
Advanced modelling techniques with instant credit decisioning.
- Fully automated, digital lending processes with instant credit decisioning
- Proprietary credit models using advanced machine-learning methodology
- Robust analytics using alternative data to broaden to thin-file customers
- Rapid cycle testing with deployment of multiple lending strategies
Digital lending process.
Digital loan application only takes three to five minutes to complete, and can be submitted via the Internet, Mobile App and TymeKiosks
Watch how our South African customers apply for a personal loan
Advanced analytics underpins robust lending.
Tyme’s lending strategy is driven by using proprietary credit models, rich alternative data, best-in-class lending architecture and robust risk governance.
ROBUST CREDIT
DECISIONING MODEL
We use advanced modelling techniques with superior forecasting accuracy of loan payback probabilities compared to traditional models. It enables true risk-based pricing that determines customers risk and prices credit on individual level, lowering cost of credit and expanding availability to a broader base. Our models provide transparency on underlying drivers and how they discriminate risk.
RICH DATA &
UNIQUE INSIGHTS
We generate unique, in-depth credit insights from not only traditional data like un-summarised credit bureau data, bank transaction data, but also from alternative micro-data, at the most granular level, from retail, telco and municipal sources. With rich data sources, Tyme has created new model attributes with resulting insights that would otherwise not be possible by just using traditional credit data.
RAPID DEPLOYMENT &
REDUCED OPERATIONAL RISKS
Our best-in-breed lending architecture and decisioning engine inherently allows seamless unrestricted deployment of machine learning models and multiple champion / challenger strategies to optimize lending decision logic. User-friendly interface of our decision engine also enables credit teams to manage lending strategy directly and react quickly to changing market conditions.
TEST & LEARN
APPROACH
We continually test and learn the model in partnership with you – from model validation prior to scale and after. We enable faster modifications by adopting vintage-based views of actual performance versus expectations. This use of advanced analytical tools enable the identification of toxic populations for removal of under-performing credit strategies, as well as opportunities to penetrate deeper.