top of page
Writer's pictureCarolyn Röhm

Perhaps I should have started with “Kia Ora”

Updated: Dec 7, 2023

Kia Ora,

Thank you for stopping by and taking the time to let us introduce ourselves.


I thought I would start by throwing out a few conundrums that I’ve tackled when I was the client, looking to grow my portfolio, without also growing delinquent accounts and bad debts.


I'd love to know whether some of these resonate with you.


👉 We had an application scorecard, and it needed to be replaced.


👉 We looked at several options:

  • Develop our team internally so that they could develop scorecards moving forwards.

  • Contract a scorecard developer to develop our scorecards for us.

  • Continue using our current scorecards, despite their deteriorating performance.

👉 We knew (in our bones) that this was a project in the making, a B I G project…


👉 We knew we needed data, lots and lots of data.


👉 And we knew our data was not sitting nicely and cleanly in a data warehouse.


👉 We knew our data would need a lot of TLC.


👉 Fortunately, we had a team of analysts that we could call on to help us. Although there were a few questions around some of the data elements that may be required – depending on the project details, we may have had to involve IT for data extracts.


👉 And we knew that our teams were already at capacity – scorecard redevelopments usually only come around every few years, so these project did need to be carefully planned to ensure that BAU didn’t suffer.


👉 Getting the project approved was not guaranteed (even if those of us in the Risk Team thought it ought to be a no-brainer). There were hurdles and committees and approvals required.


👉 We decided to use a vendor to develop our scorecards – to upskill the team would come later.


👉 So, the vendor would develop and deliver the scorecard to us, and we would implement it. That meant IT resources and testing resources, part of the project, a very important part of the project.


👉 And then we looked into monitoring the new scorecard(s) (At this point we still didn’t know whether we would need one scorecard, or several, that would come out during the project initial analysis by the vendor). Reporting was another part of the project, and in my experience, one that gets descoped far too frequently 😔.


👉 Scorecards need to be monitored – if anyone tells you otherwise, I recommend taking everything else they say with more than a grain of salt.


👉 And if those monitoring reports identify or highlight a concern, we needed some sort of plan. How bad was bad? If things were a little dicey, what was the call to action?


Image of Blue and green dense forest
Kia Ora - Image by Carolyn Röhm

Over the years I have worked in several organisations and in all cases, the above is a fair summary of the hurdles we faced when we (the risk team) knew that we needed new scorecards.


✔️ Effective scorecards are one part of a credit risk strategy.


✔️ They’re one very important part of this strategy.


Having effective scorecards in your credit risk assessment arsenal means that your team is in the position of helping the Operations and Underwriting Teams be more effective.


By having great scorecards in place, it means that you can design your credit risk strategies to automatically decision applications that are either very low risk (assuming they meet affordability and regulatory requirements), or very high risk and have an extremely low probability of meeting future payment obligations.


Having effective scorecards in place means you can implement pricing strategies and price for risk, offering your best-risk customers more favourable terms and conditions.


Having effective scorecards in place means that your customers are treated fairly and won’t have a different experience depending on which underwriter is reviewing their application.


Having timely reporting in place means that you’re always across the health of your scorecards, and can plan scorecard redevelopments in advance – so that it isn’t a last-minute scramble for resources.


It shouldn't be this difficult to live in a world where credit risk managers can effectively manage credit risk and resources.


How awesome would it be if scorecards could be developed, implemented and monitored without the need for complex historical data requirements and the commitment of resources?


😔 At Scores4All we understand just how difficult implementing new scorecards can be. We understand the frustration; we’ve worked with clients, who, nine months after deciding to embark on a scorecard development project are still working through the data extract requirements (because not everyone has IT resources readily available).


😔 We’ve worked with clients who know that a scorecard would be able to help them, but the ROI doesn’t meet their internal thresholds, and so the project doesn’t go ahead, and that acts as a limitation to growth (because in order to process more applications, more highly skilled staff are required).


At Scores4All, we’re serious about bringing the gold standard of retail credit risk assessment to all organisations.


🔥 To score your records, connect to our API.


🔥 And let our expert credit risk models score your records. We’ve built these models based on our wide and deep retail consumer credit risk knowledge, built up over years, working across multiple regions, countries, industries, and products.


🔥🔥🔥 The best part? We don’t need your customers’ personal information to score them.


🔥 We use customer attributes, rather than any personal identifying information, and we use this data to score your applications and customise your models – the more you use your models and the more data you submit, the faster your models will adapt to your specific population.


🔥 We use our proprietary data pipelines, feature engineering and machine learning algorithms to continuously refine your models. Then we seamlessly deploy these models; this means that your models are always performing at their best.


And how do you know that your scorecards are performing to expectations?


We send you monthly reports, detailing the number of records scored, their score distribution and how your applicant population is changing over time.


Because we continuously refine and deploy your models, you know that your organisation is making the best possible credit risk assessment on each and every application.


If this sounds like something you would like to know more about, please contact us for a confidential chat.

12 views0 comments

Comments


bottom of page