During the weekend of 25-26th September we all witnessed unprecedented scenes at the UK petrol stations. The week that followed has not offered much improvement with a considerable share of petrol stations reporting lack of fuel and massive queues forming. The Fuel Crisis, as this situation was termed, has been widely covered by virtually all media outlets (e.g. BBC, Reuters etc.).
A shortage of HGV drivers is commonly mentioned as an underlying root cause but the recent fuel shortage itself was caused by the panic buying that followed BP’s announcement about their forecourts lacking fuel (BBC and Reuters). A team of analysts at Atheon Analytics combined a number of data sources to shed more light on how panic buying unfolded across various retail channels; we hope this will provide an interesting data-driven perspective of consumer behaviour.
The aim of this short piece is to present insights which describe both the build-up and the crisis itself. Through a unique combination of traffic data from NTIS, anonymised ANPR sensors and fuel sales data from multiple sources, we are able to paint a historical picture along with a detailed account of the relevant two weeks in September.
Figure 1. Visitors count at motorway service operators over time.
Figure 2. Supermarkets fuel volume sold (01/01/2020 to 03/10/2021).
The figures above demonstrate the count of visitors to motorway service operators (“MSOs”) and supermarket fuel sales volume since 1st January 2020. The two indicators follow very similar patterns. We clearly see dips for the national lockdowns and the slow recoveries during times with a substantial level of restrictions up until more relaxed periods, such as the summers of 2020 and 2021. It is noteworthy that during the summer 2021, both traffic level and fuel sales have returned, more or less, to pre-pandemic levels. Despite most Brits choosing to spend holidays within the country, we do not see any out of ordinary patterns in road traffic or fuel demand.
However, everything changed substantially following BP’s announcement on the 23rd September about the closure of a number of its forecourt outlets. An approximate doubling of fuel sales can be seen in the overall trendline but a closer look reveals a more nuanced behavioural pattern from fuel customers.
Figure 3. September supermarket fuel volume sold by fuel type (01/09/2021 to 03/10/2021).
Following the announcement, we notice that supermarket fuel sales almost doubled on the 24th September (see Figure 3). The spike is of equal magnitude for both diesel and unleaded fuel. However, while sales of regular petrol and diesel experience a slump on the next day, premium unleaded fuel sales continue to rise into the weekend more than doubling the volume sold.
Figure 4. September service station fuel volume sold (01/09/2021 to 03/10/2021). Top lines indicate private fuel volumes while the bottom line shows fuel sold to lorries.
In Figure 4, we observe fuel sales at MSOs. The peak there begins on 25th September which is when supermarket sales started to level off due to the decreased stock levels and increased queues. Similar to supermarkets, the run on MSOs begins with regular fuel and is followed by the premium type. Interestingly, the volume of premium fuel sold returns to usual levels much quicker than regular fuel. At its peak, diesel sales quadrupled while unleaded sales more than tripled compared to average. What is more, fuel sold to lorries also experiences a peak but it only starts four days after private consumer peak and lasts only during the working week; returning to roughly normal levels for the weekend.
This account paints a fairly accurate picture of how the growing desperation modifies price tolerance. Being unable to source the regular fuel, customers readily switched to the premium one (even when the fuel was unsuitable for their car type). And once local supermarket petrol filling stations started lacking fuel, customers again sought it out at usually avoided service stations accepting the extra mark-ups.
This short analysis presents a detailed picture of how panic buying of fuel unfolded across various retail channels. Our conclusion is that the buying behaviour was triggered by BP's announcement on 23rd September about driver shortages and their temporary closure of a handful of petrol stations (September's announcement). Interestingly, BP has published an almost identical announcement two months earlier (July’s announcement), which had no effect whatsoever on customer behaviour despite very similar wording. The only material change that we can observe between is that the reported HGV driver shortfall has increased from 60,000 to 100,000 between the publishing of those two reports. It seems that other factors - perhaps the political landscape, a continuing interest in supply chain fragility or additional media attention - conspired to ignite fear in the mind of fuel buyers. What is clear is that once panic buying starts it is difficult to contain and a tipping point is rapidly reached; at that point the effort involved in sourcing fuel, and the price paid for it, become inconsequential to would-be customers.
Our mission at Atheon Analytics is to humanise data so that people can make complex decisions with confidence. Our focus is applied to grocery supply chains, helping suppliers to supermarkets and convenience stores to make better decisions - ensuring that they get the right product to the right place at the right time so that shelves are stocked and shoppers can purchase products at their best. For more information visit www.skutrak.com/solutions.
