In some cases, automation can even be an obstacle, drawing attention away from the content and disorienting the user from his task. Automation is carried out by mechanical code, which does not think or know any context, and as such, can very well make mistakes. And it's correcting these mistakes that can take even more time and effort from the user than it would've if he had initially done for himself what the automation did for him. A convenient timesaver turns into an inconvenient frustration.
For instance, a website could automatically assume that a user connecting from the United States wanted the English website. The user glances at the website and finds something interesting, but as he happens to be a German immigrant called Hans, he'll want the German version instead. He'll find and click the language selection, and that takes him to a splash screen that automation had skipped for him earlier. Hans selects the German version from the list and the website is completely different and the interesting thing isn't there anymore. He leaves.
This is really the ultimate question, and this can only be solved case-by-case: Is it good practice to assume the language of the user? Would Hans be less disappointed if he hadn't seen the interesting thing on the English site first, but gone straight to the German version if there was no automation? Is it fair to not assume the language for the convenience of the many people in the United States who speak English, for the sake of the few who don't? Would you put every user through one redundant but no-brainer click rather than a small percent of users through three redundant clicks? Would you automate for the convenience of many when it's correct, for the price of the frustration of few when it's not correct?
When deciding this, take into account the following factors. The more you answer yes, the more you should lean on automation.
- Is it necessary for getting targeted content (e.g. books only in the selected language) instead of just being a convenience for the user interface?
- Is your method of guessing the user's language reliable 97% of the time?
- Are you prepared to lose that 3% that might get fed up with correcting the mistake and leave?
- Is it not imperative that every user understands the content 100% (e.g. non-government websites)?
- Is the international content human-translated and certified to be verbatim and proper?
- Is it possible for you to create a way that the user can change the automated choice with one click?
Of course this is just specific to the automation of language selection, but these factors can in one way or the other apply to any type of automation, where the website guesses something instead of letting the user make a choice for themselves. There is no automation to decide this for you, but enough user testing with not your aunt will reveal if the automation is in place or not.
