From spreadsheets to autoscheduling: How Gap revolutionised the retail experience by ditching rotas
Can computers really beat humans at rota building? And, more importantly, what’s the effect on employee and customer experience if they do?
These were just some of the topics on the virtual table at , a specialist retail conference bringing together some of Europe’s leading voices on retail and digital transformation.
In a panel session chaired by angel investor, Nick Pendleton, discussion centred around how , one of the world’s leading clothing retailers, went about replacing their old-school Excel rota templates with an automated scheduling solution. Marcus Davidson, Senior Operations Manager at Gap and Rotageek’s Chief Product Officer, David Payne, and Head of Data Science, Daniel Chamberlain went through the transformative journey that took Gap from spreadsheets to autoscheduling.
“We had 4 purposes – simplify the process, improve labour to traffic ratio, improve employee experience and simplify the payroll experience” Marcus Davidson, Senior Operations Manager at Gap.
The issues with Excel or paper rotas
Using Excel spreadsheets to make a rota is very common. Many large retail businesses still create schedules this way and Gap was just one of them until they decided to go digital and automate their scheduling process. Marcus explained that it was taking too much time for teams to create rotas, they had to look at several documents to look at availabilities, holidays, staff requests and how to plan certain tasks.
We already know that Excel and manual schedules are hugely time-consuming - in fact, just switching to digital scheduling usually up to 80% of their time.
More than just a burden on already stretched managers, Excel rotas also aren’t very people-centric. Gap employees didn’t have input into their own hours, they had to rely on WhatsApp for updates, plus basic tasks like booking holidays and asking for much-needed time off was a chore.
Man vs machine - can computers beat humans at rotas?
With the challenge on the table, it was time to test the assumption: seeing if Rotageek’s game-changing autoscheduler could truly create better (and quicker) schedules than the store managers.
As data scientist Chamberlain explained, “humans are good at looking at patterns but in a retail environment, things are always changing. There are two key metrics we look at: one is the number of violations of legal and schedule rules an employer may have and the second is the efficiency to match to demand. Almost universally, when we go to clients, we see some level of violation of their rules.”
Gap initially went into a 4 week pilot across 6 stores before going into a fully-fledged trial. Marcus explained that the pilot showed that they were already good at scheduling and they were matching quite high but when they went to autoscheduling – the solution beat it by 11.3 percentage points matching on the day at about 95%. When digging into the hourly matching of labour to demand, Gap managers got it right 76% of the time, while the Autoscheduler delivered at 90%, a 14.3% improvement rate across the 4 week period.
As expected, using data-driven AI scheduling delivered more dynamic shifts, with shifts now running between 4 and 10 hours and starting every 15 minutes, versus the slightly stiff 4 or 8 hour shifts starting every half hour.
“That got us super excited because it was a push of button after we built some rules.”
Getting your people to trust a rota they didn’t build
With the initial pilot results in, how did Gap get the managers to put their faith in a computer, and give up control in a comprehensive trial and subsequent roll-out?
As Davidson explained, the store managers themselves have a huge interest in customer satisfaction and in increasing basket size. By aligning them, and involving HR in the change management approach, it was easy to collaborate and configure the system to meet their requirements.
And those requirements are critical for success. For an AI tool to work, you need to feed it the right workforce management challenges like rules or European working directive regulations. This offers up one of the reasons that automated scheduling works far better than people at rota-making: it’s almost impossible for a human to balance everything from supply and demand, to fairness and optimisation rules.
“The algorithm will generally be better at fairness as well as optimisation than a human.”
As long as you've got the right rules, and you're capturing and communicating that in the right way, there should be direct wins for managers in terms of their direct interactions with their employees, as well as better outcomes in their store.
Scheduling for success
So, the golden question: did it work and how did Gap measure it during the trial?
Davidson confirmed that there were four measures that Gap looked at: reduced admin time, labour to traffic improvement, and payroll simplification. All of which the automated scheduling solution delivered in spades.
More than that, smart scheduling was a win on theside too, with significant improvements in their NPS (Net Promoter Score), basket size and conversion rate.
“On the employee experience side, that's the biggest shining star.”
“They've got an app. They can see their schedules, they can see when they're working, they can see when their lunches are. They can see who they're working with. They can request holiday when they're at home,” said Davidson.
What really matters when it comes to scheduling is not the technology for its own sake, it’s about return on investment, ease of implementation and measuring the metrics that matter; like whether an employee can take time off for their child’s doctor appointment or how happy customers are when stepping out of their local store.
If you would like to find out more about Rotageek’s for your business, get in touch with our experts - we’d love to see what smart scheduling can do for your people, your customers and your bottom line.