The trick to getting data infrastructure and governance right
Mary Patricia Hall, Head of Regulatory Reporting & Oversight at Insight Investment explains why technological innovation is only useful if you can get company-wide buy in.
Insurance Investor editorposted on Thursday, October 24, 2019
Insurance Investor: How can firms create a solid data infrastructure?
Mary Patricia Hall: One of the key elements when you are faced with a bevy of issues that you feel you can't swim out of is that you start defining responsibility and who owns what.
A lot of people start with the technology and, whilst there is a huge role for technology and you can build all of these great platforms, if you don't have the responsibilities around who owns what so they understand the impact, you can't really build a great data infrastructure.
"If you don't get ownership down then you can have all the technological bits and pieces that you want but it will just cost a lot of money and people won't buy in."
It is an issue that we have as well and we are trying to learn from others who have done this in the past.
It has been my experience that if you don't get the issue of ownership down then you can have all the technological bits and pieces that you want but it will just cost a lot of money and people won't buy into the infrastructure and controls that you need.
II: How can data flow be controlled?
Mary: Technologically speaking, as you are bringing in data, you need to have standard ways of knowing what to do if there are problems with the information and then implementing the correct controls; but having data standards and data governance is a very large part of the puzzle as well.
Whilst you need one standardised solution to bring in all the various items there is also the issue of the transformation of data.
"Having data standards and data governance is a very large part of the puzzle."
Data comes in its raw form but then it gets transformed within an organisation and again one of the questions we are grappling with is how do you know what the information that has just been transformed is.
In regulatory reporting, one of the issues we struggled with was that the old infrastructure was not understanding where pieces of data came from so that you had people taking pieces of data from various sources, it didn’t provide any comfort around understanding what the answers were in your reports as an end user.
II: How are you setting data standards?
Mary: Where we are moving towards and what we see as a fundamental is in having data governance. I realise that this can sound like just another bureaucratic thing but you do need people within that governance group to be influential enough within their areas to buy into what you are trying to do.
I know that with other SMCR (senior managers and certification regime) you are supposed to have one person who is in charge of it, but there is something to be said for starting with a collective decision and then also having one person who is in charge of implementing that collective.
"It can't just be a technological change and has to be an organisation one as well"
However, their role needs to be made clear. Of course this wouldn't work in a 10,000 person organisation and would be more for the medium to smaller sized firms where you can have people be main influencers.
You also would need to have people at the more senior levels, such as the executives, buying into this solution in order to empower the governance.
I have built something similar within regulatory reporting where the governance was about getting the main influencers to make sure that they are happy with the decisions we make and then as head of the regulatory and reporting oversight, I am empowered to make sure that these things happen throughout the organisation and not just with my own team.
"It is all too easy to get excited about the technological change."
This allows you to keep those standards and because they have this specific role it is much easier to do. People can feel that having to do a bunch of various tasks is just part of their day job but actually they shouldn't have to do this if you have the controls put in place before it even gets to the end group.
This is why it can't just be a technological change and has to be an organisational one as well.
It is all too easy to get excited about the technological change with new infrastructure and I have worked with these in both client and regulatory reporting but if you don't have the buy in from different groups and support from influential people who make sure that the values get embedded within the business, it isn't going to work very well. You will be left wondering why it doesn't work.
II: Have you seen this particular trend developing across the market?
Mary: The rest of the industry knows that data is a very key aspect and they are putting their money where their mouths are.
The big houses know this and have their CDOs and perform data science and from my perspective it is more a question for the smaller and mid-sized firms in terms of how they deal with this.
It isn't necessarily about them trying to catch up to the bigger players, but how do they implement a solution where they can feel comfortable that the information that their front office is using is reasonable as well as what they are sending to both the clients and the regulators.
"The industry knows that data is a very key aspect and they are putting their money where their mouths are."
The reason I have taken such an interest in this issue is because so much of the data is going to the wrong regulator and they are getting cranky that the data quality of what is being sent to them is rubbish.
Earlier this year the FCA sent out MarketWatch 59 where they discussed some of the problems with the data that they are receiving. So the regulators are starting to look at this issue as well and this is why it is important.
From a worst case scenario perspective you should want to be able to control the information that you are sending to the regulator but of course that leads you to want to look at everything else.