My hypothesis (lightly supported by user research):
When a user loads an application screen with 3 modules (arbitrary number more than one) containing data and info, the user assumes 2 things — The data is real-time and the page will continue to automatically refresh if left open. Anything less requires work by the user experience team to communicate what the user has to do to get as close to that assumption as possible. You might call them real-time hacks (or real-time’s little helpers).
They are the “20 minute delay” this and the “as of 2 days ago” that that we pepper into experiences in order to smooth out the rough patch between technological reality and what users assume.
key assumption is that all 3 modules are in sync
If the data is not refreshed automatically:
Users are pretty accustomed to seeing a refresh button. Upon seeing the refresh button in a recent usability test, the user said “oh, i guess i have to refresh.” Yes…yes you do. Now we must ask ourselves “Where is the refresh button best placed?” and the second question is “Can we rely on the browser refresh instead?” The first will at least convey to a user that she indeed does have to refresh manually.
If the data is not real time:
Ask what the expectations of the user’s are and decide on a need-to-know basis what to tell them. When I load Twitter and a Twitter 3rd party app, stuff appears at slightly different times. I’m ok with the application not explaining exactly what the delay is.
If the page is about the trend in the last several months, the story of that page may not be horribly altered by the data being delayed by a bit. Maybe were talking footnote here.
But, trying to decide to buy or sell a stock requires clear disclosure about when the data was last updated. The latest updates in Google Finance (auto updates of news and key values) show a trend toward my assumption. The standard, usually prominent, delay message has been reduced to a mere “may be delayed” in a light grey footnote… with a link to the need-to-know stuff.
key assumption is that all 3 modules are not in sync
Gah!… this gets hairy:
This is the scenario where users can really begin to mistrust the application. If numbers don’t add up and things are too difficult to understand, the story of the screen gets lost. You can try to bandage up the patient with real-time’s little helpers, but it could get ugly. I will talk about some of the solutions i’ve come across for this scenario in another post. For now, well… I just say “good luck.”