Mac App Store: Year Two

Today, the Mac App Store turns 2.

Last year, I concluded my retrospective of one year of Mac App Store wondering whether 2012 would see more developers struggling to get their Mac apps approved for sale.

The Mac App Store is not without its flaws, and questions loom ahead as to whether Apple will lock down the system eventually — allowing customers to only install apps from the Mac App Store (as also recently suggested by Adam Engst on Macworld) — and cause a general confusion among developers and consumers as sandboxing is enforced and apps will need to comply to a stricter set of rules to be accepted on the Mac App Store.

A full year after that post, I believe it’s safe to say one word epitomizes Apple’s 2012 with the Mac App Store: uncertainity.

2012

The past year saw Apple “doubling down” on iCloud integration with its mobile and desktop operating systems. In July, the company released Mountain Lion, an evolution of OS X Lion – launched just 12 months before – further bridging the gap between the end user experience on iOS devices and Macs.

As I detailed in my review, Mountain Lion brought important improvements and changes to the Mac App Store. First and foremost, with 10.8 Apple unified the Software Update mechanism in the Mac App Store: users can still launch the Software Update checker from System Preferences, but now both system and third-party app updates happen through the Mac App Store. And alongside multi-touch improvements, Mountain Lion also made the Mac App Store more sharing-friendly as part of the OS’ new and integrated “share sheets”:

Every app’s menu now comes with buttons to share on Twitter and Messages, and Facebook will be enabled as an option later this year. When sharing on Twitter, the OS X tweet sheet displays the application’s icon alongside an iTunes short URL.

Customers’ response to Mountain Lion was, according to Apple, fantastic: Apple sold 3 million units of the operating systems in the first four days (July 25–29, 2012), and, at its iPhone 5 media event on September 12th, announced they had sold 7 million units. For comparison, OS X Lion sold 1 million copies in the first 24 hours; over six million copies by October 2011; and at WWDC 2012 in June, Apple announced 26 million copies of Lion had been shipped.

Apple CEO Tim Cook stated on multiple occasions that Mountain Lion is Apple’s “most successful” OS X release to date. Mountain Lion was, obviously, released exclusively on the Mac App Store. Unlike Lion, Apple didn’t offer a physical copy of the operating system.

In terms of numbers, it’s worth remembering that the majority of the 4.9 million Macs Apple sold in Q4 2012 (Q3 2012 numbers were reported the day before the release of 10.8) are running Mountain Lion out of the box. I’d also point out how Mac App Store-specific numbers were absent from the “state of the business” introduction at Apple’s most two recent media events.

Sandboxing and Gatekeeper

2012 saw the official rollout of two new OS X technologies that have profoundly affected third-party software: Sandboxing and Gatekeeper.

We have covered both extensively here on MacStories: Gatekeeper, a new user feature of Mountain Lion, was given an entire section in my 10.8 review; Sandboxing, after two delays (from November 2011 to March 2012, then extended to June 1, 2012), has proven to be quite difficult to implement for developers of fairly complex Mac apps already approved on the App Store. Developers, who had their apps already available on the Mac App Store, found themselves forced to bolt Sandboxing onto apps with previously-defined feature sets that, in some cases, were incompatible with Apple’s requirements. Lex Friedman published an excellent overview of what Sandboxing meant for users and developers; at MacStories, we’ve been following the story in our dedicated Sandboxing hub.

There’s no doubt that Sandboxing has been, by far, the biggest change that occurred in the third-party OS X developer community in 2012. The somewhat forced adoption of the technology had – and keeps having – far-reaching consequences that have shaped the Mac App Store marketplace and set a precedent for future developers to keep in mind.

Developers feared that Sandboxing would inevitably compromise the nature of Mac software sold through the Mac App Store even prior to Apple’s deadline. But it was after the deadline enforcement on June 1, 2012, that tangible effects started to appear as direct consequences to Apple’s actions.

  • Smile decided to release TextExpander 4 outside of the Mac App Store. Version 4.0 of the software wasn’t compatible with Sandboxing rules, and Smile couldn’t work around them. As of this writing, TextExpander 3 is still on the Mac App Store.[1]
  • After the release of Mountain Lion, Postbox pulled their email client from the Mac App Store.
  • Kevin Walzer said goodbye to the Mac App Store.
  • Longtime Mac developer Manton Reece went back to selling website-only versions of his app Clipstart, citing features that were “difficult to do in the sandbox”.
  • Many Tricks, developers of Moom and Witch (among other utilities), decided to start advising customers on migration from Mac App Store versions of their apps to standard, non-Mac App Store licenses. Many Tricks, too, pointed to Sandboxing issues and even created a comparison table to outline differences between Mac App Store and “website copies” of their apps.
  • Droplr 3.0 took two months to get approved on the Mac App Store due to Sandboxing issues.
  • Many Tricks is also an example of another trend: downplaying Mac App Store direct links on product webpages with smaller, often buried “also available on the Mac App Store” links. Atlassian prominently features a direct download of SourceTree; Bare Bones Software does the same for BBEdit 10, as the Mac App Store version comes with less features due to Sandboxing.
  • Speaking of feature differences: Panic’s Coda 2 webpage clearly points out the differences between app versions. Panic features a Mac App Store link for iCloud support on the top of the page; directly below it, a “Or, get it directly from us” section explains that the panic.com version is suitable for volume licensing and “Mac App Store haters”, without iCloud. And there’s more: Panic was unable to add new features to Coda 2 unless they sandboxed it. In order to achieve a painless experience for customers, they released a Coda Sandboxing Test on their blog.

These are just a few examples of what some Mac developers had/have to go through to ensure a clear and smooth customer experience, not to mention getting approved by Apple in the first place.

Sandboxing has undoubtedly left a scar that won’t go away any time soon. While the world hasn’t ended, Sandboxing – and Apple’s vague stance on some technical aspects, not the best policy when combined with multiple delays – has led to a dichotomy: in spite of Apple’s best efforts, developers are still dealing with two ways to sell their apps – the Mac App Store, and their own websites. Some developers decided to embrace a Mac App Store-only release, but, usually, these apps have moderately “advanced” functionalities (e.g. they are “simple”, one-purpose utilities or games); I challenge you to find a developer of a Mac app with deep filesystem hooks, scripting functionalities, or integration with third-party services who hasn’t decided to put up a version of his/her app on a website as well. Reality is, developers who prefer to take this route – supporting two versions of the same app – are the rule, not the exception.

Matt Gemmell released a new app called Sticky Notifications earlier this year. His experience, which he documented in a blog post, reveals a common trend among independent Mac developers: supporting two separate versions of an app is worth the extra effort.

As it turned out, there was no need for me to be scared about the stuff I needed to do to recreate all the stuff that the App Store provides. If you don’t manage to get onto the App Store (or even if you do), there’s another way you can get your apps out there, and it’s not much extra trouble.

However, I believe there is a bigger issue that goes beyond technical difficulties and best practices to release an app on the Internet. An issue of trust. Perfectly detailed by Marco Arment a few months ago, the changes that Apple forced upon the developer community in medias res have made him lose confidence in Apple’s ability to maintain an OS X storefront that won’t, eventually, cause more problems to customers and developers.

In the first year of the Mac App Store, before sandboxing, I bought as much as I could from it. As a customer, the convenience was so great that I even repurchased a few apps that I already owned just to have the App Store updates and reinstallation convenience. And, most importantly, when an app was available both in and out of the Mac App Store, I always bought the App Store version, even if it was more expensive.

But now, I’ve lost all confidence that the apps I buy in the App Store today will still be there next month or next year. The advantages of buying from the App Store are mostly gone now. My confidence in the App Store, as a customer, has evaporated.

As he later also explained, it’s naive to assume these are problems that only “geeks” notice – as if so-called “normal people” won’t make a problem out of going through hoops to license software again, or worse, having to buy a piece of software twice.

Even the “geeky” apps that get excluded, such as TextExpander and SuperDuper, aren’t used exclusively by geeks. I know because I run a web service that uses a Javascript bookmarklet that people manage to install in Mobile Safari (this is not a simple procedure) that solves a problem that’s mostly only encountered by people who browse the internet all day, don’t want to read what they find while browsing, and want to instead read it on an expensive portable gadget in the future. And they’re not all geeks. One look at my support email makes that extremely clear, and I bet Smile and Shirt Pocket would confirm that the same is true for the “geeky” TextExpander and SuperDuper.

And here’s Neven Mrgan, designer at Panic, on his personal blog:

And the Mac App Store, in its current incarnation, just isn’t built for us. It’s built for people looking for casual apps and games. (Sorry, there’s one more category: Apple’s own apps, which don’t have to play by Apple’s rules.) And that’s also fine. But put the two facts together—the loss of casual users to iOS, and the loss of non-casual apps on the App Store—and it starts to look like a problem.

Personally, I have gone back to buying Mac apps from the websites of developers I trust whenever possible. It’s not that I believe Apple is an evil entity that’s out there to steal my privacy and money: I, like others, simply don’t like the consequences that the Mac App Store brought for apps that had been approved and that I had bought.

I don’t know if future changes in Apple’s App Store strategy will once again force developers to remove apps for sale, or come up with hacks to recognize App Store purchases and turn them into regular licenses: I just know that I don’t have time for this. For this reason, whenever possible, I prefer buying from the websites of developers I already know and trust. It helps that, in doing so, I’m also probably giving them more money, because there’s no 30% Apple cut to account for.[2]

Improvements

It’s not all bad, though. The Mac App Store remains the easiest and most user-friendly way to find an app and install it on a Mac. Thousands of apps are still available on it, and Mac users can now enjoy games on their computers like never before thanks to the Mac App Store.

Delta updates, introduced in 2011, have allowed developers to push lighter, faster updates that don’t require users to wait several minutes to install a software update. There’s still no wish list, but the Mac App Store’s dropdown menus have gained a handy series of shortcuts to share an app’s direct link on Twitter, Facebook, and iMessage.

In my opinion, Apple’s best work on the App Store – not just the Mac App Store – in 2012 has been a renewed focus on curation and, most importantly, human curation. This is a subject I carefully analyzed in my Four Years of App Store article back in May 2012, and which I’ll certainly reconsider for the fifth anniversary of the App Store this summer. Regardlessly of the platform, Apple has started pushing out custom sections on a regular basis, they rebranded the App of the Week to Editors’ Choice, and began updating the App Store’s homepage weekly with custom banners, links to sections, and rotating galleries. Apple has focused more on the iOS App Store – which gained a fresh new look, updated Categories, and a first glimpse at sections inside search – but the curation effort was shared across the OS X and iOS teams. As the number of apps grow, it’ll become fundamental for Apple to carefully curate its App Store to direct customers to the best apps they can discover and buy.

Gatekeeper, initially a source of wonder as to whether it could be a sign of Apple “locking down” the Mac, has actually proven to be a substantial security improvement that, so far, developers appreciate. With a default setting that accepts apps from the Mac App Store or, in their signed form, from a developer’s website, Gatekeeper has been largely well-received.

Last, review times. Earlier this year, several Mac developers reported they were struggling with approval delays for their Mac apps; some of them even had to wait an entire month for an update to show up on the Mac App Store. It was common belief that such delays were related to the release of iOS 6 and the iPhone 5, which required Apple to shift human resources to sift through the incredible number of app submissions (that took advantage of the software update and new device) on the iOS side. It’s a credible explanation, but we’ll likely never know for sure. The good news for developers is that review times appear to be shorter now – within the range of 10 days, in fact. Mac review times still aren’t up to speed with their iOS counterparts, but they’ve gotten better.

Year Two

Last year, I collected some Mac App Store numbers and facts in a section at the bottom of the article. I’m doing the same this year, albeit with less data than 2011. For instance, Apple hasn’t posted a press release similar to December 2011’s Apple’s Mac App Store Downloads Top 100 Million.

  • Total Available Mac Apps: 13198. Last year, they were 8912 after 365 days of Mac App Store existence (via AppShopper).
  • Mountain Lion was released 566 days after the launch of the Mac App Store.
  • Lion went from 0 to 6 million downloads in 76 days. Mountain Lion went from 0 to 7 million downloads in 49 days.
  • Top Paid chart as of today: Mountain Lion, Pages, Numbers, iPhoto, FaceTime, Keynote, Disk Doctor, RollerCoasterTycoon 3 Platinum, Velvet Assassin, iMovie.
  • In Top Paid chart, 7 apps are from Apple (same as last year). Average selling price is $14.59. Of 10 apps, 2 are Games and 4 belong to Productivity.
  • Top Free chart as of today: Evernote, Wunderlist, Kindle, Full Deck Solitaire, The Unarchiver, Mint QuickView, CSR Racing, Xcode, Pocket, VirtualDJ Home.
  • In the Top Free chart, 5 apps are the same as last year, and Twitter is notably absent. There’s only one Apple app in the list (Xcode). There are 2 Productivity apps and 2 Games.
  • Top Grossing chart as of today: Mountain Lion, Pages, Final Cut Pro, Logic Pro, Aperture, Numbers, Keynote, iPhoto, RollerCoasterTycoon 3 Platinum, Adobe Photoshop Lightroom 4.
  • In the Top Grossing chart, there are 8 Apple apps (same as last year). Average selling price is $85.49. There are 4 Productivity apps, 3 Photography apps, and 1 Game.
  • There are currently 21 categories in the Mac App Store (23 on iOS, but no Newsstand on OS X).
  • Currently, 14 sections are featured on the Mac App Store’s homepage.

Based on this data, there are three interesting facts we can extrapolate. Apple is, at the moment of writing, featuring two more sections than last year on the App Store’s homepage. The number of Apple apps in the Top Paid and Top Grossing chart is unchanged, proving that the company’s own apps are consistently in the top 10 positions (therefore they are also featured on the Mac App Store’s homepage more than third-party apps). And last, if we consider AppShopper’s tracking accurate, Apple’s number of approved Mac apps has gone from an average of 24.4 apps per day to 11.7 apps per day.

There are a number of reasons to explain the drop in average approved apps, the most obvious one being that Apple sells more iOS devices than Macs, and thus developers have probably less interest in developing Mac software because they’re a much smaller addressable market. In the fiscal year 2012, Apple sold 58.2 million iPads and 125.03 million iPhones, or 183.23 million iPhones and iPads combined. In the same fiscal year, the company only sold 18.1 million Macs. It is no surprise these days to see new apps coming out for the iPhone and iPad first, with developers “considering” a Mac version. Gone are the days of popular Mac apps becoming available on iOS: nowadays, aside from some exceptions (“professional” software), developers are obviously going “iOS first”.

However, Mac sales numbers are also up from fiscal year 2011, during which Apple sold 16.73 million Macs. As I explained in my Mountain Lion review, people are still buying Macs because they need and they like the functionalities and apps of OS X. So why doesn’t the Mac App Store have more apps?

My best guess is – a combination of all the factors mentioned above. In 2012, many developers’ resources shifted to updating their iPad app assets for the Retina display, then Mac Retina displays, and then the iPhone 5. Some developers had to rethink their sync strategies due to iCloud instability, and others had to face issues with Sandboxing, slow review times, and a general uncertainity in regards to selling Mac software for a profit on the Mac App Store. This, combined with an overall (and perfectly justifiable) trend of making iOS apps first, leads me to believe that, while the Mac App Store is clearly growing slower than the iOS App Store, there’s still another platform that has been serving Mac developers just fine for the past year: the Internet.

I am thus going back to my initial point: developers seem cautious about the future of the Mac App Store and the restrictions that Apple could, in theory, bring up again. The numbers and facts speak for themselves: developers are still using the Buying Direct option. And when the download numbers of the Mac App Store aren’t so great, who can blame them? And when you start adding Sandboxing, iCloud, and Top Charts dominated by Apple apps (and the operating system itself, listed under “Productivity”) to the list of complaints, who can blame an indie development company for their decision to sell software through the good ol’ Internet? Should we be surprised that they welcomed Gatekeeper and, at the same time, worried about it?

Perhaps it’s time for Apple to give back to the OS X developer community. Maybe in the form of paid upgrades and trial versions of apps[3], or maybe by removing iWork and Mountain Lion from the charts and put the spotlight more on software made by others. Apple has plenty of options to consider: power user mode and more developer-oriented features come to mind.

Looking ahead at 2013, it’s not clear what we should expect. On one hand, a major name like AgileBits is working on iCloud integration (therefore Mac App Store release) for the future 1Password 4 for Mac. On the other hand, there’s no mention of iCloud at all in The Omni Group’s blog post regarding OmniFocus 2. How will these (and other) developers handle Mac App Store upgrades? Is this really the best Apple can do to ensure an easy, plesant upgrade experience for customers? Heck, what will Apple do when they will (eventually, we hope) have to release an upgrade to iWork?

The Mac App Store isn’t going away, but it needs to keep growing with quality software. I’m looking at 2013 with a wish: that Apple will start listening to the OS X developer community again, and that they will make changes aimed at facilitating a single, fundamental aspect of the OS X platform – discovering and buying great apps.

But will that be enough to win back customers who aren’t using the Mac App Store anymore? I don’t know.


  1. For context, Smile is the same group of developers who had to come up with a clever trick to enable iCloud sync for the regular version of PDFpen, as iCloud is still exclusive to Mac App Store apps.  ↩
  2. “Probably”, because running your own web store and content delivery network has other costs.  ↩
  3. How many of those 13198 Mac apps are actually the same app in Lite and Pro version?  ↩