Hospitality Industry | 4 mins
Better revenue forecasts: planning for profit improvement
by: Pius Steiner, CEO
Start strong to finish stronger
Many owners start by looking at the past. That's a good thing generally, as discovering what happened can help us make good decisions for the future.
But what's the most important number to start with when it comes to running a continuously profitable business in hospo?
Let’s cut straight to the chase. The most important data point to get right from the start is tomorrow’s revenue forecast.
Maybe it seems counter-intuitive? Some would say that a forecast number - essentially something that doesn’t exist yet - can’t be more important for profit than our real staff cost, or our real COGS, for example.
But after three decades of successful business I’ll tell you the real first step: the number to start with is a really accurate revenue forecast. Then we plan all the other numbers around that. Doing it this way means that we can effectively plan our own profit margin ahead of time.
Past actuals are important, of course. We can learn from them. But they’ve already happened, which means you can’t change them. And what really matters for profit is what you can change: the future. You need to understand the past and manage the present … that’s the only way you can control the future.
It's harder to change revenue than to control profit margin
Unfortunately, we can’t just decide to change tomorrow’s revenue. We can influence our ongoing revenue by continuously offering high quality products with great service (my cafe teams use the mantra 'Service, Quality, Speed'). In this way we can change our revenue over time. But we can't simply wish hundreds or thousands of dollars into tomorrow's takings.
What we can actually change more easily is our immediate profits, if we can control how we allocate our costs within our revenue, and ensure there's a margin left over to take home and/or reinvest. To do that, we need accurate revenue forecasting and a robust internal benchmarking system that allows our breakeven point, targets and budgets to be completely dynamic to revenue.
Better revenue forecasting
The thing is, revenue “forecasting” is usually far from accurate. It’s a made-up number. It’s usually a good guess, based on a series of logical factors like:
“what happened yesterday?”
“what happened this time last year?”
“is a large event being hosted in my city right now?”
“did we do a big promo/ marketing recently and hope it's going to start boosting turnover?”
“what does my gut tell me?”
The number we get this way is probably going to be more accurate than simply averaging your last few weeks. Especially when the person guessing tomorrow’s revenue is an industry veteran or has a lot of experience in the site they’re running. We’ll call that an educated guess, and these can be pretty good.
But we are human, and we have human fallibility … including forecast bias. Whether it’s consistently over- or under-estimated, there’s a human bias involved in guessing (even educated guessing). Usually our forecast bias leads us to think we’re going to take more revenue than we actually are. So if we’re calculating with educated guesses, we’d also hope to take our forecast bias into account.
Not to mention the time required to actually sit down, analyse your data and come up with specific day-to-day forecasts, then compare them to your actual outcomes afterwards to determine accuracy and forecast bias.
But the truth remains: the more accurate the forecast, the better we can use dynamic benchmarking and change our profit margin.
If we want to use dynamic benchmarking to increase profit:
We need very high accuracy, at least 90%, because we need to be able to craft a roster around the forecast.
The forecast must predict not only expected daily revenue, but an accurate breakdown of what each service of the future days will look like, so that the roster can be perfectly staffed and consistently productive.
We need that level of accuracy a few weeks into the future, so we can give our people warning of their upcoming rosters.
Likewise, to set a high-profit purchasing budget/ COGS target for our teams:
We need to have an accurate prediction of total daily/weekly revenue.
We need to ensure we have the Goldilocks budget - just the right amount of stock without being over-budget or underprepared.
We need a forecast/projection that continuously adjusts to shifting conditions, to improve our accuracy.
So if humans are fallible and our forecasts aren’t accurate enough to do all of the things above consistently within 90% accuracy, this is the moment where a smart machine can take the lead and give us better data to work with. A smart machine doesn’t let gut feel influence the data. It uses cold, hard facts. And it can detect patterns in your revenue data that you didn’t even know existed. It can analyse your entire history, every scrap of data, in an instant, and understand how factors coincide to influence revenue.
Our forecasting algorithm asks questions, just like the educated guess method. But these questions are more complex, and the answers are calculated based on your entire history and every external factor. The questions analyse your historical data to understand the most likely answers for tomorrow.
“How will the forecast air temperature influence sales volume?”
“What has been happening on site the last few months?”
“How are my sales trending on this particular day of the week?”
“How is revenue distributed throughout the day, service to service?”
“Is it going to be sunny, raining, overcast, etc. and how will that affect revenue?”
“Is it a school holiday/ public holiday and how does that affect revenue?”
All these questions are processed by a machine learning algorithm, allowing it to produce the most likely outcome mathematically based on the many, many different variables that can affect revenue and how these variables have historically influenced the site.
Basically, a machine allows us to get instant, highly accurate forecasts with very little work. So all the work can be put into changing the costs - the numbers that really influence profit.
That’s it: our golden rule. We start with better revenue forecasts and go from there. Rather than adding up our costs and crossing our fingers for a revenue that will pay the bills and leave us a profit at the end, we look at a smart revenue forecast first, and allocate costs within. It’s like slicing up a pie. Profit is just the tastiest slice of that pie.