B&M: why we ditched Excel for AI

One such business is B&M, a long-standing staple of the British high street. From humble beginnings, with its first store opened in the seaside town of Blackpool in 1978, B&M has enjoyed exponential growth. Its reputation for providing four million weekly customers with top-branded products at sensational prices has seen its footprint expand to over 700 stores, employing over 35,000 staff nationwide — and this growth is only set to continue.

B&M is renowned as the UK’s leading variety retailer. However, with a fixed product count of 12,000 products (SKUs), keeping its inventories in shipshape condition is a challenge for its team of 55 merchandisers. In recent years, the team’s only ally in this battle has been Microsoft Excel — with its capabilities, at times, pushed to near breaking point.

“Over the last 10 years, the easiest and most cost-effective way of running our business was Excel — we’re talking basic spreadsheets and look-ups,” says Simon Buckley, B&M’s Head of Merchandising.  

“As our business evolved, we needed more reports and more analysis. We had more SKUs, bigger sales, bigger networks and that made the job a lot harder. 12,000 fixed SKUs means 12,000 linecards, all managed in Excel.”

This ever-increasing reliance on spreadsheets has seen Simon and the team lose valuable hours of productivity to what he describes as “data wrangling and molding.” B&M’s merchandisers and merch controllers were spending anywhere between a day and a day and a half each week just building reports to feed into their linecard spreadsheets — just to get into the position to actually be able to do their jobs and think about what to re-order or what to markdown, for example. 

While spreadsheets remain an effective, powerful tool, they aren’t cast in iron. Rules, formulas, forecasts and pivot tables can all be undone by even the slightest error — sometimes human, sometimes hardware — and these can have a hugely detrimental effect on business performance if they’re not spotted in time.

“On occasions, we have found that too much stock has been ordered, or not enough stock has been ordered and we’ve ended up out of stock — just because someone mismanaged it or over typed a formula,” says Simon.

“Excel, on the whole, is just too fragile for a company as ambitious and as fast-growing as B&M to rely on.”

Embracing change

As B&M’s journey from coastal local to high street stalwart continued, it became increasingly clear to Simon and his team that Excel was rapidly becoming no longer fit for purpose. Something had to change, and had to change soon. 

“It’s been widely documented that we want to get to 900 stores in the UK over the coming years. But we want to do that with the same business model we’ve got now,” explains Simon.

“The SKU count will remain the same and the merchandising team numbers will stay the same. We’re just dealing with bigger numbers, greater cash flow and larger depth. We needed to change to something that was more automated, that would allow our merchandisers to analyze, forward forecast and actually do their jobs.”

But where does a business like B&M start with such a change? How do you tell a large team of merchandisers and planners that the tool they’ve known all their working lives — one that they’ve spent their careers mastering and know inside out — is no longer going to be used in the same way? 

This would be an incredibly tricky transformational project at the best of times, yet alone during a period in which British businesses — in particular those that ply their trade in the retail space — are facing unprecedented levels of disruption and uncertainty.

The COVID-19 pandemic sent shockwaves across the world of retail, with many businesses forced to drastically pivot their operating models in a bid to not just thrive in “the new normal,” but simply survive. This new normal, though, is still yet to materialize — the macro-economic climate continues to disrupt the status quo, both inflation and customer expectations continue to rise and profit margins continue to be squeezed like never before. 

However, in Simon and the team’s eyes, it was a change that would be worth it — providing it could be implemented in such a way that ensured team efficiency remained at the very highest level. 

Software and AI can alleviate a lot of the headaches Simon and his team were feeling. By its very nature, AI is great at making informed predictions and optimizations over vast amounts of complex data and, given the right software, can automate a lot of the more difficult, manual and time-consuming tasks faced by B&M.

Of course, saying that AI and software can help is one thing, but how do you pick the right partner to undertake a journey like this?

“When we first set out to look at a more automated solution and explore areas like AI, there were many, many providers we spoke to that had an off-the-shelf product,” says Simon.

“But we didn’t really want to go down that route because that would mean we would have to completely change how we work.” Making the move towards automation and AI required B&M to take a leap of faith.

Understanding the world of B&M

B&M set about finding a partner that they could trust, and one that understood their world and existing ways of working. 

With retailers across the country feeling the pinch and with teams being asked to do more with less, it was imperative that whatever solution B&M decided on, its implementation would need to keep disruption to a minimum. 

Step forward Peak. After completing a transaction for AI applications via AWS Marketplace, B&M is now leveraging multiple Peak applications to gain a single source of truth on all product performance, in a bid to avoid hours upon hours combing through spreadsheet data row by row, column by column. 

Every business is different. Different data, different people, different processes, different tech stacks and different goals. Peak believes that every business needs its own AI, unique to its own ways of working — which is why Peak applications are 80% ready-built and 20% customizable, powered by the Peak platform and built on best-in-class AWS infrastructure.

“We chose Peak because you saw our business model, you saw how it works, and you knew straightaway how to automate that and make our day-to-day lives easier,” explains Simon.  

“So Peak was easier to implement based on the fact that you were technically re-creating our processes with feedback from all our merchandisers. With Peak, there is very little change for our merchandisers; it was all about how we actually enhance their working lives.”

Right stock, right place, right time: handling inventory complexity with AI

AI gives the B&M team game-changing functionality to effectively manage stock at every single point of its vast, complex network. Peak’s data connectors automatically pull data from across B&M’s business, which is then checked, transformed and fed directly into Peak’s merchandising applications.

This robust data pipeline provides its team with an invaluable ‘single source of truth’ across the entire product portfolio to optimize decision making. It means that B&M can use all-encompassing product linecards to simultaneously monitor things like rates of sale, projected stock levels, seasonality, market trends, prices, margins and more for each individual SKU — all in the same place, with metrics and layouts customized to their world — with the team empowered to step away from hours of manual data pulling.

B&M’s inventory is complex. With 10,000s of products, 20,000 containers on order at one time and hundreds of suppliers located across the globe, ensuring that the re-ordering process is as efficient as possible so that a sale is never missed is crucial.

Peak’s application can predict stockouts or when a product will reach its chosen safety stock level. These factors, combined with lead times, minimal order quantities and supply level agreements, determine when a product needs to be reordered, and how much is needed.

Peak’s AI then simulates every possible combination of how to best fill a certain container with products from a specific supplier. In an ideal world, containers need to be 100% filled for maximum cost efficiency — but this gets difficult when you want every product in that container to be needed in a certain location at a certain time, in order to minimize expensive distribution center holding costs.

Finding the best products and qualities to put in an individual container is a complex optimization. It gets even more complex given that B&M is buying across multiple locations, and ideally wants all of its products to sell out in a similar timeframe. Plus, when you factor in different sized containers, different pack sizes and different product dimensions, no human has time to do this optimally for thousands of SKUs and manage that many stores. Even the most sophisticated spreadsheet would buckle — but Peak instantly gives B&M these decisions, the POs the team needs to be raising, and all of the explainability behind it.

“In order to have a greater amount of stores without expanding the DC network, we need to turn the stock faster — and that’s where Peak comes in,” explains Simon.  

“Peak allows us to run a much leaner business with better stock management, which helps right across the board from stores, staff, distribution center fill and the entire supply chain in general. With Peak’s single source of truth powering our decisions around re-ordering and container filling, it will allow us to literally have a conveyor belt from the time a product leaves the factories in the Far East until it hits the stores — with perfect timing throughout.”

This may have been the first step on B&M’s AI journey, but with the team already beginning to explore opportunities to layer on additional AI capabilities, it doesn’t look like being the last.

In a similar way to how Excel showed the world the art of the possible with spreadsheets, to quote Simon, “Peak has actually opened B&M’s eyes to the art of the possible when it comes to AI.” 

Long may the journey continue.