Pokemon Go - a more complex version of Monster Cache

This week Niantic (aka Google) released Pokemon Go.  The new game uses location-based augmented reality to allow players to find and catch Pokemon (which is short for Pocket Monsters) at locations that players must physically travel to.  You can then train/upgrade the Pokemon, then use those Pokemon to battle other trainers.  The first part of that sequence -- using a location-based model to find and catch monsters -- is exactly the concept of Monster Cache, our application that was released a few years ago.  Of course, Pokemon cartoons and other games existed way before Monster Cache did, so I'm not at all saying that Pokemon Go copied Monster Cache.  And I know that Pokemon Go will be more fun for a lot of people who are already familiar with Pokemon and the monsters and the battling.  Its just disappointing to be an independent developer with an application and then see a partnership of two huge companies like Google and Nintendo release something so similar.

I still believe that there's a place for Monster Cache alongside of Pokemon Go.  Not everyone is into Pokemon, and so lots of people won't care about the particular monsters.  And many people are only interested in the finding monsters and not so much in the training and battling of the monsters.  So Monster Cache will live on!

2015 Recap

Unfortunately, 2015 was fairly frustrating year for Monster Cache, mostly from a development standpoint.  

At the very end of 2014, I released an almost completely rewritten Android version that no longer uses PhoneGap.  I did that because Android was requiring a change to how In-App Purchasing worked, and Phone Gap didn't have a plug-in for the new method.  I thought rewriting the app in Java might not be too hard since only the phone/client code was changing, not the server, and I had all the logic completed so it was just a translation.  Well, it was way harder than I thought it would be.  There were some bugs initially, and of course there have been some Android OS updates over the year that caused minor problems.  So there was a lot of Android bug fixing that I did all through 2015.  The good news is things are stable and working well on Android.

Apple keeps upgrading iOS too.  And their last upgrade broke our registration form.  To fix it, I need to buy a new Mac, since Apple requires newer Macs to develop for their newer iOS versions.  I've been putting off that computer purchase for a long time since there weren't any major problems, but I may finally have to make that purchase soon since the current bug is more critical.  In the meantime, if a person with a new iOS version downloads and can't register and emails me, then I direct them to a temporary web-based registration form I built for that purpose.

I did run a Facebook ad for a few weeks last June.  It brought in some new users and purchases, but as usual, not as many purchases as the ad costs.  I thought I'd run another ad this spring, maybe try Twitter, but then the new iOS registration bug happened.  So now another ad is postponed until I fix Apple version.

I guess its good news that there is still a slow trickle of registrations and purchases.  But its pretty much only a couple purchases a month.  Here's a sad but funny Apple story.  They charge $3.50 to transfer money to my account (Google charges nothing).  If I only get $2.00 worth of Apple purchases in a month, then they transfer the $2.00 to me and charge me $3.50 for it.  (Apple is not nearly as developer friendly as Google.  They also charge $99/year to keep your developer account open.  Again, Google charges nothing)

Other good news is we still get the occasional nice piece of fan mail or good review. This from an email from someone in the Netherlands: "This app is the best!!!. Me and my brother go outside almost every evening to catch some monsters. this app is just fantastic!"  And a 5-star review from Germany, translated as: "That's fun".  47 ratings on Google and average rating is: 3.7 stars, which I consider respectable.  Munzee, a similar app, has an average of 3.5.  Ingress, another similar app by a big studio, has an average of 4.5.  

Latest Stats

It's been a long time since I posted any stats, over a year and a half.  Things have been steady over all that time.  Still the occasional in-app purchase, and the occasional user climbing the Legends board catching an impressive number of monsters.  At some point this summer, we started only showing 200 total players on the Legends board, so its not obvious when someone new catches a monster.

Back in June, we ran another (3rd?) Facebook ad.  Right before then, I recorded these stats:

# "active" (not test or failed registration) users in database: 1716
# who have found monsters: 501

So that was almost double the active users from May 2014, or about 2 new users a day.  And appropriately, it was also almost twice the number of people who have found monsters.

Now that its November, another 5 months later, here are the latest stats:

Apple downloads: Apple changed their analytics and I seriously can't find this
Android downloads: 197 current / 1278 lifetime
# "active" (not test or failed registration) users in database: 1845 (not quite 1 per day since June)
# who have found monsters: 536
Monsters found: about 18,000 (about 5x that of May 2014)
Vimeo video views: 1285 (about 2x year and a half ago)
YouTube video views: 787 (in Eric's account: 366)

And the reviews.... seem consistent.  We have about twice as many Android reviews as June 2014, and the average is exactly the same: 3.7

What's interesting is that since May 2014, installs and active users and video views have all about doubled, but total monsters found has gone up 5x.  I think that means we've acquired some passionate users in that time frame.  And the Legends board shows it.  11 people have caught over 500 monsters, 4 over 1000!

And one last interesting stat, 4 of the top 21 players are from Germany, and #'s 22 and 23 are from Belgium!

Monster Cache reviews

We finally made it to the point of having a somewhat significant number of reviews for Monster Cache.  A few are from people we know. of course, but most of them aren't.  After 21 Google Play reviews, here are the results:

# of Ratings
21

★★★★★48% of ratings are for 5 stars10

★★★★★14% of ratings are for 4 stars3

★★★★★14% of ratings are for 3 stars3

★★★★★10% of ratings are for 2 stars2

★★★★★14% of ratings are for 1 stars3

 Average Rating

3.71

On Apple we've only had 4 reviews.  I don't know why so many less than Android.  Of those 4, we only know one person.  All 4 are 5 star reviews.

An average of 3.71 on Android is okay, but I would have liked it to be over 4.  3.71 is kinda like a B+, I guess.  It's great to see when people love the application, and of course its a bummer to see the bad reviews.  As I mentioned when I posted about Hawaii, one of the 1-star reviews was because someone from Hawaii installed it and I never generated monsters there. I only recently discovered that I could reply to the reviews if they had a comment. When I generated monsters in Hawaii and responded to that review, the person increased their rating to 3 star, that was good to see.  I just replied to the other 1-star reviews today.  Here are all the reviews I've replied to:

★★★★★ ALCATEL ONE TOUCH Fierce (Rav4)

Maddie Linsacum on Apr 5, 2014 at 10:46 AM They make u pay!!!!!! Got the app so exited left it on while in the car I have no friends= no money= no nets=no more monsters :'( You replied on Jun 8, 2014 at 2:59 PM Hi Maddie, Many free applications rely on advertising and/or In-App Purchasing to generate revenue. We chose not to use advertising in Monster Cache, so once you've used all your free nets, you can get more through In App Purchasing or by referring friends. We're sorry that neither of those options will work for you.

★★★★★ App version 1.2.1 Galaxy S5 (klteatt)	steven mills on Jun 7, 2014 at 9:31 PM This sucks All of the "monsters" are on private property, behind gates, and places you don't need to be. And when you do use all of you nets, you have to buy more? It's free to geocache, munzee, ingress, I'm not paying to drive around just to look at all the monsters you can't get. You replied on Jun 8, 2014 at 2:54 PM Hi Steven. Although some of the randomly placed monsters end up on private property, certainly not all do. Also, the net's long range often allows a monster that is in a residential or commercial lot to be caught from the sidewalk. We're sorry that you didn't find any yet that were accessible, but there are millions out there so please try again. Edit your reply
★★★★★ App version 1.2.1 Nexus 4 (mako)	Lance Shimabukuro on Jun 6, 2014 at 10:09 PM Ok, no monsters in Hi fixed, will try later. Got the apology email within a day, tried briefly but @ work, so will look again later. Revised rating up already and will adjust again after I give it a real tryout. You replied on Jun 5, 2014 at 8:10 PM We are sorry for the inconvenience. You are correct that monsters had not made their way to Hawaii...until now! Now there are thousands of monsters all over the islands, so we hope you will try the app again, catch many monsters, and reconsider your rating. No need to re-register, the account you had made before will still work. Thank you!
★★★★★ App version 1.2.1 htc one (M8) (htc_m8)	Kelli Deems on Jun 1, 2014 at 3:14 PM Love the idea! Caught one monster. So many more are out of reach. I seriously cannot go walking into a freshly planted corn field, and that seems to be where they like to hide. There is no way to reset the map, so that is where they stay. I haven't seen new, more accessible ones appear. Nice idea, but not working out very well. You replied on Jun 3, 2014 at 5:30 PM Sorry that many monsters in your area were out of reach. We agree that you shouldn't try to catch monsters on private land without permission. If you can find more Level 1 monsters in a more urban or suburban area, then you'll gain access to the Level 2 monsters, and hopefully more of those will be in accessible areas near you. Good luck!

Azure osFamily="1" discontinued

As I mentioned in my last post, we had one technical issue to overcome last week.  The server for Monster Cache is hosted on a Windows Azure server.  I set that up when we first started testing, around 2 years ago. At that time, I had a 6 month free trial, or something like that.

It was probably a couple months ago that I got my first email form Microsoft warning that Guest OS 1 was being retired.  That was totally meaningless to me.  I've never heard of "Guest OS" of any version.  Later, they sent more emails saying things like "You *may* be affected by the retirement of Guest OS 1".  So I kept the emails in my inbox and thought I'd look into it more later.  The day I looked into it happened to be Sunday June 1st.  I opened the email again, and noticed that coincidentally the OS was being retired that day,Sunday June 1st.  I followed the instructions to determine if my server was affected or not.  Sure enough, it was.   Uh oh.  The site hadn't stopped working yet, but it was time to get to work.

I had no idea I was using Guest OS 1.  What determined that was a simple attribute in one of the configuration files that said osFamily="1".  It seemed it might be as simple as changing that to osFamily="4" to sue the latest version, but not quite.  The different osFamily version required a different version of the Azure SDK, so I would have to download that.  The new SDK required that I use Visual Studio 2013 instead of the Visual Studio 2010 I had always used for my Azure project.  Thank goodness I had only recently purchased Visual Studio 2013 for the sake of a different project.  So I installed the new SDK, opened the project in Visual Studio 2013, let it perform a migration, then tried to run it locally.  It failed.

When I tried to run the project locally with the debugger, the emulator would just hang.  So I started doing some research for the status messages I was seeing.  Found lots of posts of people saying that happened with the debugger, but not if you ran without debugger.  I found that was true for me too, but it didn't make me confident to release it that way.  I eventually found it was an emulator setting.  More details are in an answer I posted on Stack Overflow.  Then I got another error on my static pages after that, which required more research and another change to one of my configuration files.... so 3.5 hours later, I finally had everything running again with the debugger in Visual Studio 2013.

As it turned out, Microsoft extended the cutoff date for Guest OS 1 to September 1st.  So the server never went down, and I released the changes with plenty of time.  But just one more example of hours of work needed for a Google or Apple or Microsoft upgrade.

Monsters in Hawaii!

With all the activity these past 3 weeks, we had one technical issue (with Azure) that I'll describe in another post, and one non-technical issue.  The non-technical was that someone  in Hawaii posted a one-star review on the Play store, saying they installed the app, saw no monsters, then uninstalled.  Its true, I never generated any monsters in Hawaii.  The reason I never generated any there is because I had to do separate boundaries and generating for that area, and I wasn't sure the small area and population were worth it.  I told Kathleen this, and she said she'd sent emails to geocachers in Hawaii too.

So I quickly started generating monsters in Hawaii.  Using my new monster viewer, I could see the ones I was generating.  Of course, I'm bounding the islands with rectangles, so many are ending up in the water.  And many more end up in huge areas of forest, since that's what most of Hawaii is.  I started focusing on Honolulu then.  My initial numbers weren't making as many there as I hoped so I kept increasing it.  After a couple hours of testing and generating, I had the islands pretty well covered.  I then replied to the one-star review with:

"We are sorry for the inconvenience. You are correct that monsters had not made their way to Hawaii...until now! Now there are thousands of monsters all over the islands, so we hope you will try the app again, catch many monsters, and reconsider your rating. No need to re-register, the account you had made before will still work. Thank you!"

No adjustment to the review yet though.  But I also emailed all users in the database with home state of Hawaii to apologize and tell them the monsters were there now.  The good news is someone found one the next day and posted a message on the Facebook page!  That was good to see.

Second Facebook Ad

Tried some new things in the new ad.

• Separated mobile and desktop for US.  Appears that for desktop we were getting clicks for less than half the price, $.09 / click instead of $0.22
• Made a US and non-US add that include people who like Pokemon

I don't like the fact that I had to continue my old ad "Campaign".  Campaigns, I guess, are defined as who you are targeting rather than a span of time.  Anyway, here were summarized the stats for the last time I ran the ad:

WEBSITE CLICKS?

291

REACH?

15,931

FREQUENCY?

1.75

TOTAL SPENT?

$94.35

AVG. COST PER WEBSITE CLICK?

$0.32

-----------------------------
UPDATE
-----------------------------

The Facebook ad has been running for two weeks now.  Splitting the ad didn't seem like it worked well, the mobile was costing $0.35 / click, and Desktop one by itself wasn't doing much at all.  So I went back to the old one and its been doing great!  The average cost per click went down to $0.25, so we're getting a lot more volume of clicks each day for the same daily price as last time.