What scoring technique does your solution use?
Our scoring solution utilises our proprietary Mobile App, CredoApp, and was developed to enable the launch of Digital Scorecard product. The app is used during the loan application process and will provide CredoLab with a snapshot of the customer’s digital footprint. In combination with historical client credit data, this enables the development of robust Big Data-based highly predictive scorecard which the client can use for their credit decisioning process.
What data does your solution require to produce a score?
We only require the individual to download our CredoApp application to submit their digital footprint and the score will be provided directly to the client via our web based platform. There is minimal IT integration required in our plug and play solution. The client also has the option to “white-label” our application in addition to integrating the technology into their own application.
How accurate is the score in predicting default rates?
Extremely accurate. Typically an organisation will run their risk models at a Gini coefficient of around 0.3-0.35 and in certain situations slightly higher. Our developed models for clients are significantly higher than this.
How do other systems (such as online loan applications) interface with your credit scoring solution?
There are two main phases of CredoLab solution integration:
Pilot phase. At this phase you don't receive scoring values. CredoLab solution just collects data for further scorecard creation.
Scoring phase. At this phase a scorecard is created and CredoLab solution calculates scoring values for each client.
CredoLab solution contains two main parts:
Software for Android smartphones called CredoApp. CredoApp's responsibilities:
• Collect client data from mobile device
• Upload data to the web service
Web service called CredoLab scoring website. This is subscription based solution. Website's main responsibilities:
• Store collected data from mobile devices
• Display data collection process state
• Perform credit scoring*
*this feature isn't available at Pilot phase
Display statistical information
Step-by-step scoring process:
• During loan application fill-in, sales agent should register a data collection request entity at the CredoLab website. To do that he/she just need to fill applicant's phone number and loan application number. Note: potential borrower won't be able to upload mobile data if his/her phone number wasn't registered at CredoLab site previously.
• Borrower receives SMS with the link to Credo app at Google Play store.
Client must click on the link and agrees to install CredoApp.
• Once installed, Client should read and accept the consent. After accepting the consent client have to click "Proceed with credit analysis" button, to initiate data collection process.
• CredoApp collects and upload client's phone data to CredoLab website. CredoLab will calculate score based on that data.
• CredoLab will provide the score value by bank employee request.
Note: Steps 7, 8, are available only at 'Scoring Phase' and absent during 'Pilot Phase'.
You need to register an accounts for sales agents. There are two options:
Create user accounts one-by-one or in a bunch via import users functionality. Both options are available at User Management Page
Sales agent should create a data collection request entity for each loan application. This action could be performed at Mobile Data Collection page
You can review statistical data at Dashboard page.
How long does it take to deploy your solution? (Model development, training, localization, etc.)
We can be up and running in pilot phase in less than two weeks. With it being a plug and play solution there is no need for long IT integration projects.
How quickly does your solution produce a credit score for a new applicant?
We are currently producing a score in under two minutes for our clients.
Can the solution be run in trial mode alongside existing credit scoring systems?
Yes. During the pilot phase and subsequent model development a client will be able to assess and validate the scorecard before putting it into production. We have several integration methods and the one we will use will be entirely up to the client.