Back to Cases


Houdijk Holland

Brief introduction of company

Houdijk Holland is a leading manufacturer in biscuit feeding equipment, handling systems, operating software and components. Their mission is to be ‘First in Biscuit Feeding – Worldwide’, which is led by forty years of expertise and has resulted in big customers like Mondelez, PepsiCo and Pladis.


The challenge the company faced

The challenge for Houdijk was to change their service approach from a reactive model to a more proactive model, where service engineers will support the customer to optimise production, meet quality standards and avoid down time.


The solution we came up with

In order to determine and optimise the OEE of a machine, data needs to be retrieved from the machine. Analysing past data will allow algorithms to recognise patterns and deviations to calculate the performance, availability and quality of the machine in the future. Likewise, algorithms can also use past data to predict when and where parts of the machine will break down. Having this data allows algorithms to make autonomous decisions on what actions to take regarding maintenance and repair, which make the machines ‘intelligent’. Deciding on when and where a machine needs maintenance is one of the many ways operations can be run without human interference and allows the most accurate and efficient decision-making based on data. For Houdijk this means they can increase their after-sales service and improve the OEE of their machines. Effectively the value of their machines would go up significantly, leading to more margin per machine.


Used Services: Elastic Beanstalk and DynamoDB.


Read more about the vision, challenges and reasonings of Houdijk Holland in our interview.

“A machine that predicts when it will break down, orders the part that needs to be replaced and tells you to maintain it before it actually breaks down? Why would anyone say no to that?”  Gosling Putto | R&D Manager at Houdijk Holland.

Industries: Industrial equipment

Regions: Europe

Analyse historical log data

  • Identify maintenance types and unscheduled maintenance events
  • Prioritise specific event combinations
  • Derive Calculated Maintenance Categories

Model maintenance

  • Analyse historical sensor data
  • Refine Calculated Maintenance Categories
  • Define maintenance/breakdown indicators

Predictive maintenance

  • Alert in real time
  • Trigger proactive maintenance
  • Predict remaining useful life

Want to know more?

Contact Maarten de Boo