- Supports the fundamental business process or goal
- Works in the target environment
Over Christmas 2013 we saw a classic example of this not being done, or not being done well enough. Myer’s website crashed and was down for about a week during the biggest retail period! How much would that have cost? Estimates see this as costing upwards of $31 Million. But what about the additional costs of the free shipping that Myer has also promised to make up for the “frustration” to shoppers? Myers reportedly said that this would not impact on their revenue overall, but I for one would not like to lose $31 Million from any business.
Discussions reported about the issue stated “latest problem had nothing to do with capacity, or the ability of the website to handle heavy customer traffic. Rather, it had to do with an application “not talking” with the server, therefore causing the web pages to time out.” Which is a classic environmental test. Setting up clear communication lines, understanding the architecture of the site and then proving that the architecture supports the key business processes is what UAT is all about.
Sadly when we leave UAT to the end of the development lifecycle, we leave it to business people and end-users to explore in their own way the processes that they will use. We also give them very little time in which to work. We don’t spend money on key systems like a pre-production environment that emulates the live environment. So what can we do about it?
- Plan to test high-risk aspects of the solution (fundamental architecture communication) early in the lifecycle
- Plan acceptance tests that cover the key functional aspects of the business process
- Plan acceptance tests that cover the likely non-functional attributes required for the solution
- Plan to have both users and trained testers actively involved in UAT
Even if you added all this up for a major project, I am betting that it would not cost close to $31 Million! How much is UAT worth to your project or company?
Post by Sharon Robson