What Waterstones can teach us about warehouse digital migration

Deliv­er­ing a dig­i­tal migra­tion at the best of times is a try­ing expe­ri­ence. But when you’re upgrad­ing from an out­mod­ed tech­nol­o­gy, it can be even more chal­leng­ing. 

That was the issue fac­ing book retail­er Water­stones in the sum­mer, when the company’s 300 stores shift­ed to a new sys­tem to han­dle its stock from cen­tralised ware­hous­es. It found that the migra­tion was less than smooth, con­tribut­ing to stock short­ages, while frus­trat­ed front­line employ­ees were unable to locate books or order new ones for cus­tomers who want­ed the lat­est best­sellers.

Water­stones did not respond to a request to com­ment for this sto­ry. How­ev­er, the com­pa­ny is believed to have come a crop­per when it upgrad­ed to a new sys­tem that man­ages stock dis­tri­b­u­tion run by Blue Yon­der, an AI-dri­ven sup­ply chain plat­form. Water­stones believes that it can plough through its back­log of orders to get back on an even keel by Sep­tem­ber, when it can take advan­tage of what it calls a “much more sophis­ti­cat­ed pro­gram”.

A spokesper­son for Blue Yon­der said the com­pa­ny had been brought in to help Water­stones man­age stock dis­tri­b­u­tion from its ware­house. “While com­plex­i­ties arose, all par­ties involved in the imple­men­ta­tion are work­ing dili­gent­ly and coop­er­a­tive­ly to resolve the issue as quick­ly as pos­si­ble,” the spokesper­son said.

Learning lessons

For many busi­ness­es, Water­stones’ tale is a cau­tion­ary one – and a sit­u­a­tion they’ll be thank­ful they avoid­ed. “Tech­nol­o­gy is going to become a main­stay and a larg­er piece of the logis­tics sup­ply chain func­tion of any busi­ness now,” says Neil Jor­dan, vice pres­i­dent of logis­tics and sup­ply chains at Prox­i­ma. “But with that comes greater risk.”

Water­stones isn’t the only retail­er to face chal­lenges and hic­cups in efforts to improve its IT stock man­age­ment sys­tems. In 2019, online cloth­ing retail­er Asos decid­ed to update its ware­house tech stack. Algo­rithms designed to tell staff when to pick items for orders mis­fired, mean­ing those orders had to be can­celled when they were incom­plete. In all, the Asos imbroglio cost the com­pa­ny £25 mil­lion in dis­rup­tion costs and lost sales across its Euro­pean mar­kets, with France and Ger­many the worst affect­ed. 

The sit­u­a­tion came as a shock to Asos, just as it has to Water­stones. It also hap­pened at rough­ly the same time of year – an indi­ca­tion that both com­pa­nies chose to try to do their ware­house dig­i­tal upgrades at a time when busi­ness was com­par­a­tive­ly qui­et. 

The fact that the fail­ure came as a sur­prise to both busi­ness­es is itself sur­pris­ing: IT upgrades are dif­fi­cult. “Upgrad­ing a ware­house to become a smart ware­house has many rewards, as well as unique chal­lenges – name­ly, issues relat­ed to dig­i­tal migra­tion,” says Heico Sandee, chief tech­nol­o­gy offi­cer and founder at Smart Robot­ics, a ware­house robot­ics sys­tem.

No one real­ly knows what’s gone wrong at Water­stones and the secrets of Asos’s botched ware­house upgrade are kept with­in the com­pa­ny. Still, there are some clues on com­mon pit­falls that can be avoid­ed. “Main­tain­ing the unchanged con­tent, con­text and struc­ture of the data is absolute­ly crit­i­cal,” says Sandee. “Smart ware­hous­es rely on accu­ra­cy in data to main­tain a smooth flow in oper­a­tions. If dis­turbed, huge cap­i­tal loss­es are cer­tain to fol­low.”

Practice makes perfect

Jor­dan believes that Water­stones’ issue like­ly stemmed from some­thing sim­ple. “If some­thing slight­ly goes wrong in the process­es, and the map­ping has­n’t been done, it can have a sig­nif­i­cant impact,” he says. “There will be some­thing that has just been missed.” He advis­es that prac­tice makes per­fect – and prepa­ra­tion is key. 

When advis­ing busi­ness­es, Jor­dan sug­gests that they look at robust­ly test­ing envi­ron­ments that will be used for real in ware­hous­es to ensure that every sce­nario that could occur in a nor­mal week­day has been tri­alled, and the sys­tem shown to han­dle it. “Com­pa­nies I know who have done this very well take a lot of time and invest a lot of mon­ey in this,” he says. They’ll often run the two sys­tems in par­al­lel – old and new – to ensure that there are no hic­cups that can occur in the new one that could take the whole sys­tem offline. The old sys­tem is left run­ning as a fail­safe to fall back on in the event of a cat­a­stroph­ic error. 

Sandee also thinks this is impor­tant. “A migra­tion with all con­tent, users, and inte­gra­tions mov­ing in mass to the new sys­tem is nev­er a good idea,” he says. Best prac­tices involve plan­ning for a grad­ual, hybrid or rolling migra­tion, where small parts of the entire process and sys­tem are moved onto the new soft­ware bit by bit to ensure noth­ing goes wrong.

Backups are key

The abil­i­ty to go back­wards as well as for­wards when imple­ment­ing a ware­house dig­i­tal migra­tion is essen­tial, cau­tions Sandee. “Eas­ing into relo­cat­ing data is key,” he says. “It doesn’t mat­ter if you’re mov­ing data from one appli­ca­tion to anoth­er or from hard­ware to the cloud, ware­house opti­mi­sa­tion relies on the safe­keep­ing of the orig­i­nal data sets.” 

Not doing so is fool­hardy, he says. “If the new sys­tem fails, be sure you can fall back on your orig­i­nal sys­tem to ensure ware­house process­es keep mov­ing for­ward,” says Sandee. “As long as data can be retrieved, you can roll back with­out any dis­rup­tions.” 

It’s not a fail­ure of a migra­tion if you end up being forced to revert to what you had before to iron out issues. But it is a fail­ure if you’ve pur­sued a migra­tion so whole­heart­ed­ly that there’s no way to reverse a change with­out los­ing all your sys­tems and the good run­ning of your busi­ness.

And be aware: there will be dis­rup­tions. “I’ve nev­er seen a sys­tem launch where there’s been no issues,” says Jor­dan. “There will always be issues. It’s the mag­ni­tude of the issue that caus­es the chal­lenges.” 

It’s about how com­pa­nies respond to those prob­lems, and whether they’re able to swift­ly solve them, that will make the dif­fer­ence between viable stock­ing and shelves stand­ing emp­ty – with the resul­tant effect on the bot­tom line.