Retail 2020: ‘Of robots, shelves and packs’ by James Woudhuysen
Ahead of the launch of our white paper for Retail 2020, it’s our delight to introduce a guest blogger, Professor James Woudhuysen. The esteemed futurist and forecaster shares his thought-provoking account of robots and their use in retail today and beyond.
‘Of robots, shelves and packs’
In all today’s euphoric-apocalyptic hysteria about how the robots are coming, one sector is woefully neglected (1). While commentators wax lyrical about Amazon’s use of Kiva robots in its warehouses, nobody stops to ask why, in the aisles of supermarkets and other shops, machines are not replacing the youthful, often part-time staff who stack shelves.
That’s odd. Research by Broekmeulen, Zelst and others, from the Technical University of Eindhoven, the Netherlands, reminds us that handling stock in-store forms no less than 38 per cent of a shop’s inventory, transport and warehouse-to-store handling costs (2). So why don’t Britain’s retailers do their bit for their country’s perennial problem of productivity, while at the same time saving themselves a lot of cash?
Perhaps the wages of people in uniforms on the floors are so low, they make capital investment in shop automation seem a fool’s errand. But look at the effort that’s today being put into payments systems at the checkout, all laudably aimed at shaving a few seconds off shoppers’ exit times. That effort is enormous; so the contrasting failure to mechanise the stacking of packs remains a puzzle.
In fact the puzzle can be solved. Despite British retailers’ conservative attitude to smart cards and all that, it’s simply much easier to bring IT to in-store financial transactions than it is to apply IT to the noisy, disorganised business of breaking down and disposing of secondary packaging on trolleys, and getting packs on to display cabinets. The point about Kiva is that it moves warehouse shelves around; handling merchandise is too exacting for machines, and, at Amazon warehouses, is done by human beings. Only human beings have what the philosopher Karl Polanyi called tacit knowledge; and such sensory-motor knowledge is much harder to computerise than is high-level reasoning.
Indeed, the non-computerisation of retail stacking is likely to continue for a long time. We must wait until advances in what’s called machine learning allow computers and robots to muster and generalise from – I kid you not – massive, structured and fully captioned Big Data sets of YouTube films of human success stories in shelf stacking (3).
Still, stacking could be improved now – this side of robots attending to retail gondolas. The Eindhoven academics found that increasing the size of the cases in which products are stacked can, per stacked individual product, bring time savings of 24-49 per cent. Stacking lots of case packs of one product at once, rather than repeatedly stacking individual case packs? That can bring time savings of 8-31 per cent. Also, moving items on to shelves by tray, or into shelves with loose-fitting crates in them, can save 12-42 per cent of the time spent stacking. Oh, and there’s probably a lot of excess shelf space lurking behind your products, ready to take multiple case packs as inventory.
It’s time to ask some tough questions about exactly what it is robots can and especially cannot do. And it’s time to ask some tough questions about the pay, processes and packaging around every retailer’s shelves.
Retail 2020 launches in December, delving deeper than a blog to examine the challenges facing retailers, brands and marketers. It depicts the most pertinent trends that will shape tomorrow’s shopping experience, so you can quickly and easily grasp what’s hot, what to look out for and what to avoid, if you’re to attract, delight and ignite future shoppers to act. Register to receive your free copy.
(1) See for example Carl Frey, Michael Osborne and others, Technology at work: the future of innovation and employment, Citigroup, February 2015, on http://www.oxfordmartin.ox.ac.uk/downloads/reports/Citi_GPS_Technology_Work.pdf
(2) Susan van Zelst, Karel van Donselaar, Tom van Woensel, Rob Broekmeulen and Jan Fransoo, Models for store handling: potential for efficiency improvement, on http://w3.tue.nl/fileadmin/tm/Capaciteitsgroepen/OPAC/Zelst_et_al.pdf; also published as ‘Logistics drivers for shelf stacking in grocery retail stores: Potential for efficiency improvement’, International Journal of Production Economics, Volume 121, Issue 2, October 2009.
(3) A detailed account of Kiva’s work, Polanyi and machine learning are in David Autor, ‘Why are there still so many jobs? The history and future of workplace automation’, Journal of Economic Perspectives, Volume 29, Number 3, Summer 2015, on http://pubs.aeaweb.org/doi/pdfplus/10.1257/jep.29.3.3