Like Cassandra - the Trojan prophetess - you can now use Heron Data to make true predictions about the future. Unlike Cassandra, our predictions are trained and tested against thousands of datapoints and specifically relate to the future of small businesses.
Hey Hunters! Jamie here from Heron Data. Really excited to be launching our new predictions product - Cassandra! Quick background to Heron: 120+ B2B lenders and fintechs (including the likes of Pipe, Capchase, and Ramp) leverage Heron to assess the financial health of applicants for business credit.
Until now this primarily involved calculating true revenue, debt, balance, and profit from bank data. Gleaning companies' historic performance is no easy feat but our most common request was to look past historic performance and into the future!
Since it has very high correlation with loan outcomes, we decided to start with a model to predict NSF (Non Sufficient Fund) events first. You can test Heron with any business bank data format.
Send us bank statements (we’ll parse it for you in real-time as well as detect fraud) or stream your Plaid data, and we’ll parse the transactions, categorize them, recreate the P&L and give you a probability of an NSF event occurring. Our model today works best on businesses in the USA.
How does this help me?
- Lend with confidence: with another risk variable you can better predict loan outcomes and price risk
- Save time: Add a rule to automatically kick-out applicants who are more than X% likely to NSF
- Monitor your portfolio: our predictions update whenever you send us new data - predict negative outcomes in your portfolio and remediate them in real time
What's next? We are working on predicting other company events like revenue and balance. If there's a variable you think we should be predicting, reach out! Likewise if you want to hear which of your portfolio companies Cassandra thinks will NSF in the next 30 days, reach out to sales@herondata.io for a free trial
@olga_neudakh Thanks! NSF in the next 30 days is the first predictive feature we are adding and there are many more in the works. Initially we'll adjust the time window (60 / 90 / 180 days) and then add in other predictions e.g. balance, revenue, etc.
If there's anything you think we should be building let me know!
Heron Data