That's a nice bit of legwork to show that poor UI design translates directly into erroneous user interactions.
In conjunction with other recent Google stories (https://news.ycombinator.com/item?id=7107325) it's both cautionary, largely exculpatory of users (Google caused this problem), and speaks more than slightly to Google's hubris (I've seen plenty of poor UI from them over the past several years).
An interesting aspect of this dynamic is that the object and subject of the error are asymmetric in this case. It's one thing to accidentally out yourself through "not reading instructions" or "just winging it" or "accidentally hitting the wrong control". It's another to accidentally take an action that hurts another person -- here in a significant way as it directly affects their reputation.
Hacker News is of course hardly blameless. I've more than once downvoted a comment or flagged or upvoted a story, especially on mobile, trying to access other controls on the site. At least the story flagging can be undone, though the other actions cannot be. But that's only if I care enough to do so.
Google design sucks. They've never in the history of the company made a solid effort to nail UI and it's not part of their culture at all. Every one of their products is marred by UI inefficiencies, clutter, or inconsistency.
* They've never in the history of the company made a solid effort to nail UI and it's not part of their culture at all.*
Regardless of how you feel about how it turned out, they clearly have spent the past two years focusing on unifying design and UX across their products.
It's helpful to have meaningfully consistent metaphors and themes across applications. For the UNIX / Linux commandline, that's processes and pipes, BSD and GNU argument flags, and the like.
There are various competing keyboard accelerator schemes: emacs, vi, Wordperfect keyboard templates, DOS, Windows 3.x, Windows 9x/NT, Apple's Mac hotkeys, GNOME and KDE....
There are fundamental layout, font, and color palette schemes.
But one of the worst overall design features I see is taking a given scheme and applying it without regard to suitability across a wide range of applications with a wide range of use cases. For Google this results in the insanity of persistent browser elements occupying 25-30% of vertical screen real estate in Gmail, Groups, Book search, and other products. When the workaround is for me to either push my browser window half-height above my viewable monitor frame and drag down the bottom of the window, to be able to present my work front and center, or to open the elements inspector and delete elements from the DOM just to be able to see the relevant work, something's very, very wrong.
Microsoft of course was infected with similar idiocy with its modal, non-resizeable dialogs and wizards. Even GNOME, for all its UI/UX brain death, generally provides for resizable dialogs, and allows text to be copied from its error pop-ups (under Windows the only recourse was and to the best of my knowledge remains screenshotting the damned thing).
The worst thing for me is that they design for big screens and I have small one. Even when you pick up compact scheme.
The record was stats screen on blogger which shown me 3cm of information panel and the rest was taken by huge sticky !mostly empty! top and bottom rows.
You see that everywhere, unfortunately. I'm on a 2012 Macbook Air, and the Bootstrap documentation's sidebar navigation doesn't fit in the height of my screen.
If that doesn't demonstrate how hard it is to get responsive web design right, I don't know what does!
Ironically: I've got a reasonably large screen (1920x1080), but with a 16x9 aspect ratio, height is constrained. The physical size of the display is fairly small, so px-specified layouts (particularly for fonts) are often painfully small.
I haven't made any real use of Windows in over a decade. I know that functionality didn't exist as of WinXP and probably later. When was it added / where do you see it?
This goes back to at least XP, if not 2000. You don't see it since you're just pressing a hotkey and get something in the clipboard. In the same way you would make a screenshot which is also not a visible operation per se.
Their websites were much more usable before the UI makeover. I guess the focus of the new UI is to make it appear shiny and vibrant and spacious at first glance. But it has come at the expense of usability.
Having used DDG as my primary search engine since June 2013, I prefer its presentation.
I find Google's dynamic "oh, we're going to clear the page while you're refining your search terms to focus on higher relevance based on the results you're looking at now while you're typing that up" particularly egregious.
Google's win right now is in specific categories and sub-classes of search:
⚫ Google Books. Particularly the Ngram viewer. My frustrations with the company have reached a point that I hate to admit I love any of its products, but I love Google Books search and the Ngram viewer specifically. GBS makes serious amends by having one of the most spectacularly useless and annoying interfaces for actually reading online content. Contrast with the Internet Archive's book reader, which is a thing of joy, beauty, grace, and perfection: http://archive.org/stream/industrialrevol00toyngoog#page/n6/...
⚫ Google Scholar. I'm not aware of any other resource which has comparably deep coverage of research papers. Again, what the search interface giveth, the content interface taketh away, with the same online reading interface as Google Books, in the rare cases where journal publishers of recent content actually make full text available.
⚫ Date refinement. If DDG offered this my use of Google Search would fall by another 50-75%. Being able to search within specific bounded ranges of time is hugely useful.
⚫ Maps, though I'm finding/researching alternatives. Mostly OpenStreetmap though "maps.google.com" is still tied to my muscle memory. Streetview remains tough to beat.
⚫ News, though I'm finding/researching alternatives. Twitter's actually pretty good for breaking stuff, with search. Reddit isn't bad as a filter either. I've also got a few conventional news sources (NY Times, Guardian) on my rsstail feeds.
⚫ Financial data, though there are lots of alternatives. This is an area rife with perverse incentives, however, and for the data I'm typically interested in, research tools such as FRED (Federal Reserve of St. Louis), Wolfram+Alpha (despite its complete and utter lack of a definitive syntax), and upstarts such as Knoema and Gapminder are promising. I believe this period is actually showing signs of being a renaissance in advanced online data analysis and access tools.
Increasingly though, Google's my 2nd or 3rd tier search destination.
Don't forget the best new feature: Google+ integration right in your search results!
Now you can see all the Google+ post blurbs right at the top of your results even if they aren't good ones! Also, Google+ notifications about random people adding you? That's right, it's right at the top of every Google search page with a nice red highlight so you don't miss it!
For numerous reasons, I've adopted a hit-and-run policy for Google services. I'll log in, do my deed (mostly pimping my content posted elsewhere), and log out.
Thankfully this greatly reduces the annoyance of G+ intrusion to all aspects of the Google empire.
I'm dreading the prospect of having to use Google tools in a work environment given the disease.
That's a simple one, I just don't use Google+. I toyed with it for a while and deleted my account. Yes they nag me about joining Google+, but this is it. My search results are clean and no red square.
Agreed on Google Scholar. I actually prefer the interface of Microsoft Academic Search, but the coverage is just not good enough. Google Scholar indexes nearly everything, often within a day or two of it appearing on the internet. Whereas Microsoft has only managed to find about half my papers, and has somehow not managed to find any of them more recent than 2010. That's part of the difficulty in beating Google here, that before you can compete on user experience, you have to at least come close to matching their indexing.
Chrome, as Google Search (google.com main page), was very well-designed at the start. From there, it only went downhill, as soon as the marketing people interfered.
That's just because they weren't left with a lot of elements that they could change (most of Chrome is really minimal). However, the welcome page/new tab page has gone through some really bad iterations (app store, Google+, search box, ...) that removed most of the minimalism and simplicity.
So here's the UI in question - http://blog.testfairy.com/wp-content/uploads/2014/01/quick_s... - and apparently the large stars represent not a suggestion rating, but an app rating. So dismissing suggestion as a bad one you'd basically give a 1 star rating to an app. Or so the devs claim.
Coming from the iOS side, I just don't get it. How can you rate an app that you haven't downloaded? And then if you did download it, why is it being quick-suggested to you again?
Furthermore, if you did in fact download it, removed it and then reacted negatively to a suggestion to download it again - how's your low-star reaction is NOT really a rating for an app itself?
Also, if you look closely at the graphs, the "crashing" involves rating slide from 4.67 to 4.38 and from 4.79 to 4.74. It's still an unpleasant development for sure, but that's not a "crash" or a "drop", that's a dip. The matter appears to be completely blown out of its proportions.
I'm a developer, and I have the same rating fall in my apps too.
Re-scaling is good, as long as it is fair.
I have no problem with making it easier for unhappy users to vote, but I do have problem with a suggestion box which misleads at least some users to rate by mistake.
And I also have problem with Google surfacing apps deleted so long ago that the user has no chance to remember what it was. And once the user taps on any star, the rating is registered - no way for the user to correct a mistake or to explain a negative vote, so the dev could do something about it.
It's just Google being careless about the devs.
And it was a surprize for me to realize that. I thought better on Google before it started.
Also, if the dip affects everyone equally, then its more like a rescaling (presumably on average these reviews are lower than before, I'd guess because no-one bothers to review a 2-3 star item, only if they love it or really hate it)
If the dip affected everyone equally it wouldn't be that bad, as most of the saturation found above the 4.8 would be leveraged... However these suggestions will be rated, in average, as greater as the overall empathy and knowledge users have regarding the kind of app in question. For instance, regarding my app Tabata Trainer (which went from solid 4.57 in December 9th to 4.21 at the time of this post), how many persons identify with the word "Tabata"? Most of them will never give a 5. If someone creates a game with a very similar name to a widely known and adorable game, then people will probably rate the suggestion highly. This kind of stuff will just lead apps to be rated unequally... and bad (really bad) apps will raise their ratings, which is shameful.
Right now my Play store app is asking me to rate apps I have neither bought nor installed. If I dismiss one, it shows me another such "not purchased" app. I've seen 3 already. Is this happening to anybody else?
I understand the frustration of the developers affected by this UI change, but I think that actively asking users to rate an app they have downloaded may lead to better ratings in a statistical sense.
User ratings greatly suffer from self selection bias (http://en.wikipedia.org/wiki/Self-selection_bias). Asking users who didn't initially rate an app do so as part of some other workflow might reduce that bias a little bit.
That said, I'm very distrustful of user ratings in general, both as a developer and as a consumer, because people often rate stuff based on criteria that don't concern me at all.
actively asking users to rate an app they have downloaded may lead to better ratings
But it's not asking them to rate the app, it's asking them to rate the suggestion. And since rating a suggestion is usually a thumbs-up/thumbs-down, it primes users to select the extreme ratings.
Sadly, many users just don't read past the title.
And users familiar with other suggestion engines (pandora, etc) expect to teach the system their taste and relevance.
It would also help if the graphs in the post started at 0 stars on the y-axis, because the drops look much more severe than they really are.
I imagine in an ultra-competitive niche that the drop in the third graph from 4.72 to 4.54 could be a painful drop, but the first graph has a drop of .03 stars, and the second is a drop of .05 stars. Perfectly in line with a very slight adjustment (though I'm sure it's alarming when you first notice it happening).
To be fair, the graph formatting doesn't look like the author's fault, just the analytics page automatically zooms and crops like that.
When Google started to require users to use G+ and real names to post a review on the Play store, a number of reviews dropped significantly. The poor design decision described in the article may be a result of an effort to fix this dropped user engagement rate.
On my 2.3.3 phone, the quick rating widget takes the entire width of the phone and about 15% of the height. Any flick that originates on a star is irreversibly treated as a rating. Once the star is touched, there is no way to cancel the action. My flick style never results in any accidental 5 stars (because it is near the right edge). Wonder how many of my app users are inadvertently screwing my rating because of this design. I am surprised that Google's has such a low regard for UI design. They hide their contact forms, and don't respond when you happen to find the form.
I'm Android dev, and I think the "Want quick suggestions?" feature is a good addition to Google Play, because it encourages users to rate apps.
Yes, the average rating for my app has gone down since Google made this change. But rankings are relative, not absolute, and all apps are being treated equally, so it shouldn't make a difference in the rankings.
What wasn't mentioned is that the number of ratings has increased substantially. I get at least twice as many ratings as before.
In my opinion, more users giving ratings will result in more reliable rankings.
I think legislation should be made (maybe anti-trust laws?) which requires companies to provide adequate, free 1st-level user support for paid services, and reasonably priced user support for free services.
The current way of big companies "just doing stuff" which deeply affects other people/businesses who do not have a single way to contact the offender is a nightmare which has to stop. As the market obviously fails here, legislation is required.
I added a dialog that asks the user rate the app, remind later or just ignore it.
The trick is to add the dialog to show up on like 5-10 startup off the app.
If the user is starting the application after 5 tries, he is going to like it. And provided the way to rate it in couple of clicks, the user usually gives positive ratings.
Maybe it's better if you have a different version of Android but when I search for anything broad I see a bunch of free shovelware apps. I'd like to have an option to see just the paid apps or the most expensive apps just to clear out the noise and the spam.
I see similar behaviour on my app.. Many 1 star ratings without any comments.. Initially I was convinced it must be competitor playing games.. But the trend continued.. Not sure how long Google will take to address this issue..
I might be an overly conservative old fart, but there was a time not so long ago, where your success wasn't depending on the UI decision (or any other decision) of a corporate overlord.
Back then, you made a product and it was immediately available for all the world to use. Yes, you had to spend some time/money/effort for promoting it and yes, the success might have been dependent on your ability to promote it.
But you didn't have to ask for permission to build it and there was no way to get screwed because the UI design of an entity you have no control over has changed slightly from one day to another.
I think it's a real shame that these days seem to be over more and more.
The number of times I have seen the database connection error for an obscure site which appears on Hacker News is alarming. It is not like millions of users are trying to get to their site that they go down. If you can't host it don't - why can't sites just use something like Posthaven or something - to ensure more reliability.
In conjunction with other recent Google stories (https://news.ycombinator.com/item?id=7107325) it's both cautionary, largely exculpatory of users (Google caused this problem), and speaks more than slightly to Google's hubris (I've seen plenty of poor UI from them over the past several years).
An interesting aspect of this dynamic is that the object and subject of the error are asymmetric in this case. It's one thing to accidentally out yourself through "not reading instructions" or "just winging it" or "accidentally hitting the wrong control". It's another to accidentally take an action that hurts another person -- here in a significant way as it directly affects their reputation.
Hacker News is of course hardly blameless. I've more than once downvoted a comment or flagged or upvoted a story, especially on mobile, trying to access other controls on the site. At least the story flagging can be undone, though the other actions cannot be. But that's only if I care enough to do so.