10/1/2021
Forecasting 101
Katie Elzer-Peters
What’s your optimal profitability? Selling everything you can at full price, obviously,” says Ken Lane, owner of Hathaway & Lane, a firm specializing in forecasting and planning for wholesale, retail and growing operations with an emphasis on horticulture. “Of course, that’s not 100% achievable,” he continues. But the more you plan, the closer to that goal you’ll get. “Nobody wants to throw away plants because they grew too much, but they don’t want bare shelves on May 15 either.”
I’ve worked with Ken since 2008 and watched him create amazingly beneficial inventory, labor and other forecasts for businesses of all sizes, both brick-and-mortar and direct mail transitioning to e-commerce. This summer, I caught up with him to get some pointers for businesses starting to do more forecasting to maximize profits, with an emphasis on how COVID-19-related changes affect forecasting.
Every Forecast is Wrong
Ken wants to make sure we all approach forecasting with clear expectations. Forecasting is, he says, “A very detailed and inexact science. One thing I can tell you about every forecasting is that it is wrong. However, it isn’t the forecasting process that is wrong, it is the assumptions that go into it.” (More on that in a bit.)
When it comes to inventory forecasting, Ken says you should NOT expect a forecast to do or be:
• Static
• 100% accurate all the time
• A replacement for intuition (Ken says intuition is, in fact, a “variable” that can go into building a forecast.)
Ken recommends creating three forecasts, each with slightly different assumptions. (What if freight costs triple? What if they double? What if they stay the same?) It won’t surprise you that one of his favorite sayings is, “If you don’t have a plan B, you don’t have a plan!”
Whether you make one forecast or three, you can’t let the forecast sit on the shelf. You have to constantly review what you thought might happen (assumptions), and why you thought those things might happen against what actually is happening, and why those things might be happening. Then you have to update your forecast—by updating the assumptions used to build it.
Assumptions = Building Blocks of Forecasts
Every forecast is built with assumptions, which you could also call historical data. For instance, “How does demand for your product relate to the unemployment rate?” asks Ken.
That’s not something people probably thought about at length before COVID-19 and I’m not sure what we would have assumed before living through 2020.
“It turns out that higher unemployment means more people gardening,” Ken says. Other assumptions or data used to create forecasts include:
• Last year’s data, such as number of orders and SKU sell-through rates
• Trends in your business data
• Fluctuations in business KPIs—items per order, dollars per item, dollars per order
• Forward thinking
• Changes in price
• Changes in marketing activity
• Seasonality changes
• Geopolitical/world event distractions
The way you use assumptions in forecasting is with a lot of “if/then” thinking. For example, IF you’ve sold 50% more tomato plants than what you forecast (acquired to sell) for three weeks in a row then … what? What do you assume will happen in the next three weeks? Why do you think that’s happening? And if you expect to continue selling more tomato plants than you forecasted you would, how do you then change what you’re doing to meet the moment?
Get Started with Forecasting
Ready to put some pen to paper and build your own forecasts? Here are four steps to follow:
Step One: Lose the pen. Your forecast is going to change after you’ve reached a point of statistical significance. Use a pencil.
Step Two: Organize your historical data, which might be limited based on time in business or what types of data you
collected.
This is the type of data to gather and to eventually start tracking:
Revenue, broken down
Orders
• Average order value
Items per order
Average dollars per item
• Source of orders (for e-commerce)
Paid search
Organic search
Email
SKUs
• Prices
• Sell-through rates (both from LY and what is desirable)
Step Three: Review and learn from the data you’ve gathered. You learn by looking at your forecast and observing how you missed or exceeded the forecast. Ask yourself, “What number, or what assumption, was wrong and what can I learn from that?
“But, Katie, I didn’t MAKE a forecast. I am new to forecasting.” To that I say, “Oh, but you did make a forecast of some kind. One forecast you made without knowing it was the volume and type of inventory you brought in to sell, when you had it arrive, how you priced it and the dollar amount for which you eventually sold it.” So look at that data and answer these questions:
• What do your KPIs look like for the past two years?
• What are the trends inside of the past two years?
• What does that mean for the next two years?
• What does that mean for your supply chain?
Step Four: Revise your forecast and keep going. Over time, expand the types of data you track as time goes on, based on what you’ve learned so you can continue to tweak your assumptions and fine tune your forecasts.
Plan, collect data, analyze and keep going. (Rinse & Repeat.)
Getting the Right Help
If you’ve never produced a formal forecast and used it to make business decisions, the idea of doing so can be daunting. There are people who can help, though, and many of them have been interviewed or have written articles in this very magazine. When you call up a business consultant and ask for help with forecasting, here are some questions Ken says you should ask them:
• What do you think are the KPIs of my business?
• What’s the highest number of variables you’ve used while creating a forecast. What did you learn from them and why?
• How much data do you need to get started? (Hint: Ken recommends at least three years of data, or if three years aren’t available, as much as you have. In his words, “Three’s a trend.”)
• How do you know it’s time to update a forecast? (How do you know it’s time for plan B?)
This is just the tip of the iceberg when it comes to forecasting. Simply getting started, even if you’re just looking at performance of a few top SKUs and a few trends in order value can make a big difference in your overall business. I challenge you to take a stab at it over the holidays, a slightly less crazy time than spring. Plan, collect data, analyze and keep going.
A Forecast in Motion
How does all of this work?
Here are two sample scenarios:
Units Sold
Say your season is 10 weeks long and you’ve planned to sell 100 units of something per week. You want a 90% sell-through rate overall.
You open for the season. During the first three weeks you sell only 8 units. The assumption that you would sell 100 units per week was wrong. Ask yourself, “Why was this assumption wrong? Was the weather different? Did you sell units at $10 last year and they’re $12 this year?”
To meet your sell-through rate, you have to break price. Now you’re in a different forecast.
Like Ken Lane says, “Three is a trend. If you’re missing or exceeding your number three weeks in a row, there is an assumption that is wrong and what does it mean for the rest of the year? How do you change your forecast?”
Lower Revenue
Here’s another scenario to help you get in the mindset of if/then thinking.
Ken says, “In a vacuum: if 2020 was 50% greater than 2019, and 2019 was 3% over 2018, what do you think 2021 will be?” If you project that you’ll receive less revenue in 2021, are you:
• Getting fewer orders?
• Seeing lower spends per order?
• Experiencing a lower retention rate?
• Serving fewer customers?
If you project that your revenue will be lower from one year to the next, there will be a reason (or two or three). If you can identify the reason(s), you can make better plans. GP
Katie Elzer-Peters is the owner of The Garden of Words, LLC, a green-industry digital marketing agency. Contact her at Katie@thegardenofwords.com.