3 ms·

I'm not sure that the total amount of money you can make by providing such a platform to customers is greater than the amount of money it will cost to build it.

I'm not sure that the total amount of money you can make by providing such a platform to customers is greater than the amount of money it will cost to build it. The main problem from a cost perspective is that there isn't a whole lot of engineering leverage to bring to bear. Every integration is custom, which means that the number of integrations built scales roughly linearly vs the number of people you have building them. That said, Yodlee has already spent an enormous amount of money doing it. (And, from what I can tell, is de-emphasizing providing this sort of service as a component of their revenue mix.) Yodlee can price the service at a margin above their marginal cost. (i.e., they have paid to build the integrations, now they just need to operate the servers & do maintenance). I doubt very much that any startup can compete on price. It would be a layup to beat Yodlee in terms of developer friendliness (even caring the tiniest amount would do it), but delivering the functionality is IMO probably not going to happen. (Note: I have written a Yodlee integration from scratch and am working on a different (large) project using it now.)

steveg 15y ago
Valid points. There is no question that competing on account coverage is nearly impossible since Yodlee/CashEdge claim to support 10-12k interfaces. While I do agree that building out really wide coverage initially would be extremely capital-intensive, what about just building out the top 20, 50, 100 institutions? I would imagine that accounts have a very long tail, but supporting the top banks would yield the majority of accounts. Of course with aggregation, limited bank support can ruin your value proposition. But what if you build-on-demand? I mean, you build out support for the top 50 banks, then rapidly build out new ones as they are demanded? Still costs a lot of money, but if you can get revenue to roll in then it's more manageable. A lot of online bank UI's offer the same data options (download CSV, MS Money, Quicken, etc). I really see 3 big activities that can be abstracted, then customized based on URL structure and UI: 1. Login and security verification 2. Gather account information (nicknames, balance, limits, interest, etc) 3. Gather account transactions (download statements in one of the provided formats) Many of these formats even include categorization text or SIC numbers. Normalizing the data would be a pain the ass, but could get better as you have more raw bank data to analyze. Of course, there will be tons of outliers, but a lot of banks could be supported with slight modifications. Just feel like this whole aggregation space is ripe for some disruption.
- joncooper 15y ago
  I would run as far and fast away from this idea as possible. Implementing this is a horrendous sucking morass of death and the best available evidence (Yodlee) suggests that it's not an especially profitable service to provide.

steveg 15y ago

Valid points. There is no question that competing on account coverage is nearly impossible since Yodlee/CashEdge claim to support 10-12k interfaces. While I do agree that building out really wide coverage initially would be extremely capital-intensive, what about just building out the top 20, 50, 100 institutions? I would imagine that accounts have a very long tail, but supporting the top banks would yield the majority of accounts. Of course with aggregation, limited bank support can ruin your value proposition. But what if you build-on-demand? I mean, you build out support for the top 50 banks, then rapidly build out new ones as they are demanded? Still costs a lot of money, but if you can get revenue to roll in then it's more manageable. A lot of online bank UI's offer the same data options (download CSV, MS Money, Quicken, etc). I really see 3 big activities that can be abstracted, then customized based on URL structure and UI: 1. Login and security verification 2. Gather account information (nicknames, balance, limits, interest, etc) 3. Gather account transactions (download statements in one of the provided formats) Many of these formats even include categorization text or SIC numbers. Normalizing the data would be a pain the ass, but could get better as you have more raw bank data to analyze. Of course, there will be tons of outliers, but a lot of banks could be supported with slight modifications. Just feel like this whole aggregation space is ripe for some disruption.

joncooper 15y ago
I would run as far and fast away from this idea as possible. Implementing this is a horrendous sucking morass of death and the best available evidence (Yodlee) suggests that it's not an especially profitable service to provide.