AutoVue User Acceptance Testing
By Angus Graham on Jul 27, 2010
Picture this... you've got AutoVue set up on your production environment. You thoroughly examined the architecture diagrams weeks ago, making sure everything would sit in the right place. All the configuration on the servers and application servers has been done. You've tested a few files and scenarios. Everything's working as far as you're concerned. You flip the switch and pat yourself on the back for a job well done. AutoVue has gone live into production!
... and then the user calls start coming in. "My designs aren't displaying the way I expect them to!" "Printing doesn't work the same as I'm used to." The calls keep coming in throughout the day. The users are unhappy. Very unhappy. All the effort you've spent to make their life easier has backfired and they're now spending twice as long to do things today. What could have gone wrong?
You didn't get any users to test the system before going live.
User Acceptance Testing (UAT) is a critical, yet often overlooked, part of the deployment process. The concept is simple: set up a test environment which is identical in architecture to the production environment, install your software, configure it, and then get actual users to test it with real data the same way they would use the system in real life.
UAT is a must-have step in any deployment. Without UAT, users are first trying out AutoVue once it has already gone live. Any problems in terms of configuration, product issues, feature understanding should all be discovered before the system has gone live.
Doesn't Oracle test AutoVue? Why should I do my own testing?
This is a great question. The simple answer is of course AutoVue does its own testing. A huge amount of testing. However this testing goes hand in hand with UAT, it doesn't replace it. There are three big reasons UAT is necessary:
1. Your users may use AutoVue in ways that no one has before
Users are great. They will find productive ways of using a product that no one has done before. At the very least they will use combinations of settings or use features in a particular order that may not be common for anyone else. The AutoVue team tests many different use cases, but your users will undoubtedly have different variations, or even different use cases altogether.
2. Your environment will be different.
The AutoVue team tests several environments, but in the real world there are tons of combinations. Different versions of Java, different load balancers, different proxy servers, different security settings, different browsers, etc., etc. It would be impossible to test every single combination of these. By testing your exact setup with your users, you can discover any potential issues due to your environment and get them fixed before you Go Live.
3. Change is scary.
Anytime users have a new tool to use, there will be resistance, confusion, and frustration. By exposing them to AutoVue before going live, you will be able to answer process questions: "How do I do this or that? Where did my button for printing go?" All these questions are easier to answer during the UAT than in the heat of post-go live. Knowing a group of users has tried AutoVue and agreed that it will work in their business process will do wonders for a successful deployment.
So how do I do UAT?
We have just released an AutoVue UAT Guide, which describes the process of planning an AutoVue UAT. I strongly recommend anyone planning an AutoVue deployment read through it and plan some user testing accordingly. You can find it at Oracle Support under document 1160792.1 (It is truly hot off the press, so if you can't access it yet, keep trying, it should be there in a day or so).