In 2016 Peter Jackson joined The Pensions Regulator as Head of Data. Here he discusses his challenges, successes and advice in developing a data strategy.

The organisation felt an overwhelming need for change that was driven right from the top. I think that you can only effect data-driven change if you have support from the top, if you have a very senior sponsor in your organisation. Ideally, the CEO is the best person.

The primary way to approach change is to create the vision and articulate that vision to all levels of the business. Start with the board and an elevator pitch, and then go right the way down to the grass roots of the organisation and explain how it will affect people in their day-to-day jobs. By creating that vision, you can put a narrative around change to excite people. You can make them see the benefits for them and the benefits for the business, and how it will grow the business going forward.

Data is absolutely crucial to the future of the organisation. It is a risk-based regulator, in that they have to understand where the risk is and then they regulate to focus on that risk. They are inundated with data. Our main source for regulation is more data about pension funds, about employers. We receive data from across government, directly from our regulating community. You can either sit and look at the wealth of data, or you can actually make some sense of it. And so data is crucial for regulating pensions – we’ve just got to use it in a smart way.

Develop your strategy and hook your strategy to the business objectives. You don’t want your data strategy doing its own thing and not hooking into the primary objectives. In a commercial world that’s easy – it’s either customer acquisition or increasing revenue or perhaps reducing cost. In not-for-profits and the government sector, it’s slightly harder to find your business objectives. But whatever you do, the data strategy has to hang on to the business objectives. In those dark winter days when you’re losing the plot, you can look back and it gives you the point of articulation to say to the board, “That’s why you’re doing this; this data strategy is important because it’s going to give you X and Y for your business objective”.

The biggest data trend for The Pensions Regulator is sharing data across government departments.

John Adams

Everybody seems to understand that data is the new oil and we’ll be a data-driven business, but nobody understands how it’s going to happen. So I think that one way of effecting change is to do incremental change, to go for some quick wins. Do a small data science project that shows value in master data management. Show quick value to get people to buy into the vision.

To gain some quick wins and prove incremental value, we kicked off a couple of data science projects, got some data sets, cleaned them up and pulled out a bunch of that data to do behavioural analytics on our regulator community. The Pensions Regulator has been around for 15 years, so we’ve got 15 years’ worth of deep data for us to go and mine. Just by looking at the data with very, very simple analytics, we could start to understand how different parts of our regulator community behaved depending upon their characteristics. The actionable outcome was that we could identify those pension schemes that were going to struggle to comply with our regulations. So rather than waiting for something to go wrong and then enforcing against them – which is an expensive and laborious process – it’s far better to pick them up early and to help them to comply.  

The biggest data trend for The Pensions Regulator is sharing data across government departments. It seems nonsensical that two or more government departments would ask for the same piece of information from one entity. We ought to be collecting it in one place and sharing across government as the digital economy builds. It reduces the burden on either the employer or the pension scheme that you’re trying to regulate. They don’t have to fill in endless forms; instead, they fill in a form once and we share that data. That to me is the biggest significant trend for the next two or three years.  

One of the problems with sharing data is different formats and different technologies holding that data. Trying to get at the right piece and bringing it in to ingest it in the format that you need is actually quite hard. There are an awful lot of trees in the woods – you don’t want to collect dark data and you don’t want redundant data, so you’ve got to focus on the data that you really need. I sit in policy at The Pensions Regulator, I don’t sit in technology. I sit close to the business so I understand what they need. That’s how I see the trees in the woods, because I’m listening to the business.

In many organisations, the relationship with IT has changed. Because of things like software as a service (SaaS), you can put data in your data warehouse, your data vault, your lake in the cloud. You don’t have to have it on-premise, and you don’t have to have database architects building your database any longer. You can actually say, “This is the data that I want and this is how I want to store it”, and then IT are really just a service supplier. Technology is no longer the barrier; it’s understanding the data that you want – that’s what you need.

When it comes to investing in a new piece of technology, the first thing is to be well networked so that you’re not working in isolation

John Adams

I’ve accepted that data is coming to our operational systems and IT run the operational systems. What I want to do is to suck the data out of those operational systems into an SaaS system that I control where I can create my data lake, wrap master data management around it, create multiple records and allow the business to access that data. The way that we’ve moved forward is separating out a technology layer and a data layer, similar to the way that 15 or 20 years ago people were separating out web design from web development.  

When it comes to investing in a new piece of technology, the first thing is to be well networked so that you’re not working in isolation. A lot of chief data officers are all plugged into each other. We’re all talking so we know what other people are using and how they’re using it and the problems that they’ve faced and the fast step forwards that they’ve managed to achieve.

I’m also a huge believer in proof of concepts and feasibility studies. These are the precursors to investment. I’ll look for a partner who is prepared to do them, because they’re going to be a good partner to work with if you’re going to invest money. It also has the benefit of showing to the business – who may not know technology and who may not understand data fully – what’s possible. It makes it easier to make your business case if you can go to the board with a proof of concept that has worked. Then you can show how you can save hundreds of thousands of pounds or do things more efficiently.