Effective project data analytics requires data; lots of data. There is a split across industry into how this challenge is tackled, divided into various camps. I’ve summarised them into 5 areas but I recognise that it is difficult to capture all the nuances within such a short blog.
1. Closed systems. This is where a few companies are heading, creating their own vaults of data to shape their own models and analysis. We are seeing these starting to proliferate, from waste to project performance, schedules to risk. Each is individual silos, with bespoke T&Cs.
Beliefs: Data is valuable and whoever has the most has an unassailable lead. Innovation is limited to those who own or collate the data.
2. Open systems. Where organisations collaborate, share and pool data for their collective Fully open systems are hard to build because some of the data is sensitive, but data trusts work on behalf of the collective and open up data as far as is practicable, with constraints defined by data owners.
Beliefs: Data is more valuable when integrated. Move quicker together. Improve data pipelines together.
3. Blended approach. This appears to be where government are heading through initiatives such as TIES. They are integrating data from across different projects into the cloud. Anonymising data on the way in, but losing utility in the process. In the case of TIES, the cloud is managed by a main contractor, which will ultimately lead to trust issues as they integrate competitor data. It is possible to reverse engineer anonymised data if you have the knowhow.
Beliefs: Open up data from across projects. Client data is sensitive and must be protected and anonymised. We will seek to extract more data from contracts in the future and suppliers will be obliged to comply.
4. Software and app development companies are amassing large volumes of data. Do you understand what your rights are regarding your data and how they are using it? How are your insights leaking into other projects?
Beliefs: I use my product to acquire large volumes of sensitive data, then sell the insights that I derive from this data to other organisations. I can build out more features, that I can sell. The more data I have, the more valuable my company becomes.
5. Access but not ownership. There are a number of organisations, predominantly project delivery organisations who have access to a vast array of data from client projects such as schedules, cost plans, risks etc. But are they the facilitator for this data or do they own it? This will be hidden deep in contracts, where dispute could be very costly.
Beliefs: I have a lot of data in my possession. Ownership isn’t entirely clear, so let’s derive insights from it so that I can add this as a service to my current portfolio. The more data I have, the better the insights.
If we follow a path of closed systems or privileged access to data do we drive a behaviour of data hoarding and/or data redaction and manipulation? Clients select suppliers who have the biggest data set, data sets get bigger and we close down competition. Everyone wants to keep their own data because it is the only way of developing an advantage.
We inadvertently incentivise a system based on distrust and protection of self interests. This is at odds with Project13 and Gemini principles that sit at the heart of current policy.
The National Data Strategy also encourages opening up data for the public interest. Not for a select company, but the benefit of the collective. I agree that there need to be controls, but is it right that government are the arbiter of what is released? I tackled this head on in 2018 and 2019 when government refused to release data, protecting narrow departmental interests. I won numerous cases with the ICO. I lost one, took it to court and won that too. But it is a costly exercise for everyone. With the release of the strategy, the pendulum has swung further in the public interest. Maybe it is time that we develop a citizen panel model, where we decide collectively what and how data should be accessed?
We have a window of opportunity to establish some principles that we all are happy to work within, for the benefit of all of us rather than the few. We create an environment of collaboration and inspire innovation. I am campaigner against closed models; the direction of travel is fundamentally anti-competitive, driving bad behaviour and flawed.
The Project Data Analytics Task Force is tackling this challenge head on through the data access work stream. Although we are piloting a data trust model in 2 sectors, there is much more that we still need to do. A facility to securely pool data under third party stewardship for the benefit of everyone. A data trust is much more than cloud services; it requires a connected ecosystem to ensure we leverage the value that results from it.
And just for clarity, the fundamental principle is that a data trust is owned and governed by its members; where membership is unconstrained. To enable it to operate effectively, the membership needs to be represented by a board, ensuring that it works in the collective interest.