Open All Hours?

What works for the Arkwrights – in terms of opening hours rather than double entendres and carnivorous cash registers – simply can’t be replicated by a major high street chain. The running costs mean that it isn’t viable to trade for long hours, hoping to hoover up every last drop of potential custom.

In some cases – particularly shopping centres, where trading hours are often part of the lease – there is no scope to consider what the optimum opening hours should be, but it has traditionally been the case that high street stores have traded standard hours for all stores in all locations. A standard pattern of 9am to 5:30pm might be very easy to apply but is it the most profitable way to trade, particularly as working and shopping patterns change?

Vodafone Retail in the UK operates a network of 300+ high street stores. Although they also have digital and telephone channels for sales and support, the stores offer an important way to attract and serve customers. Running the store network is, however, a complex business. On top of the obvious concerns to make sure that the products and packages are the right ones for customers (which applies to all channels), the stores must be in the right place, have the right look and feel and have the right staff (both in terms of numbers and knowledge/experience).

Hartley McMaster Ltd (HMcM) had worked with Vodafone analysts for many years, helping to collect data and develop models that were used to identify the national total staffing required for stores and the grade mix within this total. The work then allocated the staffing to individual stores, aiming to make an optimal fit to the trading targets for each store within practical constraints such as standard shift length. Vodafone launched a major review of the trading in retail stores – the Right People Right Place programme – and HMcM were asked to look at one specific area of the business that hadn’t been considered before; were the stores trading at the right times?

The picture was a very standard one: those stores with a contractual obligation traded in line with the terms of their lease, some stores (for example at London mainline stations) traded non-standard hours but all others traded 9am to 5:30pm Monday to Saturday, with some also trading on Sunday. There were anecdotal reports of stores full and turning customers away at 5:30pm and of stores with busy periods of trading where customers left the shop (possibly not to return) because queues were so long, while the store stood near empty during others. HMcM were given 12 months EPOS data (till receipts) for all UK stores and asked to see what these said about trading patterns. The EPOS data showed the store, type (eg contract, accessory, repair, etc), value, date and timing of each transaction.

The first question to answer was whether the sales data painted an accurate picture of store activity. For example, if customers liked to browse in stores to do their research before considering their options and returning later to make a purchase, then sales data alone would be misleading when assessing the activity levels of a store. Vodafone were able to provide footfall data (from automatic counters in store entrances) which could be compared to the sales patterns for those stores. The patterns were almost identical for each store – clearly sales activity was aligned almost exactly with store activity and so could be used as the sole measure for assessing the effectiveness of trading hours.

The next step was to create typical trading profiles for each store. This involved two areas for analysis: converting each transaction to a standard value and taking into account seasonality in trading.

For some transactions – the sale of an accessory or a Pay As You Go phone topup, for example – the value is obvious. For others – especially a new contract – the value of the transaction at the point of sale may be small (or zero) but the full-life value much higher. Working with Vodafone customer relations and finance staff, HMcM were able to define a set of methods to assign a standard value to each transaction.

Seasonality in trading could clearly be seen in the data – for instance, December was a particularly busy month for stores (in the build up to Christmas) and January was disproportionately quiet. There were also within month trends visible as well as clear patterns in trading within the week and within the day. The sales data for each store was processed to remove seasonality and to produce a typical trading pattern showing standard values of sales for 30 minute slots across each day of the week. Vodafone then provided a list of the trading hours for each store, allowing the trading patterns to be compared with these.

The final dimension to the analysis was provided by considering the typical running costs for stores, allowing a profitable trading threshold to be defined, expressed in terms allowing it to be compared directly to standard sales. The trading analysis could now be presented to show the profitability of each 30 minute slot for each store. A typical set of results might have looked like those in Figure 1. In this case, the store is (say) one that trades 9am to 5:30pm on Monday to Saturday (but opens at 9:30 on Tuesday to allow for staff training) but does not trade on a Sunday. 

The analysis shows a number of interesting things. First of all, the store is trading profitably across most of its current open hours – clearly good news. However, the first 30 minutes of trading for Tuesday, Wednesday and Friday are below the profitable threshold. Even more strikingly, the store is showing sales activity in the 30 minutes after it officially closes on all days. This can only mean that there are customers left to be served when the store closes – the implication being that this is when customers want to use the store at that location. The recommendation would therefore be that the store should change its trading hours to 9:30am to 6pm for Monday to Friday.

For Sunday trading the analysis looked at the sales patterns for those stores that were already trading this day and used their sales data to forecast what might be expected for other similar stores. In the example above, the indications are that the store should trade profitably on a Sunday, particularly after 11am – opening the store on a Sunday would therefore also be recommended.

The analysis looked at each store and recommended what their trading patterns should be. For a minority, the recommendation for no change, but for most a change – generally to later opening and closing – was proposed. Sunday trading was recommended for most stores, although in a few cases it was recommended that stores currently trading on Sunday should stop doing so.

Implementing the changes involved potential changes to staff contracts and, in a few cases, contractual negotiations with landlords. For many organisations these sorts of hurdles might, at best, have led to recommendations being watered down. Vodafone, however, valued the evidence provided and made virtually all the changes recommended within six months of the analysis being completed.

The results were clear that the shift in trading hours had been very successful

A year later, HMcM worked with Vodafone to review the success of the change programme, repeating the analysis with post-implementation sales data to allow comparison with the forecasts from the original models. Clearly many factors can influence sales performance (including shifts away from high street shopping, the general state of retail spend, the strength of competition and the range of products available) but it was possible to see if the patterns in trading had reduced periods of unprofitable trading and taken advantage of the opportunities.

The results were clear that the shift in trading hours had been very successful. Vodafone were happy that their Right People Right Place programme had been successful. HMcM were happy to have been able to see their analysis validated and to have worked for a client which valued and profited from analysis.

This article was originally published in Volume 2, 2016 of the OR Society’s Impact Magazine.