Forecasting Step 2: Grab Your Data

"Collecting data without a plan is like grocery shopping while hungry; you'll end up with a lot of stuff you don't need." — Anonymous

All right, so you figured out your forecasting project's goals (you read article #3, right?), and now you’re ready to move to the next step. Now what.

Now you need data. You need to check out what happened before so you can make some educated bets on what’s coming up.

Historical Data

So, why is old news important? Historical data gives you a peek at patterns and trends you’ll likely see again. It's like having a cheat sheet for the future. Let’s remember that historical data will not tell you everything. But it’s the beginning of the number-crunching process.

Since we are concerned with forecasting financial metrics in this newsletter, we will see quantitative data more than qualitative. But we should be aware that qualitative data are often necessary and useful too.

Quality versus Quantity

Let's remember, more data isn't always better. You’ve got to be picky about the info you collect and use. Think quality over quantity. You want the good stuff — accurate, clean, and relevant. It’s about making sure you’re not feeding your forecasting machine junk. You have heard of junk in, junk out, right?

Where Do You Even Get This Stuff?

Data are all over the place, but you’ve got to know where to look. Sales records, customer feedback, and tweets about your business can help you. The sources will largely depend on the project’s nature and could range from internal data sources and sales records to external market research, government publications, or social media. For example, if you're forecasting consumer demand for a product, you might pull data from past sales records, market trends, or customer reviews.

Mind the Gaps

Unless you’ve got the world’s most perfect record-keeping, you’re going to hit some gaps. By the way, there’s no such thing as a perfect data set; there is always a gap.? Maybe you’ll average things out or get fancy with some algorithm to fill in the blanks. Incomplete or missing data is almost a given in any data-gathering exercise. So, prepare in advance for how you will handle such gaps. Will you use averages or maybe more advanced techniques like machine learning algorithms to fill in the blanks? The method you choose can significantly affect the forecast’s result and reliability.

Play by the Rules

With some data comes responsibility. You can’t just grab people’s private info willy-nilly. There are rules that must be followed — ethical and legal ones. Make sure you’re not stepping on breaking any laws while you're out there data hunting. Ensure you adhere to all data protection and privacy regulations, like GDPR in Europe or CCPA in California.

In short, gathering historical data isn't just a procedural step but an essential part of forecasting. It shapes the quality of your entire forecasting model.

EXAMPLE: CULINARY HEIGHTS

Do you remember this restaurant from the past article? Let’s see how they deal with this step in the process: collecting data.

As a reminder, their objective is to "increase weekday dinner revenue by 25% over the next three months by accurately forecasting customer attendance and optimizing food and staffing costs."

And they want to do this without the guesswork. Then let’s get some historical data.

Old Menus Tell Tales

First up. They pull out old reservation books, check out which nights had people queuing up. They're looking for patterns like "Taco Tuesdays were a hit" or "Winter Wednesdays were slow."

Count Plates

Then, the Culinary Heights management dives into the numbers — how many avocado toasts did they sell last March? Did the fries brought in the cash? They sift through past sales like they're on a treasure hunt, figuring out what dishes were hits and which ones were misses.

Listen to the Crowd

But numbers aren’t the whole story. The team listens to what diners have been talking about. Maybe everyone’s been raving about the wild mushroom risotto or grumbling that the kale salad's just a garnish. Reviews, surveys, and even overheard chatter become pieces of the puzzle.

Spot the Gaps

In the real world, records have holes. Maybe the new server spilled soup on the reservation book, or the point-of-sale system napped on a busy night. The restaurant has to get crafty, using their best judgment to fill in these blanks so they don’t skew the story their data is trying to tell.

Keep It Clean

And they're doing this data-gathering exercise while keeping it clean and legal. No shortcuts. They’re respecting privacy like it's the secret ingredient in their signature dish.

What’s Next?

With all this data, Culinary Heights is now ready to whip up better forecasts. They’re predicting how many diners might show up for that fancy five-course meal on a random Tuesday, what dishes will be delivered out of the kitchen, and how many hands they need on deck to deliver the perfect experience.