Source: https://www.etim-international.com/

So, the first step in spend classification is to choose a product classification standard that meets your needs to classify your products.

Benefits of Spend Classification

Why should companies classify product items to standard categories? There are a number of benefits for using product classification standards:

• Standardisation – A standardised product classification can improve the purchasing costs for a company. Standardisation allows the business to have a clear picture of what its buying and from whom these goods and services are sourced. Only then it will be able to make marginal improvements to its spend performance.
• Enabling spend analysis - Without accurate categorisation of historical spend data, no analysis can yield the information necessary to fully optimise sourcing decisions, or to conduct effective negotiations.
• Optimize supplier base – Using a procurement analysis based on a spend classification you will be able to identify the commonality of productgroups across departments, business units and/or suppliers. If you combine this with information on how much of this spend is under contract you will be well prepared with data when you walk into negotiations that can lead to volumediscounts, better payment terms, and more spend under contract. A broad visibility into product groups allows for more strategic sourcing options.

Is that easy? Not really.. keep reading for the main challenges.

Challenges of Spend Classification

As most valuable things in life, Spend Classification doesn’t come easy. There are many challenges in the classification process.

• Quality– The data quality of theproduct items to be classified is often quiet poor which makes it hard to make classifications on. Furthermore, product items might have different sources like multiple ERP systems. In these situations, data should be integrated first before a complete spend classification can be performed.
• Accuracy- Companies have significant amounts of misclassified data that prevents them from effectively managing their spend. General methods to classify spend are to do this manually, or to use a rules-based system. These tasks are labor-intensive and typically result in human failures during the classification process.
• Labor-intensive -  Classifying all different type of productitems manually is highly labor-intensive. More advanced methods like Machine Learning techniques can be extremely helpful in fulfilling such tasks. Based ona labelled training set a model can be generated to make classifications automatically.  

Spend Classification Powered By Machine Learning

Zuiver ICT is developing a Machine Learning model with the goal of automatically classifying product descriptions into groups or subgroups. It allows to perform a highly-valuable Spend Analysis.

We are in the phase where a model is still being trained. If you are interested in improving your product classification, based on a pretrained model, you are most welcome.

Any labelled data available will strengthen the training process resulting in higher accuracy. Therefore, in this phase of development, participation is free in a sense that we do not charge any costs for the outcome of the classification performed. If you are interested, do not hesitate to contact us