Today I was helping someone with their android phone (they were telling me how terrible the battery was). I came to the task of navigating through their apps list to find Google’s Play, looking to download an app (U-M developed Power Tutor). Despite scrolling through the long list of apps for about 10 seconds, I couldn’t find the app until I backtracked halfway back up the list.
Later in the conversation, I wanted to show him the app, which I had downloaded previously to my own phone. There, I found it immediately even though I rarely ever navigate to, or use, this app.
After the conversation I recalled the interaction. Why wasn’t I able to find the app on his phone? Well, his version of android had a continuous scrolling vertical list while mine had scrollable pages of apps. I remember when I had a continuous list of apps (on a previous device), and I remember that it was significantly harder to find apps in it than it is for me now with the page system.
This led me to a general question.
Is it easier to navigate a wrapped list of items when viewed by page versus when viewed by continuous scrolling? If so, this could (should?) apply to traditional implementations such as Windows Explorer or Mac’s Finder. Big potatoes.
Back to the interaction: Why would my phone’s method, pages, be superior to a continuous list? But, before finding pro’s, let’s find possible confounding situational attributes.
-Perhaps I didn’t know the actual name of the app I was searching for on his phone (it was “Play Store”), and indeed I didn’t. However, before navigating I figured it was either simply called “Play” or would start with the word (otherwise, so much for their Play campaign). It could have been “Google Play” now that I think about it, but regardless, I guessed right and it therefore didn’t effect my navigation interaction. So, no.
-Perhaps in this instance it was easier for me because I have experience with my phone, and have a cognitive spatial map of the (linear) list of apps, while on his phone I obviously don’t. Checking my position against my cognitive map might be a faster check than doing an alphabet check, and thus this would reduce navigation time for my phone. However, I don’t recall ever doing either of these checks–I was talking throughout the interaction and was navigating on auto pilot, just scrolling away, hoping the app would float through the screen.
Ok, then what does auto pilot do? If you can help auto pilot be efficient, you can help the pilot when he’s in the chair, too.
Let’s take a finer look at my interaction on my phone and see just what my thought process was.
- Try to identify the title of the item I’m searching for. If I don’t know the whole thing, just the first few letters will do.
- Open up the list
- Do a quick visual skim of the visible items:
- Glance down the center of the screen
- Don’t actually read the labels, use pereferals to scan for the first capital letter of the word
- ^^Would this action be harder if designers applied the all-lower-case treatment that is so common these days? Probably.
- Did the visual skim find an item that starts with the same letter? If so, identify each item until found. Otherwise, continue…
- Do an alphabetical check to see where you are and thus which way to scroll.
- Visual skim
- Loosely compare the first letter of one of the last items on screen and the letter you’re searching for using the cognitive map of the alphabet. If gut gives strong response that you’re far away, scroll in the direction gut says. Otherwise continue scrolling in original direction.
- Repeat 6 & 7 until exhaustion, then do an alphabet check while scrolling back through list. Repeat.
How does this vary with continuous scrolling? Well, the visual skim doesn’t work as well, as it checks chunks of the list at once. To use this technique with continuous scrolling, you would need to display chunks of that list at a time… which is how I used search Windows Explorer before I realized you could just type the first letter in (here’s to wishing I could just type half the filename in like in Ubuntu). This hypothetical evidence certainly suggests that paginated lists are potentially more efficient to navigate. This would have a strong application in file explorers–especially when no keyboard is present (ie: Windows 8 on a tablet).
Of course, this calls for an academic experiment. If you’re interested, be my guest (to senior Informatics majors: capstone?). Here are a few control variables that I think would interact with such a study:
- How many apps are visible
- How many apps are in a row
- How big the text size is
- If there are icons or not
- Proper case versus lower case
- User experience, preferably measured in technique development (do they even do visual skimming or are they continuously doing alphabet checks?)