(Reconstructing this post, because TypePad ate the previous version).
Our application has been on the market now for a couple days. And so now, we have something new: users.
One thing I learned long ago about users: They don't always behave the way you expect or want!
One of the little not-so-endearing features of Google's Market application store, is that the descriptions you get are pathetic. You get to use 325 characters and two screenshots. That's it. Somehow, in 325 characters, you have to tell users what your application does, and why they'd like to try it.
This, and the ONE comment you get to submit as a purchaser of your own app, are your only means of communicating with your customers.
So ours reads:
SmartVolume by SF Smart Systems makes it easy to get the most out of your phone.
One screen to manage all your volume settings and even lock volume for each type of sound. No more fumbling around for a setting while trying to hear a caller.
Never miss an important call because you accidentally turned down the ringer.
As a major purchaser of phone apps, I've read a lot of user comments on apps, both on the iPhone (all four of us in my family have one), and on the Google Market application.
As a purchaser, I've found these comments to be all over the map. A few are helpful. Many are just insults, obscene comments, and other juvenile conduct. Others are obviously confused about either their phones or the app's intended purpose.
And many clearly show that the user has not read the app description before downloading, trying, and commenting. Clearly-labeled prank apps being taken for the real thing. People expecting the obviously impossible.
So I was prepared for the worst.
So I took the precaution of posting the first message.
Please send any support requests or problem reports to us at email@example.com. We aim to give fast, quality support.
How much good do you think that did?
Here's a summary of the first comments:
- A report that likes a feature, but complains it won't let her change something on her phone. I'm not sure exactly what she means, but I'd like to know! She also reports the FAQ link doesn't work -- but nobody has asked any questions yet.
- A guy who really loves it. (Thanks!)
- A "Nice try", but reports that while it acts like it'll work, it doesn't -- whatever that means. He reports he's on a Motorola Droid
- A guy who reports it works great on his Motorola Droid
OK, at this point, I'm scratching my head. Do I have a bug? No sign of any problems examining the reports from Flurry. There are a few exceptions that I handle and report, because it came up exactly once during my stress testing. Thanks to Flurry, I can see that it happens once every several thousand passes through that code -- about once per 1000 sessions.
But then I get one more comment.
- A guy who reports it works, but then forgets, but works again if you fiddle with the volume control
OK...this may be a clue. I go back and retest.
After some testing, uninstall, reinstall, test, retest, reboot, I come to understand that the problem only exists between install and first boot. Once you reboot your phone once, it works reliably and properly.
But if you get a low memory situation, and the system kills your app, THEN it does not restart.
When I tested this behavior -- first I tested that it starts up on booting OK. Then I tested that if you kill it, it restarts properly. That worked. But I needed to have tested them in the other order.
With this information, it didn't take long to track down and fix, and I had a new version up on the Market an hour or so later.
(The problem was that the user interface only bound to the service, which causes it to run, but didn't actually "start" the service, which is required to keep it running even after a low memory situation).
But what to do about those comments on the old version? I'd love to have the people retry, and update their comments, but I have no way to communicate with them.
So I just added the phrase to my comment, "NEWLY UPDATED!". And I found this nicely moved the comment up to be more recent than the other comments. Thus, this now marks where the comments are obsolete!
I tried really hard to test low memory. I really did.
I blew it. In my rush to get a fix up there, I neglected a necessary check. Fortunately, a few hours later I happend to be trying UStream Viewer, and it caused a low memory situation. My app blew a fuse (FC - Force Close). A quick look revealed the problem, and up went a second revision.
A quick check of the bad version (1.0.293) showed:
This is a small percentage of the user base, but I don't want any unhappy customers, even for a free app. The good news is, very few people are encountering low memory situations. As part of my update, I made Flurry report low memory situations. Out of 1206 sessions over the past couple days, only 6 have encountered a low memory situation at all.
I'm going to keep doing this trick with updating my market comment, both in a futile attempt to communicate to users who've already posted comments, and to let potential users know which comments are still relevant.
But the other thing I've learned, is I Really Need A Tool to really test low memory. If I can't find one, I'll have to write it.
So that's all the market comments we've received. Our rating is at 4 out of 5 stars, which isn't too bad for a first attempt, and a few glitches.
But what about email? Did we get any email? The Market app makes it easy to send email -- just tap the Email Developer link.
Yep. We got email. After the above Market comments, two emails came in.
The first was from a fellow Android developer, commiserating about a post about getting useful user feedback. She liked the app.
The second was from an old friend, whom I'd worked with at MIT and Symbolics back in the 1970's/1980's. Howard I. Cannon is cofounder and CTO of ModelSheet Software which is in public beta now. It generates nice, easily customizable Excel spreadsheet templates for various business modelling needs. A very clever idea. Since business is addicted to Excel, but it becomes unweidy -- do the customization separately, rather than by editing the spreadsheets directly.
Howard had some great feedback for me. I was supposed to call him back today for more, but life intervened, but we'll talk more this weekend. And I was able to return the favor by looking over his beta site as well.
But that's it. The only email has been from other developers!
That's really not that surpising. We developers understand how valuable feedback is. Even if it annoys, it's our lifeblood. Customers, especially for free or inexpensive apps, often don't have so much investment in the product. To them, a problem is just an annoyance.
But it's really a partnership. If you take a moment to contact a developer with a problem, you'll help make a better product!
I've done everything I could think of, to make it easier for users. I've used Flurry to automate reporting of errors and other things that will help me make the application better. I've put a menu item in the app to take them to our support page, where they can report problems. I've put in an extensive tutorial (TOO extensive, reports Howard, and I agree) that covers every aspect of the program, and the ins and outs of volume controls on the Android platform.
But it'll never be enough.
Remember all this, the next time you're reading user comments in a phone app store.
I really hope Google folk read this. The Market really needs some work, and #1 on my list is to improve the communications between developers and users. By contrast, the iPhone app store allows over twice the communications and a lot more tools.