show image

Sooraj Shah

Contributing Editor

Sooraj Shah is Contributing Editor of New Statesman Tech with a focus on C-level IT leader interviews. He is also a freelance technology journalist.

How Hovis improved demand forecasting with Google BigQuery

When Premier Foods sold a 51 per cent stake in Hovis to the US investment firm The Gores Group back in 2014, Dominic Howson had an opportunity to help set up Hovis as a standalone brand. Despite it being a well-distinguished British brand with a 130-year history, the scope of work was big but Howson was excited about the challenge ahead.

In the beginning of 2017, Howson’s role of looking after information systems expanded to look after Hovis’ supply chain and planning as well, covering demand planning, supply planning, materials management and customer operations. He is now supply chain planning and IS director at Hovis Ltd.

“It’s been an interesting couple of years and I feel like I’ve added strings to my bow [with the addition of the supply chain role]. It’s been a lot of fun,” he says.

In terms of IT, Hovis is essentially a cloud-only business, using all of the big three public cloud services: Google Cloud Platform, Microsoft Azure and Amazon Web Services (AWS). The company also has a HR cloud project with SAP SuccessFactors, and uses a private cloud for its legacy SAP systems.

Howson is a big cloud advocate, and believes the technology can help the organisation to better innovate. Part of that innovation is being better able to use data to improve the way the business runs.

“We’ve been working on this deep neural thinking demand forecasting tool which we call Icaris – it takes feeds in from the weather, school holidays, our promotional information, third party information, all our general ledger information and runs it all in Google’s BigQuery,” Howson explains.

The use of this information can help Hovis to plan how much bread it produces, and can help it to plan how it transports this bread across the country. For example, if it is snowing this has a big impact on sales, while returning to school in September sees an increase in bread sales as parents start packing sandwiches for their children’s lunch.

“This gives us insight into what our demand is going to be at a certain point in the year, and it gives us a much deeper understanding of what we need to do from a logistics planning perspective. Something we’ve been doing in our demand management team is using BigQuery to give us horsepower so we’ve uploaded every single good receipt (GR) transaction we’ve done in the last four years, all the delivery information from the last two years, down to a single load or a packet of muffins,” he says.

This is just some of the data being uploaded to BigQuery, with Hovis hopeful that it can capture more useful data and then query the data quickly. The aim is to make its demand forecasting as accurate as possible – into the high nineties, as Howson says being 100 per cent reliable is impossible.

Currently, one of the hardest points is getting visibility into the national school holiday calendar as different counties have different start dates and holidays, while private schools differ too.

“Getting all of the feeds from all of the councils and geographical areas of when schools are on holiday would be helpful because they have an impact on bread sales and conversely on summer rolls sales, and non-bread products,” Howson explains.

One way that could help Hovis’ bid to get to the high nineties in its demand forecasting accuracy is by getting real-time data from supermarkets’ shelves – an idea Howson would be willing to discuss with retailers.

“If we get feed of that data, we can drive intelligent ordering as well as delivering to store. A big part of that is freshness, as we want our products to be fresh rather than supermarkets holding stock so there is a real scope of opportunity there,” he states.

As well as being a cloud advocate, Howson is keen to not rely too heavily on consultancies.

“We have third parties who support us and aren’t adverse to it but we’ve been very lucky to use the talents of our people internally. Since we’ve become Hovis on our own we’ve invested in newer tools that people want to use and we’ve built our infrastructure and capabilities around our internal staff, giving them the license to take it where want to take it – and we reap the benefits of that,” he states.

The company does get support from a vendor on its process automation journey.

“I bought it as a tactical tool to knit systems together where I didn’t want to spend millions on ERP systems and actually it turned into a strategic tool because where you have got ERP investment or cloud HR investment, you can use process automation to close the loop on them,” he says.

This means certain processes such as the on-boarding process for new employees is far easier and more personalised.

Howson will be hoping this will make new employees as reliable as those he has in his team already, ensuring that the company can continue to innovate using cloud, process automation, and even more data sources.