Rabobank disputes online banking audit figures

Redirect and Akamai configuration blamed.

Netherlands bank Rabobank has shed light on why one of its online banking sites was so poorly ranked against its competitors in a one-month audit released by Compuware earlier this month.

Compuware's study of Australia's banking websites measured response times and availability.

The home page of Rabobank's personal banking site, RaboDirect, came in16th in the analysis, with an anomaly of 16 second response times compared to 0.16 second recorded against Westpac's.

RaboDirect general manager Greg McAweeney said it was not informed prior to the audit; he said Compuware tested the wrong website.

While the audit was being done. Rabobank rebranded its Raboplus product to RaboDirect, and redirected domain queries from raboplus.com.au to rabodirect.com.au.

That had an impact on the bank availability of the website, accessible 98.46 percent of the time. McAweeney said that might have in turn have impacted response times.

"A more transparent study would have benchmarked Raboplus up until May 20 and then Rabodirect thereafter," McAweeney told iTnews.

Compuware said the impact on response times of the redirection was "negligable" or less than a second.

The bank was disappointed that Compuware tested the sites against Microsoft's Internet Explorer 7 web browser. McAweeney said most of Rabobank's customers use version 8 for better performance.

Compuware will widen its browsers in future tests.

The bank said that hosting its service in Europe meant very little to the results because most of the content Australian customers required was delivered by a local Akamai proxy pool, known as a "content delivery network".

"Using the Gomez tools on [Compuware's] site, the response time on rabodirect.com.au was 4.79 seconds from Sydney. From Amsterdam it was 4.9 seconds," McAweeney said. 

"The reference made to foreign banks using overseas hosting is not a valid argument in our case. Around 70 percent of the content on our public and secure sites is not hosted in Europe."

And he questioned response times as a measure of the performance of a site as rich as an online bank: "It would be more beneficial and constructive to benchmark the perceived responsiveness, such as how much of a page is loaded to make it usable rather than stopping the clock when the very last object is loaded."

Rafi Katanasho, service management strategist at Compuware said the discrepencies illustrated the need to provide benchmarks to industry.

He said the results convinced Rabobank to reconfigure its Akamai settings, which "improved the results by nearly 10 seconds".

"The Gomez solution clearly identified the change in response and also the redirection of [Akamai] to a local server, which helped improve the result."

RaboDirect's McAweeney said he was happy to be audited or benchmarked if the methodology was transparent.

"Are there opportunities for us to further optimise performance? Absolutely," he said.