ATM testing strategies must evolve for banks to avoid downtime

ATMs have become increasingly complex as technology has evolved, yet banks woefully underestimate the testing required to properly maintain this channel and maximise network availability. As the Flickr group ATM crashes clearly shows, ATM downtime occurs relatively frequently and this can impact upon brand reputation and revenue.

The testing of the ATM itself is just the tip of the iceberg as there are numerous moving parts that sit behind it. As banks migrate from proprietary systems to a multi-vendor Windows-based environment, they must cope with the frequent updates and fixes to the Windows platform as well as updates to the numerous other applications resident on, and connected to, the ATM. Furthermore, banks are faced with new software releases and functionality upgrades at the ATM, which often involve short timeframes. This increased level of change requires frequent and extensive testing, which is no longer possible with outdated testing strategies, particularly those that rely on manual processes. The result is poorly tested products and system updates, which can lead to ATM downtime, lost revenues and a damaged brand reputation.

While banks have implemented various forms of ATM testing procedures for some time, all too often they tested the different elements of the ATM service in isolation, without employing a holistic approach. This old approach, based on a static application software and operating system environment, mainly focused on message testing between the ATM and the host. However, this strategy is no longer fit for purpose and some banks have lost control through applying these principles to an environment which now looks very different.

An ATM alone will require several hundred tests just to confirm operation of its basic functionality. However, banks must consider all of the messaging streams using all different combinations of card types and authorisation systems, as well as interaction between all the applications and the operating system resident on a modern ATM. When considering all of the scenarios multiplied by the different card types and authorisation routes, best practice dictates that a test strategy should encompass between 5,000 and 8,000 test cases.

Banks must now regain control through extensive and unified testing of the various elements as part of a unified, end-to-end and automated testing strategy. This will take account of the changes implied by a more modern technology environment and enable the bank to test more frequently, extensively and cost-effectively, ensuring greater reliability at the ATM for consumers.