Procurement function to squeeze value from supply chain analytics

Supply chain analytics promises savings for procurement. But spending data can be scattered around the organisation. Procurement Lincolnshire, a collaborative local government entity, might show the way

Procurement professionals crave data. It is the lifeblood of their craft. Without information on how much their organisation is spending, on which goods and services, and with which suppliers, it is difficult for them to get to the multi-million pound savings which are often up for grabs. 

Time for data management superheroes to don the spandex and fly to the rescue? If only it were so easy.

Getting a holistic view of corporate spending and supplier data has long been the holy grail of chief procurement officers, but, like the apocryphal goblet, it has proved elusive. 

“Imagine: there are 35 different instances of ERP [enterprise resource planning], it is all coded differently, and incorrectly half of time,” says Duncan Jones, Forrester vice-president principal analyst, sourcing and vendor management. 

“We were promised that if we implemented ERP we would get that view [of spending], and that was a lie. Then they said you would if you got a single instance of an ERP system, and that was a lie. Then it was, ‘get a data warehouse’, and that didn’t work. It is more of the same," he says.

“What people have done for 10 years or so is suck data out of feeder systems into analytics or BI [business intelligence] tools that we call spend analysis, but the problem is the data is so inaccurate. The challenge is not so much using the tool, it is fixing the data,” says Jones.

In most organisations, everyone owns master data and no-one owns master data

James Abery, vice-president, Capgemini

Taking ownership of data

But getting business managers to engage with a project to improve data governance can be hard work, according to James Abery, vice-president at Capgemini

Master data is a massive problem," he says. "Although most organisations are starting to wake up to it, it is not the sexiest topic in the world. In most organisations, everyone owns master data and no-one owns master data. There is no governance sitting in place which says where the buck stops.”

Assuming organisations can get their data into a satisfactory state, the objective is then to aggregate spending into categories, consolidate suppliers and go to market to get savings from additional economies of scale. The potential is certainly there. 

Procurement Lincolnshire, a public sector shared service organisation, hopes to save £35m over five years using this approach (see case study below).

But the limitation with this type of analysis is that it is backward looking, according to Tim Payne, research director at Gartner. “The next stage is being more predictive,” he says.

Read more on supply chain analytics

  • Learn how supply chain analytics can bear revenue and cost rewards
  • Find out how demand-driven supply chain innovation turns on analytics
  • Read tips on manufacturing business intelligence

Turning to predictive analytics

To this end, leading companies are taking demand data from their customers, performing predictive modelling, and then informing their suppliers of future demand more accurately, says Payne.

In sectors where demand changes rapidly, such as retail, predictive analytics of this type can stretch the limits of the available technology, he says.

“As soon as you start bringing in point-of-sale data, you blow out the planning model with 100 million lines of data. At the moment, you are seeing a compromise which says, ‘I have got it roughly right, but quickly enough to take a decision’, rather than getting more precision in the forecasts, but far too late to react," says Payne.

“The idea has now moved from roughly right and fast to precisely right and fast. This is where advances in in-memory analytics hold real promise,” he adds.

Master data can be the biggest challenge in procurement data projects

A natural habitat for data-driven thinkers, procurement offers good opportunities for analytics and business intelligence tools to help the business save money and avoid risk. But spending data can be scattered around the organisation, sometimes with little thought to governance. Therefore, master data can be the biggest challenge in procurement data projects. 

Meanwhile, advances in analytics technologies could allow data teams to help procurement offer suppliers a more accurate picture of future demand in a timely fashion, cutting waste for both buyer and supplier.

Case study: In an austerity age, joined-up spending analysis to save Lincolnshire councils £24m

In 2010, the UK central government announced that its funding of local authorities would be cut by an average of 28% over four years. With councils duty-bound to provide certain services, they have been forced to find new ways to improve efficiency. In some cases, this has meant sharing procurement spending between organisations and getting better deals with suppliers.

But before the public sector benefits from this approach, organisations need the right data at their fingertips.

Procurement Lincolnshire, a collaborative purchasing organisation of local authorities in the county, is using a BI and data warehouse system to bring together data from its eight member authorities, which spend around £600m a year.

“They are all using different finance systems and they have all got different features for extracting data,” says Alex Botten, senior procurement analyst at Procurement Lincolnshire, which was founded in 2009.

“We put together a common specification for the data extraction and we worked with council data departments so they fully understood the model that we needed,” Botten says. The team then used QlikView’s associative learning algorithms to avoid the process of categorising around 37,000 suppliers the councils buy from.

“Within the finance systems there is not a standard way of classifying what a supplier does. The system learns from the transactions descriptions – it works on the statistical likelihood of particular words in the transactions and points to the classification of supplier. The more data we put in there, the better the system becomes at automating that process,” he says.

The QlikView system is supported by a Microsoft SQL Server 2008 database within Lincolnshire County Council’s IT infrastructure. The application holds a static copy of quarterly spend data extracted from the database.

The object is to see where authorities have a common requirement with different suppliers and different contract rates, as well as multiple contracts with the same supplier. 

“We can then run a procurement exercise to have standard specifications, single contracts and agreed contract rates across the authorities,” says Botten.

The process saved £3m in its first year, and Procurement Lincolnshire forecasts it will save £24m over five years.

Image: Thinkstock

Read more on Business intelligence and analytics