I’ve cleaned up and open sourced three financial libraries I was using for my company. They are bindings for the APIs of these three companies:
Plaid - A financial data aggregator
Dwolla - A payments platform
Yodlee - A financial data aggregator
Here are the links to them on Github. They’ve all been published on Hex.pm.
How comparable would you say the plaid and yodlee API’s are for transactions and accounts? Did you do some level of abstraction over the two in your system to make them seem like the same source?
We’re currently using MX and the number of account linking issues is staggering. But, we’ve heard the success rate on other platforms isn’t much better. Any insights on how plaid and yodlee perform?
We used both because of the differences in institution coverage and data points supported.
Plaid is a much newer service. They support only banking data and have a finite number of supported fields. Their service is incredible, however. The end user experience for connecting to a financial institution is outstanding. Their documentation and API is also top notch.
Yodlee is the biggest player in the space. They support far more institutions than Plaid, and as a result, far more data points. However, it comes at a price. Their connectivity is less reliable than Plaid, their documentation poorer and API less user-friendly. We used Yodlee because they were the only vendor that supported the student loan servicer data points we required.
Connectivity is fickle with data aggregators. They really depend on two factors: 1) the quality of the provider, and 2) the method of data aggregation.
The quality of the provider is well represented by Plaid vs. Yodlee. Plaid is an SF based startup with a lot of talent running their application. Yodlee has outsourced all its development and support overseas. It shows.
Data aggregators can get data by one of two ways: A) an API provided by the institution, or B) screen scraping. The quality and connectivity will only ever be as good as the method used to connect and parse data from the institution. Screen scraping will always produce a poor user experience, but it’s also the only resource for obtaining data from small institutions.
In the data aggregation game, you’re only as good as your weakest link, which, is often pretty weak. Unfortunately, users don’t understand this and you take the hits for it.