Blogs & Articles

Nine reasons why you need a data model

In this blog, we’ll take you through nine reasons why you need a data model and hopefully by the end you’ll have an appreciation as to why a data model is critical to your digital aspirations. In later blogs we’ll take you through the specifics of how the data model will help, and more detail on the specific scenarios below. 

You will have lots of message types that you create a regular basis alongside the provision of lots of files, spreadsheets and regular ad-hoc requests for data with a variety of teams dealing with these. This type of ‘data’ work is likely to be very manual, hard to manage, and not very efficient. It’s also quite difficult to understand the interaction between the messages and where the data comes. 

The data model will enable you to map the content of all of these data requests back to one view (or model) and then map the sources. This will help you understand what data you’re sending out, by any means, and understanding where it is coming from. 

It is likely that you are getting new demands on your data and that your ability to respond quickly to these requests is impeded by the ability to understand where the data comes from. This is further complicated by changes to your underlying systems and platforms and understanding the consequent impact on the data extracts. Usually you’ll find out that there’s been a change only when the data in the messages is wrong because the change was not picked up. 

The data model will enable you to identify how you can quickly generate these requests knowing what data is available.

One of your concerns might be that multiple datasets are being sent out to multiple destinations and whether those are consistent and reconcilable. For example, you would want to ensure that any data you send out to regulators or to back-office processing systems for pricing or risk calculation are consistent. The symptoms you will see will include reconciliation problems downstream.

The data model will enable you to ensure you are using consistent ‘golden sources’ or ‘single sources of the truth’ for all data you are sharing. 

It is likely that you’ll see symptoms of poor data quality. These issues may manifest themselves in queries back from the destination, i.e. the regulator, resulting in iterative processes, significant manual override, and potentially overtime claims. The nature many of these requests is last-minute, time pressurised, therefore any data errors will put staff and processes under significant stress.

The data model will enable you to manage the quality and governance of all the data fields being transmitted at source. This will result in fewer errors, fewer reconciliation problems, more trust from regulators and other third parties, reduction in overtime costs, and improvement in overall staff satisfaction.

Mistakes resulting from data quality are likely to increase the risk to your business. This could include miscalculation of tax, allocation of funds, cash flow forecasts, new business planning, which if they are entered and translated incorrectly could result in fines and potentially damage your reputation. 

The data model will enable you to define the business rules, acceptable data tolerances, and validation reducing the chances incorrect data being transmitted and therefore reducing your overall Company risk.

The reputation of your business could be adversely impacted by third parties’ perception of your ability to do business digitally. This is in terms of your ability to respond to new initiatives, your data quality, the efficiency of your data exchanges and the accuracy of your data.

The data model will enable you to assess and respond quickly to new initiatives, whilst enabling you to ensure your data quality and accuracy, thus helping support a positive digital reputation. 

Increasingly regulators, such as the PRA, SEC, and FCA, are asking not just for data values but some evidence of the provenance of data and wanting to understand any transformations of your data. It is likely that your internal audit teams will have exposed some data issues too and they’ll be sitting as actions with your audit committee. 

The data model will help you define clearly your data lineage and provenance of your data from source, with clear definitions, business rules, quality definitions, and standards. 

One of your objectives of timely, accurate, accessible data will be the ability to generate insights. Insights that would include the ability to identify growth opportunities, gaps in the market, risks that need mitigation, and operational efficiency opportunities. During the onset of Covid you might, for example, have wanted to understand the impact of Covid on your business? 

The data model will help you by showing what data you have (in business terms), how you can access it and how accurate it is so that when you ask for insights they are quicker to generate and you have a view of the ‘art of the possible’.

You will need your IT, business and data teams to work together as they discuss requirements, design and build feeds, messages, APIs, platforms / apps / systems. They need to communicate in a technology-neutral way but need specific data entities and attributes to point to. 

The data model makes it easier to show the integration of business processes with data rules, data structures, and the technical implementation of your physical data in a way that everyone can understand.

In this blog we’ve tried to give you some idea of why you need a data model. There is no doubt you’ll be thinking great, but how does it do that? This is the subject of our next blog, or if you can’t wait then please do get in touch and we’ll be happy to help you bring it to life. 


Get in touch

If you need help understanding how to get the most value out of your data, get in touch with our experts by email or give us a call. We’re happy to help!