What’s next for retail data? Apple Watch and Apple Pay are upon us and everyone is trying to work out the implications for shops and consumer behaviour. Is it a weapon for the high street retailers, with wrist-vibrating alerts for time-based and geo-located coupons luring in shoppers? Or will it improve e-commerce by providing delivery alerts, making the process more convenient?
What data is needed for retailers to stay on top of the coming changes?
As retail vacancies in the UK have started to increase again to about 50,000 empty units (up to 11.7 % for retail and leisure), (LocalDataCompany.com, May 2015), investors and retailers are looking for a much more realistic analysis of the UK retail space. “Our high streets need to shrink by 30%” was the diagnosis by Lord Stuart Rose at a recent AIS independent retail event in Birmingham.
If this forecast sounds alarming, it should be – not just to retailers and investors, but also for local councils. It’s not in anybody’s interest to witness another upheaval of the scale of the Jessops or Woolworths shocking large-scale (and very sudden) closures, with their ensuing mass job losses. To improve forecasting, retailers and academics recently gathered at Oxford’s Said Business School to assess the gaps in the “Quantified High Street” data.
According to the go-to source of data for retail movements, Local Company Data, about 5% of ‘comparison’ goods including fashion, white goods and children’s wear have transited online over the last 24 months, and that trend is accelerating. During the recent BCSC North conference it was reported that 12.5% of current UK retail is online, a figure that is expected to peak at 16% in 2022 (ONS).
An estimated 12% of fashion retail closures are anticipated for 2016 as the leases of some large chains come up for renewal and the chains systematically release the stores back to landlords. Removals of visually appealing fashion stores will have negative implications on the attractiveness of the area, speeding up the decreasing footfall. How much brand recognition will multiples achieve once their shops are removed from the high street and shopping centres? “Out of sight out of mind” the saying goes, so it is highly likely that the brand awareness of the retailers undergoing large-scale closures is going to negatively affect their reach.
As we assess the number of vacancies created by this channel shift, one of the risks to retail forecasting is that our current understanding of ‘vacancies’ is not sufficient. As Professor Leigh Sparks notes, occupancy rates are often imprecise, hiding a high number of charity shops, which should signal distress rather than good news. There are also vacancies occurring due to a mismatch of size as a result of the store trend of larger locations. Often a tenant is faced with excessive business rates, a curse highlighted in Bill Grimsey’s High Street Report (2013) and a problem acknowledged by George Osborne in his Autumn Statement 2014.
One of the findings in Professor Cathy Hart’s study on town centre consumer behaviour was that parking (in)convenience is an important factor in the shift away from the high street. A real-time open data feed of available parking spaces fed to drivers’ mobile phones would go a long way. In fact, investors would appreciate an open data feed of parking spaces in each location indicating whether or not a particular site merits a store visit – a key consideration. A feed of transport data is being developed for London by TfL, but the provision of such data is not so good in other towns and cities. 2014 saw a decade high of new car sales, with the UK registering 2.7m new vehicles. However, councils are not obliged to track car sales or respond with an increase of parking options. Green vehicles account for a growing percentage of the new car purchases, so local councils can no longer say that they are against cars on the grounds of pollution.
High street is still loved
The high street is still in favour, as noted in research by Professor Hart’s research at Loughborough University, whose data gathering program uses iBeacons for ‘mystery shopping’. Through an app, customers are given deals as an incentive to provide specific, real-time and geo-located feedback on the parking, product, atmosphere and cleanliness (amongst other factors) of their high street experience. However, the Local Data Company reported that footfall on UK high streets fell in April and May consecutively. Retailers are anxious to get sensor-based, real-time and exceptionally precise footfall delivered in an anonymous and aggregated open data format. Such data would help to explain whether vacancies are due to footfall issues or, as is often the case, are an outcome of a mismatch of the retail offering to local consumers.
Where people do visit frequently they spend very little, often under £50 per visit. Nielsen Sales data and the BRC Sales Index provide a general understanding of retail on a monthly basis, but this data excludes online retailers who do not have bricks-and-mortar shops like Amazon, eBay, ASOS etc. This data set needs revision as pure-play e-tail has become highly significant across a range of sectors, and excluding it from high street sales figures is under-representing the shift in consumer multichannel spending. It also needs to include smaller retailers and independents, which are missing from the current data and are growing by about 10% year-on-year on the high street as well as online (Local Data Company, 2015).
Big data comes to the rescue
Opening up all the sources of data that shed light on the UK retail (high street as well as online, since multi-channel retail requires a multi-channel data strategy) should be made compulsory for the local authorities, following the example of TfL.
We need to understand which data needs to be open (parking availability, real time high street footfall), which data needs to be ‘shared’ with guaranteed privacy considerations, and which data sets are to remain closed, hosted securely and only made available to insight professionals (credit card spending). Details of those increasingly precise understandings of the nature of the open-closed spectrum of Quantified High Street data are emerging. We were given a clarification of these different data sets at a recent presentation at Open Tech 2015 by its pioneering CEO Gavin Starks, who has created a working matrix for retailers to discuss and join the Quantified High Street project.
Retail in the dark
One reason that investors and retailers are still in the dark is because councils don’t provide a detailed matrix and real-time database of vacancies. Appear Here, Town Teams and other start-ups that are attempting to improve the use of empty retail properties are struggling to get accurate visibility on upcoming (or indeed current) vacancies. We have investigated this issue with Hackney Council, finding that the sole commercial officer was covering over 300 retail properties and was just too busy to deal with database updating. This is a pattern across the UK, and with 1 in 5 shops in the north closed it can’t be a sustainable business model for heavily indebted councils.
Digital data exhaust
New streams of data are coming online, such as the Twitter geo-location analysis, which is creating a new, dynamic and real-time footfall picture (instead of the current static snapshots of footfall from a relatively small sample of locations provided by the BRC Sales Index).
Guy Lansley of UCL has delivered a pilot study for the Consumer Data Research Centre that analyses location and time-based Tweets from London and matches them against the 2011 census, ONS Survey data, Ofcom data, Land Registrar Data and other public data sources. As immigration increases there is fast growth in certain demographics (e.g. men aged 25-35) that are creating new retail opportunities, but these are often only picked up long after the opportunity has passed. The data mix used by the UCL study has delivered a glimpse of what is possible, i.e. a very granular, dynamic, and predictive pattern for London shoppers. As such, it would be able to help retailers to support customers with highly timed and localised service and product offers. As Twitter messages with geo-location skew towards men (women are less inclined to share their location), it may be useful to extend the analysis to other social media streams where the location can be inferred rather than stated.
The retail future is looking a lot more complex than the already multi-layered, multi-channel landscape of today. No retailer alone will be able to develop comprehensive, predictive and sufficiently adaptive datasets to answer all of their questions. Data is the future of retail; collaborating on Quantified High Streets will help us all to be more resilient.