How retailers use school holidays, severe weather warnings and college dates for better forecasting
Accurately predicting demand and footfall for retail locations is complex. Weather, seasonality, marketing (both yours and competitors), nearby events, consumer confidence - there’s a lot to factor in. In this post, we share how leading grocery retailers and as well as a couple of the world’s largest quick serve and coffee retailers incorporate school holidays, academic dates, public holidays, observances and more into their forecasting for efficient and effective staffing and inventory management.
Broadly speaking, events like severe weather or holidays have three major impacts on retailers that they can size and prepare for:
A significant change in demand based on event-driven people movement changes.
A significant surge or decline in footfall, which while less directly impactful on the bottom line, still needs to be carefully forecasted as it will impact staffing as well as the timing of marketing or discount sales.
Sudden changes to demand (increased or decreased), staff availability and safety as well as supply routes due to unscheduled events such as severe weather and natural disasters.
We will work through concrete examples of all three of these event-driven changes in demand and give real-world examples of how prepared companies can maximize demand during event-impacted opportunities, and mitigate the impact and inefficiencies events cause.
How retailers can pre-empt and make the most of school holidays, academic dates, observances and more
One of our retail customers, a major coffee chain with thousands of locations, is a good example of how school holidays and college dates need to be tracked and factored into planning.
Let’s start with school holidays, which have different impacts across stores:
Many stores experience a drop-off in regular peak demand periods (such as mornings) as parents are not driving their regular routes and stopping off for a coffee or meal along the way (either with their kids or post drop-off). They respond to this by scheduling one or two fewer staff (depending on scale of impact) as well as reducing their breakfast inventory.
Meanwhile, stores located on major transport routes experience surges in demand at the beginning and end of school holiday periods, as well as a smaller but significant uplift throughout, as families seek food and drinks on trips. Knowing about these surges in demand enables reliable offerings, fast service and happy customers, as well as the ability to only invest in more staff and stock when this peak will hit.
Stores at key holiday destinations also experience a notable increase in demand, requiring more inventory and more staff to meet demand and ensure a consistent and positive experience to maintain customer loyalty.
The challenge with school holidays is that they vary by district and they vary by substantial time periods across these districts. This causes unnecessary overspend or losses when you don’t have the correct data. We know this firsthand, as we used to offer standard US and UK school holiday dates, but upgraded our offering from state or national school holiday dates to holiday dates by district. This meant in the US, we went from ~250 school holiday dates to more than 55,000 school holiday dates, at the request and for the benefit of our retail customers.
For this coffee chain, and all of our customers who also cater to families and young people, the impact of school holidays can be skewed or compounded by college dates. College sessions and breaks drive the relocation of ~20 million young people in the US per year, while major events such as graduation also drive thousands of parents, friends and family towards colleges. Retailers need to know:
When college is in session (dates vary by institution)
When colleges break
What percentage of students are likely to remain on campus (ie near their stores) based on break intensive classes
Major events such as graduation, homecoming, exams and prom, all of which can impact demand.
Knowing about these events enables you to identify their impact in previous semesters and years, comparing your historical transactions and PredictHQ’s years of verified event information. Whether you compare both data sets or use lists of events, we can help you identify expected impact based on your business type and location. Beginning to understand this impact is key. Once you know their impact, you can set up alerts or ingest our event data directly into your models to inform your forecasts and strategies in advance.
Predicting footfall: How retailers use events such as school holidays, observances and college dates for more accurate predictions
The first section of this post explored how retailers use event data to better anticipate demand for their staffing and inventory planning. This section explores the value of factoring events into your footfall estimates, which inform decision making.
Given events can send footfall surging or through the floor, understanding scheduled events is important as is factoring in live breaking events. We identified the impact of school and college holidays and dates above, which is similar to footfall patterns, so in this section we will explore the impact of public holidays and observances, as well as event clusters.
Public holidays and observances vary by state, city and community. While you may know about the major ones, such as Presidents Day or Hanukkah, there are many smaller public holidays and observances that occur throughout the world and disrupt demand in a range of ways, such as:
Public holidays cause surges in domestic tourism, which causes surges for some locations and stores and distinct drop offs in others.
Observances cause communities to stock up or eat out in celebration or remembrance of key moments.
And both are particularly powerful in compounding the impact for attended events, such as sports and festivals. This is why they are key to understanding event clusters.
Being able to track event clusters is a revelation for retailers. Anyone who has worked in retail has experienced those days when your entire team is very busy with no warning. This is usually due to an event cluster: lower profile events that combine and generate major impact. We call these Demand Surges, because prepared companies can turn these footfall peaks into sales peaks.
Let’s take a look at one in Las Vegas in February 2022:
This is impactful because most businesses prepared only for the Marathon (the high profile event) without realizing that the three expos were driving thousands of people into the city, and the concerts would be creating surges of demand before and after them for retailers and entertainment venues.
These can be particularly impactful when you understand how a city is celebrating an observance or public holiday. For example, Chicago has three separate parades as well as many community events such as fairs and food stalls to celebrate St Patrick’s Day. Here's what an event cluster in early March in 2022 looked like:
Knowing about the events (and their predicted attendance, which PredictHQ forecasts for you) as well as major expos, sports and concerts events happening throughout the city at the same time gives you greater insight into which stores are likely to be impacted, and by how much.
Sudden changes to demand (increased or decreased), staff availability and safety as well as supply routes due to unscheduled events such as severe weather and natural disasters
The third significant kind of impact are unscheduled or non-attendance based events that are high impact events that require rapid, data-driven responses. PredictHQ tracks and enriches severe weather watches, warnings and events to enable retailers to understand the likely location and impact of these events, so they can respond swiftly and accurately by:
Alerting staff at relevant stores
Ensuring they have enough stock in the stores most likely to be impacted to meet any surges in demand or redirecting supply to stores that won’t be affected
Updating supply chains routes
Updating staffing and inventory plans based on the severe weather event’s demand impact pattern
PredictHQ tracks severe weather events as polygons, so it’s easy for your teams and models to identify which stores will be affected. While severe heat or cold impact a city fairly uniformly (and the impact will begin prior to the full blown event as the temperatures change and news alerts people), a severe rainstorm can impact one suburb but not the next one over, and floods can impact one store but not another a few blocks away. Or, as some of our grocery retail customers know, it can decimate demand for the store that’s impacted and direct not only the demand that would have navigated to this impacted store but also drive increased purchases for those still in operation.
For more on how retailers can take control of the impact of severe weather, check out this report.
Find a solution that works for your retail network and your planning process
Regardless of whether you want to manage the impact of severe weather, school holidays, academic dates or simply reduce the amount of unexpected demand surges and drops your stores grapple with, you need reliable data to build your plans and strategies on.
This is why PredictHQ exists. While each retail company has their own requirements and goals, the way PredictHQ customers use our data falls into three categories:
1. Visibility: empower your end users such as hotel managers or revenue managers to know about impactful events so they can decide what action to take in response, or to better trust and respond to directions from head office. As the pandemic blew demand patterns to pieces, this is increasingly being viewed as a must-have source of insight.
2. Integration: factor events into existing business intelligence platforms and processes to make your planning or forecasting real-world aware. This is a low-code solution that ensures you are tapping into more of the impact of events, and is more feasible for organizations with more than ten locations as it helps automate smarter decision making. This is our fastest growing approach to using our enriched and verified event data.
3. Machine-learning forecasting: ingest demand intelligence directly into your relevant machine learning models for better informed and more accurate forecasting at scale. Most of our hotel users opt for this option given their scale.
Whatever approach you need, the PredictHQ team is ready to assist you to create your most accurate demand forecasts yet. Get in touch today.