why we use data
We use data analytics to complement the human aspects of lending like underwriting and relationship management, not to replace them.
It gives us an objective view of credit risk, meaning we can provide indicative pricing rapidly. It also allows us to identify potential changes to our borrowers’ circumstances, which means we can have meaningful conversations to help mitigate the risk of default.
Ultimately, we use data to speed up the funding process, make it more transparent and maintain a long-term relationship with our clients.
Our proprietary model prism
Whilst off-the-shelf ratings from credit reference agencies can be perfectly adequate, they have mostly failed to distinguish between the stronger scores of mid-sized SMEs compared to smaller businesses. This is why ThinCats built its proprietary credit risk model specifically for mid-sized SMEs.
We have trained our model on the core universe of mid-sized firms meaning it is perfectly calibrated for our market and therefore more accurate than any off-the-shelf model.
It also allows us to explain what factors have contributed positively / negatively to a borrower’s score, rather than being a “black box” algorithm that generates a score with no ability to understand the result.
How PRISM works
PRISM has been built using financial information from filed accounts. Measures include gearing, liquidity, income, profitability, growth/decline, balance sheet structure all over a full economic cycle (2007-2019).
This data is then overlaid with credit account and current account information shared by banks and other lenders and aggregated by credit reference agencies. Information available at a given point in time is then trained to predict likelihood of insolvency over the following 12 months.
The model has been extensively tested to predict accurately and consistently across different industry sectors, sizes of business, geographies and over time. This means we can be confident that our pricing is fair and appropriate