Contractors and clients are signing up to a new initiative to allow project data to be shared and analysed. The implications of this new legal entity could be profound, its founders tell Denise Chevin.
A group of construction companies led by Sir Robert McAlpine is coming together to share data on an anonymised basis in a bid to improve industry efficiency and harness artificial intelligence.
Together, Grant Findlay, director of strategy at Sir Robert McAlpine, and Martin Paver, CEO and founder of Projecting Success, a data analytics expert with a construction background, have been the driving forces in establishing the industry’s first ‘data trust’. To date, six companies have signed up to join and 80 others have expressed interest.
“A data trust is a specific type of data institution where the data is independently stewarded,” explains Findlay. “The vision is to gather data from multiple organisations across the industry, with a view to gaining sufficient volume of data to apply advanced analytics/AI and machine learning to that data. That’s very difficult to do in individual companies, nobody really has enough data themselves.”
The group wants to use the data to benchmark industry performance to boost performance in areas such as quality, productivity, carbon emission reductions and health & safety.
How the data trust was born – Grant’s story
I’m a civil engineer by training and worked in infrastructure, roads and railways before moving into building construction. I’ve been in operational roles in construction for about 30 years.
Throughout my career I’ve always used data to do what I want to do, for design or analysis, but it’s only in recent years that the tech has allowed me to use my data more effectively.
When I met Martin a few years back, we started to talk about data analytics, how we could use data, and use it more often. Because it’s very under-used, it’s something the RICS said last year that most data is only used once in any construction project.
Our ambition is to unlock the value of the data that all the companies are collecting, and to create better insights to improve the industry.
A further purpose is to solve shared problems. “We’ve been using the principle of pooling data to run community ‘hackathons’ across the construction industry and with partners in other organisations, to try to solve challenges over the course of weekend,” says Paver.
Practical examples of what has been achieved in these hackathons include the development of apps to help with site safety. “We used analytics to identify particular features in photographs; we did one around identifying particular kinds of construction materials and we’ve done one around PPE and identifying people coming onto site in the right PPE.
“We also developed an app for a smartphone which allowed a user to report a safety incident by describing it to their phone. Anyone on site could narrate the safety observation directly into their phone, and download the structured data into a file, which could then be further manipulated via workflows and AI,” Paver explains.
Other solutions have used AI to help identify loss-making projects six months in advance, improve schedule estimates and automatically create tailored project communications. By delivering solutions in a couple of days they can accelerate the rate of progress.
The idea of setting up the data trust has been in the making for the last three years, with momentum increasing towards the end of 2020 after key collaboration partners committed funding to address shared productivity-related challenges.
Says Findlay: “The recent government national data strategy has sighted the use of data trusts, so we’re quite hopeful that government is going to be engaged. Martin and I have both been in front of the public accounts committee recently, telling them about the work we’re doing on data analytics. And they were very engaged, very supportive, and they’ve introduced us to a number of government departments and ministers, to try to help accelerate this journey.”
He adds: “There hasn’t been English law in place to allow us to set up a data trust, a trust that manages data, it just doesn’t exist at the moment. It was never my intention, but we’re being quite pioneering, because we’re going to end up defining some of the legal documentation and legislation that allows data trusts to become the norm.”
The new organisation will be run on a not-for-profit basis and funded by membership for the benefit of all. So far it has all been developed on a voluntary basis. Confirmed members include three main contractors and three client organisations. Mace is also on board, but the names of the others cannot be disclosed until the legal paperwork has been completed.
How the data trust was born – Martin’s story
I’m a chartered engineer and a chartered project professional. I’ve been working in project delivery for the last 30 years, I’ve worked on billion-dollar jobs, which I’ve led, and a $10 billion project management office. I’ve seen how we struggle to leverage our collective experience; much of this is codified in data.
I formed my company, Projecting Success, in 2014 and in 2017 I pivoted it to focus on project data analytics. We saw that a big change was coming and an opportunity to move towards data driven project delivery.
In the early days progress was hard won so we founded the project data analytics ‘meet up’ community to share good practice and help to inspire others. It has now grown to about 8,000 people, which Sir Robert McAlpine have kindly helped to sponsor.
In 2018 we held our first hackathon, with our 10th planned for 7-8 August. And in 2020 we launched the data analytics apprenticeship, for anyone, at any level of seniority, who wants to develop lifelong skills in project data analytics. It helps people to prepare for what will be an inevitable and hugely exciting future.
This is the third initiative Findlay and Paver have been involved with that is helping improve how data is being used in the sector. As well as setting up the data trust, they both sit on a new Project Data Analytics Taskforce, a group of around 12 people including academics and government officials which is working to drive up project delivery performance tenfold, from an Oxford University data set of 12,000 projects.
They have also worked closely on setting up a data analytics apprenticeship to school people at all levels in organisations on data analytics. Sir Robert McAlpine has so far put 26 of its staff through the course – and it’s paying dividends.
Says Paver: “In the last hackathon, a couple of apprentices presented a challenge on a cost planning exercise and said they’d like to extract data from a price book, and create a tool to help them to work out cost estimates much more quickly, reducing the time a task took from 40 minutes to just five.”
Paver says that simple developments like this can have an impact, but he expects to see big leaps in productivity when the group is able to look at patterns in the data, and this could have major implications for the way buildings are tendered. “Say you’re building 40 hospitals, and instead of giving one to McAlpine, one to Mace etc, you say ‘we’ll start to pool that dataset, so the next hospital builds on what’s gone before, so we get better and better’.
“So in the future as a client you may say, ‘we’ve got all this other experience from my previous 10 hospitals, what are you going to do with all of that data, so you can deliver the next one more cost effectively, and with greater delivery confidence?’
“Nobody’s asked that question before, because it’s been impossible. So then it’s not just about building the best and cheapest building, it’s saying I’ve now got a range of advanced skills in my team which when combined with a rich data set and a desire to work collaboratively, helps to unlock insights from that data that would otherwise have been invisible. It will save a lot of money as well.”
Findlay adds: “It’s supported by the government’s construction playbook, which talks about more collaboration between clients, contractors and suppliers. Normally you get sent a document that shows the end game, the end model of the hospital, the specs, but we don’t get the experience of the NHS last time they built a hospital.”
The next step for Findlay and Paver is to drill deeper into the productivity challenges and establish membership of the data trust. “More than 80 organisations have approached us about supporting this, from all sectors of the industry, including at least one government department interested in joining the board. We’ve also had support from the Centre for Digital Build Britain and interest from a large insurer, because they can see the value in de-risking projects.
“Within the next 12 months I would hope we have created a good cross-sector representative membership, and settled on a commercial model that allows us to ensure that the data trust is commercially viable.”