How it worked: Defining challenges

 

The central data sharing mechanism around which Data Pitch was built was the open innovation challenge answered by a competitive call. Consequently, the first task was to define the rules for that competition – who could set a challenge? What kind of challenges would they be? Who would be eligible to answer the challenge? How would they be selected? What timescales would this take?

Challenges were defined by public sector and private sector organisations from across Europe, who shared their previously closed data for the purpose of enabling skilled, innovative companies to provide solutions to those challenges. In Data Pitch terminology, the companies who shared the data were known as data providers, and the companies that aimed to use that data to solve challenges were known as solution providers.

Data providers received support and service in the process of data sharing, and a time bound licence to the results the solution providers produced. Data Pitch worked closely with them to help them understand the benefits and risks associated with data sharing. This was partly due to the fact that many of the data holders did not have much previous knowledge of sharing data, and therefore needed to establish its value across the organisation before things could move ahead. Typically starting from one contact in a single department, the idea of Data Pitch, as well as the terms of the programme, had to be communicated through multiple departments in order to secure agreement from a business strategy, technical, and legal perspective.

The discussions involved in both defining the challenge and selecting appropriate data to address it, required significant resources. This was because the more complex the organisation was, the more different departments needed to be involved. For example, the legal department would assess whether and which data could be shared legally; a technical or data department would assess which data was available or could be shared technically, and what processing would be required before it could be shared; and a business or strategy decision-maker would need to sign off the engagement in the process in general, which would be dependent on the potential benefit gained from this engagement, the required resources, and the challenge it addressed. Assembling a team and having established communication across all of these departments proved to be crucial for success. This process became easier the more focused the initial contact was, the more decision-making power the internal stakeholders had, and the more experience organisations had with either innovation with data, or open innovation.

Onboarding a data holder and supporting them through this whole process took a dedicated team within Data Pitch up to one year. It was not always successful:

Some organisations, where discussions had been active and fruitful for months, had to withdraw from the process. This was for a variety of reasons, including reorganisations, a focus on ‘business as usual’, or a risk averse corporate legal culture.

As tempting as it was for a data holder to simply release data sets and set a challenge such as ‘Tell us what the value is in the data set’ or alternatively, to define a challenge and then later decide which data sets would be the most useful for answering it, it was crucial for compliance with data protection regulations that the two informed each other.26

As the data holders had to agree to share their data with organisations who would address their challenges before those organisations had been identified, Data Pitch developed a bilateral, asynchronous contract, or in fact, two contracts. The terms of the data provider contract mirrored those of the contract with the solution provider, but were executed before the launch of the open call. The second contract was executed after the selection process was closed.

During the process of recruiting data providers, two things became apparent: many organisations did not have the resources to work with more than one or two startups at a time; at the same time, Data Pitch would have the capacity to support more data sharing innovation. To enable this, we introduced ‘sectoral challenges’, in which Data Pitch oversaw the definition of important industry challenges and invited entrepreneurs to solve them, using shared data they sourced themselves. This enabled Data Pitch to support a large number of innovation projects covering a much wider range of application contexts.

Recommendation:

Data sharing is an area that many parts of an organisation may feel they have a considerable stake in. Representatives of the technical, legal and business sides should all be involved in the discussion as early as possible. This will allow them to take an equal part in the conversation and framing of the data sharing relationship. Prioritising one aspect over the other may lead to increased risk aversion from the other parts of the organisation.

 

A Data Pitch challenge and associated dataset

Challenge identifier: DPC3-2018

Proposed by Greiner Packaging International GmbH (GPI). GPI manufactures and markets plastic packaging solutions for food and nonfood industries.

Background

Sensor data and other Internet of Things (IoT) technologies used in industrial processes (known as the Industrial Internet of Things (IIoT)) are providing manufacturers with increased opportunities to optimise their operations and business processes, as well as engaging with their customers.

The smart factory not only represents a step forward from traditional automation, but also rests on a fully connected and flexible system—one that can use a constant stream of data to learn and adapt to new demands.

Manufacturers who are able to access such insights are able to optimise business and manufacturing processes better than ever before. The market for big data manufacturing software is estimated to be one of the largest opportunities in any industrial category.

Description

GPI has recently invested in extensive sensors across 3 manufacturing plants. We are now looking for ways to utilise this data, along with our existing data, to best support and enhance our business. We are looking to develop applications which span across traditional boundaries of manufacturing, logistics and supply chain (perhaps even sales), providing data-informed solutions to better coordinate these processes, make them more efficient, and platforms supporting these new processes.

We are particularly interested in solutions that:

  • Define new relationships between data to provide new insights and understanding;
  • Discover new business opportunities and improve production efficiency;
  • Help ensure the integrity of interactions within supply chains;
  • Integrate data of different modalities (sensors, acoustic data, historical records, thermal maps) to produce useful products and services;
  • Use data to predict maintenance needs and create more efficient servicing and repair services; and
  • Predict and help optimise consumption and stock level, including in multi-country operational scenarios
  • Capture and interpret data to produce answers to commonly asked questions and reduce human intervention;
  • Are able to integrate poor, inconsistent or fuzzy data or information and provide interfaces that communicate key findings and effectively engage users.
  • Failure cases of machines
  • Maintenance and usage history

More detailed information about the data can be found in our data catalogue.

Expected outcomes

Examples of outcomes include but are not limited to:

  • Prediction algorithms that help decrease total stock holdings and lost sales
  • Supply chain optimisation algorithms
  • Algorithms and applications that integrate different sources of data in interesting and novel ways
  • Insights across business functions such as marketing, operations, product and development, sales, etc.
  • Repeatable analytic processes that accelerate the adoption of analytics
  • Ability to gain a better understanding of what data is currently not collected
  • Ability to develop benchmarks that over time contribute to the optimisation of future decisions
  • Waste and lead to better carbon footprints
  • Reporting, analytics and visualisation tools that help to:
    • Absorb information in new and more constructive ways
    • Visualise relationships and patterns between operational and business activities
    • Manipulate and interact directly with data
    • Allow other stakeholders to engage with the data

Expected impacts

  • Ability to make better informed decisions e.g. strategies, recommendations
  • Ability to discover hidden insights e.g. anomalies forensics, patterns, trends
  • Facility for automating business processes e.g. complex events, translation
  • Performance payoffs
  • New business processes
  • Improved decision making