Rusty Secrist, Director of Product Management
In my previous post, I discussed the need for a detailed set of common assumptions and expectations to align accountability across stakeholders. It is important for the budget to incorporate changes, such as new locations or programs, or more competition. The process should include iterative reforecasting and short-term planning throughout the lifecycle of the budget. In this post I will begin exploring the key components of a good budget and how that can be harnessed into a single budgeting process in today’s environment.
Leveraging the past, mindful of the future
We all know creating a budget is not easy. You must fully understand what happened in the past, predict what will happen in the future and then layer on any other planned changes that you know will happen.
A budget must start with reliable and accurate data and an accurate forecast. For your personal budget, it is relatively simple to understand your expenses by simply looking at your historical income and expenses and using that as your framework to create your forecasted income and expenses. However, in retail it is much more complicated. Forecasting with historical data must be thoughtful and deliberate. The impact of unique, non-recurring events must be normalised. Are you planning to remodel the same stores this year? Not likely. If you ran excessive overtime in some stores due to exceptional circumstances, do you need to normalise overtime pay to straight pay going forward? Probably. This is all part of forecasting and working the base forecast from history for the targeted planning cycle.
Forecasting: The foundation of budgeting
Once the special circumstances of the comparable past period are normalised, you can focus on what will change in the period ahead. New programs, merchandising initiatives, labour rate changes, and productivity adjustments, just to name a few, can be layered in. You sold a lot of something last year. Will you this year? There are a lot of variables that could change. Not just the varying impact of a pandemic, but also more typical but still impactful factors such as weather events and changes in competition, among a host of other variables. Ways of shopping could swing types of items bought, too. Supply and demand of items may inflate sales and cause shortages (cough cough, toilet paper). Customers may even start stocking up on more tinned and boxed items instead of fresh and prepared items.
There is so much to consider. Many companies try to manage all those moving pieces in offline systems with each stakeholder participating in the process using their own data. But it does not have to be that way. Your forecasting engine should be reliable and smart enough to detect and adjust accordingly.
Artificial intelligence in forecasting is becoming a must for retail. Forecasting engines have so many algorithms available that take historical data, recent trends and other factors to determine a forecasted value. As those algorithms run and produce a result, that result is then compared to actual values to determine how accurate the algorithm was. This is called machine learning and is critical to producing an accurate forecast that is then used in the budget. A budget is only as good as the forecast that is used in it. If you are not using any sort of machine learning to produce your forecast, how can you produce an accurate budget?
Can labour be accurately quantified?
With the foundation of accurately forecasting your sales—and the associated items, customers, pounds, cases or other volume metrics—an additional critical component in budgeting is understanding the labour required to deliver those sales. Selling an item is the goal, but there is a lot of work that goes into getting that item into place on the shelf so that it can be sold.
A labour model is an indispensable tool to translate the sales (and underlying workload of items, customers, etc.) into the hours required for employees to service those sales according to your business model. Better models use unique standards for each store for greater planning precision. The best labour models yield insight beyond total hours for department sales and also help you understand workload by position or labour task so you can associate hours at the appropriate pay rates.
Some labour models and budgeting processes work only at a very high level given the inability of the system to process the detail really required at the department level and at the weekly level. That puts an organisation at a competitive disadvantage for managing the largest controllable expense in the organisation. It doesn’t have to be that way, and you don’t need to be a Top 10 retailer to access powerful system functionality to create a better plan and to manage performance accountability.
So, about that minimum wage
Assigning wage rates to the budgeted workload starts by normalising exceptions from historical data and layering in anticipated changes. But what if the change is significant, like a change in the mix of full-time to part-time associate hours? What if the minimum wage is increased? What if you want to reduce overtime hours or pay a premium for certain work? All of these are great questions, but also things that must be accounted for in your budget. This is where understanding work content at a more detailed level enables you to assign rates modeled from historical experience but shaped by the upcoming realities you know will happen.
Plan, adjust, react and adjust some more
As we saw with the pandemic, some retailers navigated better than others. However, it’s guaranteed that every retailer was impacted. Those that had systems in place that allowed leaders to quickly reforecast, replan, adjust labour spend, and manage accordingly were at a competitive advantage. Others reacted as best they could but not very quickly or, in some cases, completely.
A robust budgeting system should have all those elements in place to allow you to put your best plan in action and adjust as needed to manage to your goals. So, the question is, does your budgeting process give you a competitive advantage or not?