about. It’s about going deeper in your efforts to describe what’s going on in your business.
When thinking about describing your business with data, we must first identify what our goals are. Outside the basics of knowing how much we sold at the end of the day, we usually look at data to make decisions. When making any decision, you naturally need to pose a question. An example of a question using our financial data is ‘How are sales correlating with our payroll numbers as the dog days of summer set in?’ Since we can’t operate with the same payroll in July as we do in May, we reduce hours. How many hours to cut and when to reduce them are the questions you’re asking. In this case, the data you’re using to make the decision are sales and payroll.
Let’s peel the onion back another layer and get more complex with our questions. The right questions take aim at one or more entities. I tend to think mostly about two primary entities in our business: customers and products. Like primary colors, you can mix them and get more colors. When customers and products mix, you get transactions (at least that’s the goal!). A lot of the juicy insights come when you look at transactional output data, and that’s where business leaders spend the most time and energy. However, in order to generate as many transactions as possible, we have to spend time with the input data.
Customer and plant data today
Our products are a wide-ranging array of unique plants that can be described across many dimensions. They can attract pollinators, resist deer or smell good. We also want to know what sun and soil conditions they can handle. We have all this information somewhere, right? It might be on word documents, plant tags, signage or online.
You’ve heard before the importance of understanding our customers. Good businesses do this well. Despite privacy concerns, companies are pushing forward and acquiring as much data as possible from their customers. Luckily, it’s much less invasive to ask customers about their soil and sun exposure than which websites they’ve viewed lately. Our salespeople ask these questions organically when chatting with customers in the store. The most knowledgeable salesperson can call up plants on the spot that fit the sun and soil parameters as described by the customers. The least knowledgeable may do the same by using those various plant information sources. This has worked just fine for decades. That said, I don’t want to settle for ‘just fine’ as I know there’s a better way.
Digging for data in the future
I would first question how accurate the data you’re getting from the customer is. This data is a rough observation from someone who likely doesn’t have a horticulture degree. Even in the industry, we’ve settled into rules of thumb that describe something as complex as sunlight into three to four categories between ‘full sun’ and ‘shade.’ This is necessary to make it simpler to understand for the customer. However, with today’s technology, we have very accurate light meters that can put light into numbers in measurements like foot-candles. With GPS data, that light data could be pinpointed to location across time. These are mathematical equations you can’t do on the back of a napkin and would even struggle to do in Excel. Enter artificial intelligence.
In about 10 years, through the use of IoT (Internet of Things) or internet-enabled sensors, light and soil information will be detailed and accurate across a customer’s yard. If a customer wants some plants in specific spots, the data can tell us about the requirements for a plant to be planted there. Now, it’s up to us to get our plant data organized. Those Word docs and fact sheets won’t cut it. This data needs to be organized tidily in databases, complete with the information necessary to answer the questions we’ll be asking, like which plant will work for this customer? If done correctly, AI applications of the future will be able to utilize this data to answer all your questions.
To take this a step further, once you’ve organized your ‘data model,’ it inevitably won’t be perfect. You may have information from growers about plants they know in their own area, but they’ve never seen grow in your area. Just as nature is diverse, so are our customers’ yards across the country.
The model can be used to drill down more precisely by collecting data on the plant’s life after it’s planted. By logging the planting location, you’ll be able to see how well it does given the sun and soil conditions described in the data. Those same IoT sensors can monitor moisture levels over time. Inevitably, when plants die, we log that data as well. We may be able to prescribe a cause of death easily by knowing whether the customer watered their plant. Beyond that, it will contribute to the statistics of this plant as we get more and more data on that same variety.
Then, we are able to ask more detailed questions of the data. For example, “How well does this Catmint do in wet conditions?” You may get back an expected lifespan of 1.36 years. If this customer was expecting this to grow to fill a space over several years, this isn’t the plant for them. In fact, if the garden center built a program to outfit the customer’s space with sensors, they’d be able to pick a place, then look through the inventory of plants that have the highest odds of surviving and establishing for the long-term. Warranty programs can be built around this. Over 90% survival odds may garner a three-year warranty. On the other hand, maybe we don’t offer a warranty to plants under 50%. This can dramatically decrease replacement costs.
It’s not too soon
For those garden centers that are prepared, AI can make their businesses smarter and make their customers more loyal. It will represent a disruptive force in the industry. We can only hope it comes from within the industry rather than a highly funded startup. How plants are bought and sold has changed dramatically over the past couple of years, but the pace of change isn’t going to slow down. AI will be a much more profound technological force than e-commerce was before it. It’s not too soon to organize your data and start to think about the garden center business models of the future that will be afforded by AI.
