The complexities in optimising a QSR / fast casual food industry network plan have increased many fold with the huge growth in delivery post Covid.
Pre-Covid, delivery was a minor part of the sales mix and therefore network planning decisions were made around the dine in and take away sales that would come through the door. This meant that networks evolved around the natural trade areas bricks and mortar customers were originating from.
Post-Covid, delivery has increased to represent as much as 20-50% of sales, and now, optimising a delivery footprint is as critical a consideration as optimising a traditional bricks and mortar network for non-delivery customers.
Complicating the issue of planning a delivery network is the fact that different delivery aggregators use different delivery area rules, and all too often, aggregators are less than forthcoming with information about how they set up these delivery areas. Some aggregators use delivery radii that vary around population density, others use travel time limits (resulting in areas that elongate around main roads), while others appear more ‘freeform’ in their delivery area architecture.
This makes the planning and visualisation of delivery coverage difficult at best. As a result, delivery is still being too often ignored when planning site locations, and when optimising networks. This can result in a very ineffective delivery coverage.
In order to get around this, Geotech has spent significant time trying to understand the likely rules different aggregators use when setting delivery boundaries around sites. This has allowed us to introduce into our sales forecasting models and site assessments, estimates of delivery sales and cannibalisation based around coverage.
Even understanding the differences in how delivery sales at a new store ramp up over time (as compared to the traditional restaurant channel sales ramp up) has been an important lesson learnt, and one Geotech can help clients better understand.
As delivery and traditional bricks and mortar channels often possess very different factors driving performance and cannibalisation impacts, this has been a welcome addition for the brands we analyse, bringing increased accuracy and usability to the models we maintain and networks we plan.