This Week’s Banking API Highlights
- Tide in the Open Banking API – UK – Production category takes the overall title again this week with a CASC score of 9.83, down from 9.84 last week.
- HSBC in the Open Banking API – UK – Open Data category takes the overall title again with a median latency of 95 ms, the same as last week.
- Barclays (Sandbox Auth) in the Open Banking API – UK – Sandbox category takes the overall bottom spot again this week with a CASC score of 5.17, up from 4.74 last week.
- We look into why being fast doesn’t always mean that you have a high-performance banking API.
Something of interest
The great thing about the Open Banking API – UK – Open Data category is that the banks have implemented equivalent endpoints. They are thus directly comparable and there is nowhere to hide in this category. 
HSBC has the fastest latency in the open banking API category. Its median latency is nearly ten faster than Lloyds Bank, which exposes the same six endpoints. As we always say, medians can be misleading. HSBC is doing a lot of caching. And that’s fine. The list of HSBC ATM locations is not going to change much on a minute-by-minute basis (given the rate of bank branch closure, it might change on a week by week basis). But what’s not fine is, is being so inconsistent. 
The modal latency is between 50 and 100 ms, but it’s those outliers on the right between 950 and 6300 ms that are determining the mean and the standard deviation. And a high standard deviation means a low CASC score. 
We can contrast HSBC Get ATMs with Lloyds Get ATMs. HSBC is spikey because although the median is less than a 100 ms, there is a high percentage outliers at 500 ms and beyond. Lloyds, on the other hand, has fewer outliers because it is always slow. It is almost slower than HSBC even HSBC is pulling the data from the backend. This is why you need to be actively monitoring from your banking APIs and, where possible as with these Open Banking API endpoints, comparing your APIs with those of your competitors and peers. Both HSBC and Lloyds have something to learn here. For HSBC, it’s that caching doesn’t always provide your users with the reliable consistent service they need. Could they pull from the backend less often? For Lloyds, it's not that you can be slow and consistency, but are there ways that you optimize your endpoint to provide a better service. Three seconds is an awfully long time to wait for a list of ATMs. 
Banking API Analysis: Overall Performance
Week Ending 3 May 2021
Category
Organization
CASC score
Cryptocurrency Exchanges
Gemini
9.22
Fintech
Square (Sandbox)