🎶 Pick n Pay Hypermarket, bring your car you can park it 🎶
If you’re a child of the 1980s like I am, you probably now have that jingle stuck in your head. Sorry! You probably also remember the red Beetle convertible, which swung into a gigantic parking lot outside the country’s latest consumer craze: a warehouse-sized shop that sold everything.
‘Hypermarkets’ haven’t changed much over four decades. They’re still large-format stores that sell a wide variety of products, with an emphasis on value and bulk savings. Back in the 1980s, however, the differentiator for such a store was that a customer could get everything in one place, with the added benefit of ample parking.
But times have changed and trends have largely shifted towards day-to-day convenience: more local stores, smaller ranges, more frequent purchases and, most recently, personalised deliveries to your door.
Still, PnP has 21 Hypermarket stores and their competitor Checkers has 38 Hypers. We estimate that Hypermarkets make up 12% of PnP branded space in South Africa, yet the stores account for only 2% of total spend in our data. (The same comparison is not possible for Checkers because the group does not disclose average store size, most likely for this reason.)
On its own, a lower trading density at a Hypermarket is not a problem. And it’s probably to be expected because of the sheer size of the store and the (presumably) lower average operating costs per square metre.
‘Underperforming hyper stores have long been an Achilles heel for grocery retailers. Analysts have been questioning the future of this format for many years.’
As we showed in a recent note detailing PnP’s performance, Hypermarkets and franchise stores (mostly Family and Express outlets) are currently a drag on the group’s sales growth. In the case of franchise stores, we showed that customers are still spending the same as before, but the change is due to fewer people visiting these stores. We hypothesised that franchise stores may be second-order casualties of the shift to online, since a large proportion of PnP’s franchise stores are located in suburbs and fuel-station forecourts, servicing the same customers who are now ordering on asap!
The news about franchise stores might be new, but underperforming hyper stores have long been an Achilles heel for grocery retailers. Analysts have been questioning the future of this format for many years.
Although the number of PnP Hypermarkets and Checkers Hypers has been stable for a few years, the retailers have been making some changes to their respective formats. Checkers embarked on a refurbishment strategy, converting some Hypers into FreshX stores, and while PnP has been slower in this regard (maybe due to capital constraints) it did close some stores and downsize others.
This report is a unique deep-dive analysis of Checkers Hyper and PnP Hypermarket stores. To our knowledge, no existing report has investigated this segment of the market – certainly, not one based on the actual spending and consumer behaviour of more than 8,000 hyper shoppers.
In the report, we isolate spending at PnP and Checkers hyper stores and evaluate average transaction value, spend per customer and share of spending. For additional context, we also include Makro – another large-format grocery and general merchandise retailer, owned by Massmart/Walmart.
Then we perform a head-to-head analysis of PnP Hypermarkets and Checkers Hypers in 12 locations where both stores are within 5km of each other. Our rationale in using the 5km radius is that customers in these zones are choosing to shop at one particular store over the competitor, and comparing spend between neighbouring hypers is therefore a strong signal of underperforming stores. The radius also controls for any distribution bias, since it’s likely that our data does not span all hypers in South Africa.
Spending habits
Our analysis covers the last three months of 2024, with transaction data from 217,000 unique shoppers who spent at the major grocery retailers. Of these, ~8,000 shopped at a Checkers Hyper or PnP Hypermarket during the period, for a total of ~40,000 transactions at a hyper.
Once we’d isolated the transactions, we scraped the store names and coordinates from the retailers’ websites, then we allocated each transaction to a store location. Given the nature of transaction data, not all transactions refer to a store location, or sometimes the location is ambiguous. Erring on accuracy, we mapped ~50% of hyper transactions to 47 PnP and Checkers stores. The average store had ~325 transactions.
The results are surprising – and always when dealing with this level of detail, nuanced…