Sezzle Results Presentation Deck
IMPROVED UNDERWRITING WITH NEW RISK MODEL
心
3.2
Dec '21
PROPHET
MODEL
AVERAGE PROPHET SCORE
(PROBABILITY OF DEFAULT) ¹
Decrease of 56%
1.8
Dec '22
1.6
Jan '23
●
A proprietary machine learning model
designed specifically for the holiday season.
Internally developed throughout FY2022 and
deployed during the 2022 holiday period.
The Prophet Model (probability of default
predictor¹) has been a significantly better
barometer in identifying high-risk consumers
compared to FICO and alternative third-party
providers.
Since the introduction of the model, the
Company has been able to better control risk.
RESULT: In 4Q22, Provision for Uncollectible
Accounts as a percent of UMS improved by
232bps compared to 4Q21 (3.5%), reaching a
record-low 1.2%.
¹The Prophet Model predicts the probability of default at the time of underwriting. As a result, it does not correlate one-to-one with reported loss rates.
Source: Sezzle Internal Data.
SEZZLE INC 4Q22 Presentation | 9View entire presentation