Approaches to scheduling have evolved over time, as operators strive for the nirvana of having the right people in the right place at the right time. Each evolution has brought new challenges, but has also brought more predictability and more accuracy, and brought us closer to the perfect schedule.
In this method, managers allocate labour across the week based on their allowed spend (e.g. a percentage of expected revenue). They assign team members to set shifts with little thought to staggering or activity levels by hour, and the resulting day has times of overstaffing when quieter and understaffing during peak trading. Often, too many staff are scheduled at the beginning of the week and, to try and stay on budget, shifts at the end of the week are cut.
While the weekly budget is an important matric, all of this impacts employee and customer happiness. Employees earn less tips and have less to do when it’s quiet, or are run ragged during peak trading. For customers, they struggle to get adequate service they it’s busy, or find staff bored or overbearing during quiet times.
Sales Per Labour Hour (SPLH)
With SPLH, the revenue forecast is divided by a value that has been attributed to each labour hour, to decide how many employees are needed. Different roles have a different SPLH but as a rough example, if the total sales for the week are £4,000 and your target SPLH is £50, then the manager has 80 hours to work with.
Again, SPLH is an important metric (especially for benchmarking performance) but this approach has two main challenges. Firstly, workload differs depending on what’s ordered. £100 could come from fifty £2 soft drinks or a single bottle of champagne, but the time to serve is very different. Secondly, the forecasted spend often doesn’t account for tasks that don’t generate revenue, such as prep work or restocking.
While scheduling based on covers (the number of seats filled on a given shift) brings in a little more science, the main challenge is again that workload differs depending on the guest behaviour.
For example, if guests at one table order soft drinks and starters, while a different table orders a three-course meal and a few rounds of cocktails, the amount of time needed to prepare, serve, clear and tend to each set of guests is different.
Demand (or item-based) forecasting is currently the most scientific approach to scheduling. This method fully harnesses the power of data, with self-learning algorithms using data including historical sales, recent and year-on-year trends, weather forecasts, and events to predict the individual items that will be sold across the day. Managers can tweak this forecast based on their local knowledge, which results in an accurate shape of the day so staff can be scheduled to meet demand. By taking an item-based approach, the amount of time needed to deliver each item (including non-revenue-generating work) is also factored in.
This means managers can confidently get the right people in the right place at the right time, leading to happier employees (through calmer shifts) and happier guests (getting great service). It also leads to better profits by reducing costly over/under-staffing, and maximising sales of second drinks or desserts.
Find out more about demand forecasting and how to bring more science to your scheduling in our complimentary white paper.
Your subscription has been confirmed and you've been added to our newsletter list.