Walk the Talk – Machine learning and big data for procurement analytics | Sievo


Walk the Talk – Machine learning and big data for procurement analytics

Sammeli Sammalkorpi

Vice President, Customers; Co-Founder

In procurement analytics, there are two clear trends:

  • Data is the new oil.
  • Big data and advanced analytics, such as machine learning, deep learning, and neural networks, definitely create new opportunities.
    The way we can automatically extract insights from vast amounts of data will both unearth new opportunities and move tasks performed by humans to machines.

As these great opportunities exist with a lot of buzz, the big question remains – How to walk the talk?

The market tends to get so excited about opportunities that it forgets about the real challenge. When you think about traditional spend analytics, the challenge is not really around analytical capabilities – SQL, visualization tools and skills are commodities. The challenge is around data acquisition (e.g. how to get valid data out of ERP and market price sources) and cleansing (like how to map different data sources into coherent category structures) than around analytics and visualization.

Even though new technologies do bring great opportunities, the data acquisition and cleansing challenges remain.

Today, applying advanced analytics does bring a certain competitive advantage – for example, neural network-based algorithms can be used to both support in data cleansing and to identify opportunities. However, these capabilities are becoming commodities so fast that they will not provide a sustainable basis for competitive advantage. The same thing applies to external data – there are, and increasingly will be, companies that provide structured data sets around market prices, supplier risks etc. at fairly low prices.

The challenge remains – how to (i) acquire internal data sets, and (ii) combine internal and external data sets to drive value. These challenges will actually be the most difficult to crack since they, by definition, involve internal data sets and therefore cannot be solved by cloud providers independently. And even if source system consolidation (whether ERP-based or procurement suite-based) will drive down the number of sources at one end, the proliferation of new solutions and data sources will imply that there will be the data combination challenge for the foreseeable future.

With this, for procurement organization to create competitive advantage for their companies, they need to master three complementary capabilities:

The companies that were able to create a consistent spend analysis based on internal data have had a great head-start. They are well positioned to really leverage external data sources and advanced analytical technics. For the rest …well, it’s going to be lot of talk, but very little of actual walk.

All articles

Related articles

Sammeli Sammalkorpi

Vice President, Customers; Co-Founder

Big Data

Where the World of Procurement is Heading

Having more data generated in the last 2 years than in history of mankind is an opportunity to recognize how analytics is…

Pasi Viljanen

Head of Procurement Big Data


Procurement Big Data – from Data to Value

Big data is nothing new – we all have heard about it and have some idea of what it means. But when…

Never miss news from Sievo

Get updates

Never miss news from Sievo