A common practice in product development is assuming that a successful execution can be dropped into an unrelated context.

Take the iOS sliders as an example:

These work well with a touch-based OS, as you ‘grab’ the raised portion of the component and literally slide it from left to right (or vice versa).  You’re moving the component from one location to another and the completion of that movement correlates with the value being modified.

Take a look at the American Express communication preferences:

I haven’t seen a control like this on a web site, though I assume the analogue is the iOS slider control, so here’s me attempting to grab and slide it:

What we have here is a component that relocates when you click it.  This is different than the iOS slide where you take responsibility for the relocation.

These are on/off options and our vocabulary is for those is well-understood – checkboxes!  Spend less time on this sort of thing please :)

Continue Reading

A few weeks back I was in Youngstown, Ohio.  Figuring Verizon’s reception would be better than the wireless in the hotel, I picked up some GBs on the run (a little over 1000) and signed up for the iPad data plan.  As it turns out, wireless contention isn’t an issue in rural Youngstown hotels.  Regardless :) I’d like to cancel the recurring 3G data plan so here we go.

Which one of these will help me cancel my plan?

1. I’d like to think the button labelled Cancel next to Cellular Data Account would do the trick.  Unfortunately that just closes the window.  Given that this is a status window, “Cancel” isn’t appropriate – you aren’t “cancelling” any action in progress or pending.  E.g. if this were a dialog that said “Printing… 1 of 3… 2 of 3…”, an option to cancel would be appropriate.
2. Hmm… well I don’t want to change my plan.  I want to get rid of it.  In some wacky way, a “cancel” is a form of “change” so this is also a candidate.
3. This is viable as well – perhaps I have to delete my account to cancel and sign up each time?

If you guessed (2), you’re the lucky winner!

How about some slight changes to remove ambiguity and conform with iPad standards?

Notice the replacement of the Cancel button with a Back button, positioned as one would expect on the iPad.  All plan changes (including cancelling) are clearly indicated on the first option.  This isn’t a grand redesign, just a slight tweak :)

Continue Reading

Or “How Not to Re-launch Your Website.”

AT&T recently unveiled their new and improved customer portal.  I’ve only been able to experience the front page though, so I’ll have to ask around as to what the rest of the experience is like.

Here’s a screenshot of my failed login attempt:

A few points of question:

1. This error message seems to indicate that there’s a transient error with AT&T’s system.  I would like it to tell me whether or not this is a maintenance window or when I could expect to be able to login again.
2. This error message seems to contradict the first error message.  Is there a system error as described in (1) or did I enter my password incorrectly?
3. Which “drop down menu” should I be choosing on this page?

I’m going to guess that the people designing the site are not the people writing the error messages… that the team responsible for internationalization and localization (i.e. translating the error messages, etc.) were working off of a different specification of what the ultimate site would look like and hence the confusing language.

If this is indeed systemic, I would expect alarm bells to be going off in AT&T’s call centers for fielding more calls because the portal is busto.

Then again, years of substandard reception, dropped calls and poor call quality for their iPhone users didn’t prompt them to do anything so I’m not surprised :) (but also :( since I’m a paying subscriber!).

Continue Reading

Pokin and I are frequent users of Priceline.  Being 45 minutes from LA and a bit over an hour from Seattle, we head down (and up) a few times a year.  We’re savvy enough travelers that we can get Priceline to give us decent deals in great locations.  After each stay, we receive an email from Priceline asking us to rate the  hotel.  That link takes you to a page that looks like this:

The first visible page of the form contains 92 radio buttons.  Across the three screen-fulls of this form there are a total of 125 controls for 25 questions.

When faced with a wall of controls like this, I feel overwhelmed.  I mentally switch off for a few reasons:

• There’s a disconnect between when I expected and what I received when I followed the link.  When I click that link I’m expecting a tight query for my thoughts on the Hotel Andra.  Instead there is (for example) a question about whether or not I’d recommend Priceline.

Take this feedback form from OpenTable.  Note the use of background contrast to indicate which section is important/requred.

Back to Priceline:

• They’re asking for questions that make their lives easier, yet have nothing to do with my desire to giving feedback.  These questions are so that my input can more easily fit into the various slots on the Priceline website where it’s likely to appear.  In addition to asking for my time to make their website more valuable, they’re asking me for additional time to make it easier for them to do so.  For example:
  • “Tell us something you liked about this hotel in 20 words or less.”
  • “Tell us something you did not like about this hotel in 20 words or less.”
• I  notice that the “on a scale of 1 to 10…” style of rating, which is far too granular.  How can I choose between 6 and 7?  Or 2 and 3?  It’s easier (and faster) for me to make gross assessments like the difference between a 5 and a 10.  They’ve decreased the chances they’ll get anything because of analysis paralysis – there’s too many notches on that scale.

If Priceline wants to increase the chance that I’ll fill this out, how about a bit of progressive disclosure?

Progressive disclosure is an interaction design technique that sequences information and actions across several screens in order to reduce feelings of overwhelm for the user. By disclosing information progressively, you reveal only the essentials and help the user manage the complexity of feature-rich sites or applications.

There are two questions here worth answering, or at least the two that I’d like to answer:

• Rate your overall experience at the hotel
• Write a short review

How about starting with those, and giving me a reason to keep going?  Here’s a quick suggestion:

By prompting with the shorter version to start, you increase the chance you’ll get any feedback at all and you’re giving people a reason to continue.  Instead of monetary incentives, you can use social prompting as well, e.g. “The more you tell us, the more you help other travelers like yourself!”

If Priceline’s desire is to capture as much feedback as possible, I suggest they start small and stick with sound principles to decrease abandonment and give their users (and themselves) a fighting chance when faced with such a daunting form.

Continue Reading

Pokin and I have had a lot of trouble with Wells Fargo through the years.  First they submit my name to the Secret Service for counterfeiting after I returned from a trip to Africa (don’t ask) and then they accidentally charge me for 4 years of monthly bill pay fees. The list goes on…

Where they have excelled is in the usability of their ATMs.  I’ve joked about this in the past though since then, things are much improved.  You can now receive scanned images of your cheques, send receipts directly to your email, etc.

Here’s another example from today.  Check out the previous iteration of the ATM login screen:

This is the new login screen:

Wait, how is this a login screen?  Where are the login buttons?  They have been replaced with what are essentially navigation items.

“Do you want $60 in cash with a receipt from your checking account (which is what you’re most likely to do)?  Or would you like to do something else?”

Contrast this to the previous workflow which required  you to login and choose from a hierarchical menu (withdraw – select account – how much?) to complete your transaction.  Wells’ ATMs have gotten better at saving and promoting your most commonly used actions and this is taking it one very large step further.  This means that most of the time I won’t ever have to make it past the login screen.

Won’t people be confused by the lack of a ‘standard’ login process?  In this case, here’s what Wells has going for them:

• Your interaction vocabulary with the ATM is one of touch, either via the screen or physical keypad.  Yes, something is different about this prompt but since navigation never occurs on the keypad, one of these two buttons is surely the path forward.  Even if you took a guess, you’d have a 50% chance of doing the right thing :)
• The most likely action is called out with the use of color.  I would prefer if the less likely action had a bit less contrast (jet black on white has the highest contrast ratio).

This is a subtle and interesting change that’s worthy of consideration in other domains.

Continue Reading

A little while ago I wrote about Netflix and being puzzled about having a system that suggests movies it thinks you won’t like.  Like an airliner, where are the backup systems that go into place when the system is about to do something seemingly so strange as to suggest you do something you won’t enjoy?

Here we have FlightTracker Pro, a generally great application… except when you need it most.

There’s been some inclement weather around San Francisco and FlightTracker informs me that a flight to Vancouver, Canada has been delayed by 1h50m.  About an hour later, FlightTracker updates:

I know what you’re thinking, “Four bars?  With AT&T?!”  Don’t let it distract you :)  A flight departing early?  I mean, let alone on time but early?  Amazing! Time to hustle to SFO!

Checking the Air Canada site clued me in – nope, it was still delayed.

FlightTracker might be pulling “40 minutes early” from its web service although the iPhone application is certainly free to report on that data as it sees fit.

Ticket fine print states that you have to be at your gate (and therefore checked in) some amount of time prior to the scheduled departure, usually 15 minutes.  Now, through some strange runway antics, you might have been bumped up in the takeoff order but that would mean your boarding time would have been ridiculously early (again, not possible) for this to happen.

I’m surprised the logic exists to report an early flight.  It had to be in there because the determination was made to color the message green.  Is there value in reporting an early departure?  Should I leave early for the airport? If not, it would seem as though FlightTracker has (at least) three options:

1. Report “on time”.
2. Suggest that I check the airline’s site.
3. Present a link in place of the message that calls the airline’s automated flight check line.

Of the three, I would prefer the first.  If there’s no need to change my behaviour (because the chances of a flight leaving materially early are nil) then I prefer to be blissfully unaware of the strange information FlightTracker is getting about my early departure :)

Continue Reading

…the Internet is just messing with me this week.

Continue Reading

Proving once again that uninstallers don’t get the testing love they deserve :)

Continue Reading

Took a crack at re-imagining an infographic from the Economist, from an article on life expectancy in America.

The study, published in PLoS Medicine, looked at four preventable risk factors: smoking, high blood pressure, elevated blood-glucose levels and being overweight. It then examined how these risk factors reduced life expectancy in eight population groups. … (or, put another way, could expect to gain those years if they were to live healthier lives).

Here is the original graphic:

From the title of the chart, I was expecting lifetime measured in years.  Scanning the horizontal axis were numbers in the single digits, off by a factor of 10 however they were close enough that it could be part of a surprising conclusion – perhaps life expectancies for some segments of the population were far smaller than I had imagined?  At the bottom of the chart, this would mean some life expectancies in the 30s and 40s, which is clearly incorrect.

Now searching the chart for clarifying data, I read the subtitle which explained the bar chart values.  Although to truly understand the values one has to do the math – add the “at-birth” life expectancies (shown in the left-most column) to the potential years gained (“value” of the bars).

It also seemed the exception that the male/female difference was significant enough to warrant adding so much contrast to the chart (the blue/yellow per-gender breakdowns).

Having just finished Wong’s Wall Street Journal Guide to Information Graphics, I challenged myself to simplify the chart and improve the understandability.

Lots of changes here:

• The major difference between the two charts is that I’m using a stacked bar to represent the additional life expectancy gained by mitigating the preventable risk factors.  It allows the chart to be expressed at a scale one might expect when thinking of life expectancies in America (i.e. total years, somewhere above 70).  It also helps you understand the baseline life expectancy for the group combined with the relative potential improvement.
• The national average is now presented with additional contrast, so it’s clear where each group stands relative to the baseline.
• I chose not to label the horizontal axis from 0 – 60, assuming that readers  would understand that life expectancy is given in years.
• The male/female discrepancies are averaged out, which caused one change in the sort order.
• Potential gains are called out on the rightmost side of the chart, as I wanted it to be clear that the sort order was from the most potential gain to the least.

Room for improvement:

• The most important area for improvement I see is that the title implies the “how” will be prominently featured, when it’s a footnote (literally).
• There is no legend for the chart.  I’m relying on the implication that life expectancy is frequently communicated in years, and that the rightmost label is linked to the size of the darker-shaded stacked bar segment.  This appears to be the riskiest part of the overall change?

Having spent a few hours on this, I have a new-found respect for those info-graphing as a day job!  Very fun, very challenging.

Continue Reading

Yesterday, our Comcast cable modem began resetting the Intenet connection almost continuously. Time to swap it out! From speaking with Comcast support, this is as easy as heading to your nearest service center and trading it in – no technician necessary.

A quick Google search landed me here, at what appears to be a pretty standard “store locator” on Comcast’s site:

(1)  I’m a Comcast subscriber, accessing the Internet from my Comcast Internet account. Comcast knows I’m physically inside of their network and it knows my modem is associated with my account (to activate your modem, you have to provide your account number). If Comcast knows who I am, and where I am, then why do I have to provide any information?

(2) Is the apartment number query necessary? Do service centers split service down the middle of an apartment building?  I realize it’s marked as optional although I can’t see why providing it helps the person filling out the form.  Taking it one step further, only the zip code might really be necessary.

(3) Whenever you see this option you have to wonder in what context the creators of this form imagine you to be using it. At a coffee shop? In school? When would it be appropriate to forget the information someone provided in a form, unless it was their username or password, neither of which are present here.

(4) (not numbered) What purpose do the gigantic colon characters between form fields serve?

Points (2) – (4) are not as important as (1) because they are improvements that assume prompting you for your address is necessary.

Given what Comcast knows about me from (1), this is what I hope to see the next time I’m looking for a service center.

Of note:

• What you don’t see: No prompts for location.  Comcast should figure this out from the modem/account pairing.
• What you didn’t know: The nearest location for each type of service center is shown.  Previously you only learned that there were different types of service centers once you started searching.
• What you didn’t know: Your nearest service center might not service you.  Each service center only “serves” different accounts.
• What you didn’t know: If you have a Comcast voice account (as I do), you cannot exchange your hardware.  In that case, this page would show a different message rather than your nearest service center – a UI to schedule an appointment for a technician to replace your modem.  I drove from Mountain View to Sunnyvale with my modem to find this out.
• Thumbnail map view should be zoomed out enough such that you can make gross location decision making (“Oh I see Sunnyvale there, that’s 2 hours away.”)
• If the guess is accurate, one-click access to directions (although I could see directions provided here too).  If the guess was inaccurate, one-click access to finding additional service centers.

There’s certainly room for improvement here.  What are the service center hours?  Are they currently open?

All of those improvements assume Comcast buys into the notion of not asking their customers for what they already know about them.

Continue Reading