Forgot your password?
typodupeerror
AI Businesses Math News Science

The Artificial Life of the App Store 106

Posted by timothy
from the how-about-global-thermonuclear-war? dept.
mikejuk writes "How does the Apple App Store actually work? What is the best strategy to employ if you want to get some users and make some money? There are some pointers on how it all works from an unusual source — artificial life. A pair of researchers Soo Ling Lim and Peter Bentley from University College London, set up an artificial life simulation of the app store's ecosystem. They created app developers with strategies such as — innovate, copy other apps, create useless variations on a basic app or try and optimize the app you have. What they found, among other things, was that the CopyCat strategy was on average the best. When they allow the strategies to compete and developer agents to swap then the use of the CopyCat fell to only 10%. The reason — more than 10% CopyCats resulted in nothing new to copy!"
This discussion has been archived. No new comments can be posted.

The Artificial Life of the App Store

Comments Filter:
  • by Anonymous Coward on Sunday April 22, 2012 @04:49PM (#39765397)

    Perhaps they can simulate how to make slashdot summaries make sense next?

    • That sort of translation technology doesn't exist, as of yet.

    • by Sarten-X (1102295) on Sunday April 22, 2012 @05:09PM (#39765549) Homepage

      This one makes a decent amount of sense to me, though I did do previous work in artificial life simulators...

      Effectively, they built a simulation of the app store, and filled it with developers following several different strategies, and presumably a feedback function that models expected consumer behavior. The simulation was left to run, and interesting results were gathered. TFA is actually a rather well-written explanation that's worth reading.

    • by plover (150551) * on Sunday April 22, 2012 @05:21PM (#39765645) Homepage Journal

      Perhaps they can simulate how to make slashdot summaries make sense next?

      Seriously, if you can't understand this one, go play on facebook or whatever the kids are doing these days. Your life is wasted here, as is a fraction of ours for reading your inane drivel.

    • We do get the privilege of using more than 160 chars here kiddos. Time to work on the parseable sentences.

  • by allo (1728082) on Sunday April 22, 2012 @04:51PM (#39765407)

    Of course a copycat can be minimum efford maximum profit in a simplified model, but this strongly depends on the calculation of the fitness-function. I think it can be hard to match the real world fitness-function, because some of the factors that are relevant to an actual user are hard to calculate.

    • Re: (Score:3, Insightful)

      by goombah99 (560566)

      There's more than one utility function. For example, if wrote an app I would not expect to profit, it would be for fun. Thus I'd give it away for free or a dollar. Someone else might be hoping to make a living at it. too bad.

      • by plover (150551) * on Sunday April 22, 2012 @05:38PM (#39765749) Homepage Journal

        There's more than one utility function. For example, if wrote an app I would not expect to profit, it would be for fun. Thus I'd give it away for free or a dollar. Someone else might be hoping to make a living at it. too bad.

        This simulation was built to identify profit models, not to maximize developer happiness. But the two are related, and profit will be an element everyone can measure.

        Consider if the app you created turned out to be really fun and truly innovative, and it went viral and sold five hundred thousand copies at $0.99.

        If nothing else, you'd learn that half a million users can be awfully demanding. You might find yourself mired in support requests, and have to decide whether or not you can support it yourself or if you want to sell it to a game company so they can manage it. If nothing else, you might be surprised when you discover you have to pay taxes on a whole lot more income than you thought. The point is that at some financial threshold, you will probably have to take it seriously. My threshold might be higher or lower than yours, but in this simulation, it doesn't really matter. It would change your personal view of profiting from your work.

        • by tlhIngan (30335)

          If nothing else, you'd learn that half a million users can be awfully demanding. You might find yourself mired in support requests, and have to decide whether or not you can support it yourself or if you want to sell it to a game company so they can manage it. If nothing else, you might be surprised when you discover you have to pay taxes on a whole lot more income than you thought. The point is that at some financial threshold, you will probably have to take it seriously. My threshold might be higher or lo

      • by Kergan (780543) on Sunday April 22, 2012 @07:50PM (#39766531)

        There's more than one utility function. For example, if wrote an app I would not expect to profit, it would be for fun. Thus I'd give it away for free or a dollar. Someone else might be hoping to make a living at it. too bad.

        Too bad? If you're not meaning for the customer/end-user, I'd wager you never wrote any such app; or any OSS app, for that matter. And that those who tagged you as insightful haven't either.

        In the real world, app development is just the beginning. Unless you decide to accept no feedback whatsoever, which is a losing proposition, you're in for a lot more feedback, emails and/or forum posts than you ever wish you'll ever read in your entire lifetime if you're even remotely successful. It's absolutely insane. Your success will destroy you unless you've an adequate means to scale -- whether monetization or extra funding.

        So here you are, quacking that you'd happily share an app. For free. You'll keep your day job as you do. Someone out there actually wants to make a living off of a similar app. But he or she will get less or no business because you released that -- soon to be unsupported -- app at in an inadequately low price point.

        Look... It's one thing to be competing with a Chinese team who can field $500/month coders to support their app, or with crap hobbyists who only have a slight clue of what they're doing. Those are mostly manageable in practice. It's an entirely different thing to compete with hobbyists who distribute good products without any interest in having a sustainable business.

        Think of it this way: for every $100/month "cool, I got some pocket money I barely couldn't care about" app out there, an actual person who does care might be losing his job. So please do yourself a favor, do that guy a favor, and -- most importantly, in the long term -- do the customer a favor, and don't release it unless you work out your business model first. Else you're just building a mine field for those who do care.

        • by Jeremi (14640)

          Look... It's one thing to be competing with a Chinese team who can field $500/month coders to support their app, or with crap hobbyists who only have a slight clue of what they're doing. Those are mostly manageable in practice. It's an entirely different thing to compete with hobbyists who distribute good products without any interest in having a sustainable business. [...] [they're] just building a mine field for those who do care.

          You've misread your talking points, Mr. Ballmer. Free software is supposed to be a 'cancer', not a 'mine field'.

          • by Octorian (14086)

            Pretty sure this rant isn't about *real* free software, of the open-source variety. Rather, its about free-as-in-beer (or priced close enough to it) closed-source software, with an end-game that has it likely headed straight for abandonware.

        • Interesting argument, having written an Android app that was the first of its kind, I disagree with your argument that hobbyist development takes away jobs. If anything, it can create jobs, exactly because of the "copycat" phenomenon. I wrote an app that was the first of its kind available in the Android Market. There was at least one similar for iOS before I wrote mine. About a month after my first release, I discovered that another company had hired a developer (possibly more than one) to basically copy m
          • It seems my mod points ran out as I was reading your post, so I will comment to confirm this. If there is an app that is needed, I may download and try out the free or cheap apps, but will be more than happy to pay more for a non-buggy version that is supported. Various fields I have hobbies in such as RPGs and photography use various apps like PDF readers and model consent form apps. There are many free or cheap versions of both but word gets around pretty quickly on forums which ones are the best ones to
          • by Kergan (780543)

            Oh, don't take it personally... competing with a hobbyist who doesn't care much that isn't an issue. They fall behind in UX, maintenance, etc.

            The trouble makers are that special breed of hobbyists who seem to have some kind of David vs Goliath point to make. It's to their credit, in a sense, and I don't mind it at all when they've a business model -- they're in business to stay and good for them if they're good at what they do.

            But most don't have such a business plan. They learn the hard way that it's a lot

    • by Zerth (26112) on Sunday April 22, 2012 @05:04PM (#39765513)

      I wonder if they included the idea that frequently it is the copycat that takes off, while the originator languishes in obscurity.

      • by gstrickler (920733) on Sunday April 22, 2012 @05:23PM (#39765659)

        That is why I refuse to use any Zynga owned game. Even if they purchased the original rather than copy it, they'll use the profits to copy others.

        • Re: (Score:2, Interesting)

          by tepples (727027)
          Do you also refuse to use Linux because it copied UNIX? Or Windows because it copied the Mac? Or Mac because it copied Xerox?
          • by gstrickler (920733) on Sunday April 22, 2012 @06:34PM (#39766107)

            Are you just trolling or do you not know about Zynga's practices [wikipedia.org]? Google them, they copy most of their game ideas, while filing suit [techcrunch.com] against anyone who copies one of their game ideas. They're completely amoral, or worse.

            • I just wanted to know how strongly you clung to the "don't patronize reimplementors" mentality. The Tetris keiretsu has similar practices: copying someone else's concept to make Feevo while suing anyone else who sells a falling tetromino game. The problem here is that casual games tend to have so few core rules that the line between making an original game in the same subgenre and misappropriation of original expression is harder to discern. What do you think of Quadrapassel?
              • Never heard of it before. Had to google it. Ok, a Tetris knock-off 20+ years later? Tetris itself isn't even available for any modern platform.

                As to your earlier questions about Linux and Mac OS X, they're not relevant. Unix [wikipedia.org] was widely available as source code and encouraged alternative implementations [wikipedia.org]. Mac OS X is derived from Mach, a BSD [wikipedia.org] variant. And Linux [wikipedia.org] started as a clean room implementation of MINIX [wikipedia.org], itself an independent implementation of the original AT&T Unix spec.

                • by shmlco (594907)

                  "Mac OS X is derived from Mach, a BSD [wikipedia.org] variant."

                  And that's what comes from getting al of your "knowledge" from Wikipedia. First, reread for comprehension. He didn't say Mac OS X, he said Mac OS, as in the original 1984 Mac OS that was NOT based in any way, shape, or form on Unix.

                  Mac OS 1.0 was based on original work and UI work done on the Lisa OS, and both drew heavily from work licensed from Xerox's work on the Alto (which Xerox later tried to market as the Star).

                  The original Windows code w

                  • Re: (Score:3, Informative)

                    by gstrickler (920733)

                    Yes, Mac OS/LISA were inspired by what they saw at Xerox PARC. However, if you've read up on those, the people who actually worked at PARC said that LISA/Mac went far beyond what they had done or envisioned. [wikipedia.org]

                    "the Apple work extended PARC's considerably, adding manipulatable icons, and drag&drop manipulation of objects in the file system (see Macintosh Finder) for example. A list of the improvements made by Apple, beyond the PARC interface, can be read at Folklore.org. [folklore.org]"

                    My knowledge doesn't come from Wikipedia, I've been in this industry for 30+ years. I just use Wikipedia to support my statements.

                    • by shmlco (594907)

                      So? I've been in it since 1972 (40 years). Appeal to authority won't work here.

                      Changing the subject won't help either. The OP's comment was about using one thing because it was copied or based on another. Windows/Mac, Mac/Alto.

                      And if you're going to provide links to articles, you might avoid cherry-picking facts to prove your point, "There is still some controversy over the amount of influence that Xerox's PARC work, as opposed to previous academic research, had on the GUIs of the Apple Lisa and Macintosh,

                    • You wrote: And that's what comes from getting al of your "knowledge" from Wikipedia.

                      I replied: My knowledge doesn't come from Wikipedia, I've been in this industry for 30+ years. I just use Wikipedia to support my statements.

                      I didn't use an appeal to authority, you need to check your attitude and your definitions. A direct reply to your statement is not an "appeal to authority", I provided a fact that contradicts your statement.

                      I didn't cherry pick facts to prove my point. The facts I cited are sufficient to demonstrate that the Mac was not just a copy of the ideas from PARC, as implied by previous comments. Go back and re-read the entire thread. I never denied that the Mac was influenced by Xerox Alto, in fact, I began my resp

                    • by shmlco (594907)

                      "I didn't use an appeal to authority."

                      Right, "My knowledge doesn't come from Wikipedia, I've been in this industry for 30+ years. " Which sets yourself up as the authority.

                      "Now, go annoy someone else."

                      Nah.

                    • No, it establishes that my knowledge comes from experience, and that I cite wikipedia as backing, just like I stated. I made no claims about my expertise or authority, only that my knowledge comes from experience, not from Wikipedia as you asserted. If I wanted to set myself up as the authority, I would have referred you to my credentials.

                • by tepples (727027)

                  Ok, a Tetris knock-off 20+ years later?

                  The exclusive right that The Tetris Company claims has a 95 year term, not 20.

                  Tetris itself isn't even available for any modern platform.

                  TetrisFriends.com disagrees with you, as do the official Tetris app by EA for iOS and Tetris Axis for 3DS.

                  Mac OS X

                  What about Mac OS I, on the 128K Mac?

      • by Anonymous Coward

        Yeah but often the reason that happens is because of slight improvements in the new copy. And often those improvements were requested and ignored or dismissed from the original developer.

      • I wonder if they included the idea that frequently it is the copycat that takes off, while the originator languishes in obscurity.

        First Mover Disadvantage is usually considered in simulations like this.

      • by timeOday (582209)
        That idea isn't quite concrete enough to implement directly, but the tendency is always to tweak model factors until a known or suspected outcome is reached, in this case "fast follower." This is unavoidable at some level, since otherwise it wouldn't be a "model" - just a bunch of random dynamics. But the value a model must be judged by novel and correct predictions it makes, and I'm not sure there's really a "prediction" here.
  • Interesting (Score:5, Funny)

    by Anonymous Coward on Sunday April 22, 2012 @04:53PM (#39765421)

    In my simulation the best strategy was to take 30% of everyone's revenues.

    • Re: (Score:2, Insightful)

      by Anonymous Coward

      Your study is flawed. For many App Store developers, Apple gets much more than a mere 30% of the developer's revenue.

      To develop and market a $0.99 or $1.99 app, the developer needs to drop many thousands of dollars on various pieces of Apple hardware, and then there's the annual fee that needs to be paid, too. And that's all in addition to the 30% you mention.

      For every developer who does turn an actual profit, there are likely many thousands who suffer very significant losses, even if they try to deny it. Y

      • You should log in next time so we can help you with your sense of humor.

      • Re:Interesting (Score:4, Insightful)

        by gstrickler (920733) on Sunday April 22, 2012 @05:26PM (#39765675)

        That's no different than any other development model, or most other businesses. Starting a business is a risk, welcome to reality.

        • by Anonymous Coward

          His point was it wasn't worth the risk. Not that all businesses don't have risk. The majority of people lose money. I sell free software. I make money. We develop little to nothing though. However we do fund it. Many of the developers whom write this free software are being rewarded. I rely on that software for which without it my business model doesn't work. Despite many competitors I have very little real competition. We are cornering the market in a sense without preventing competition. I probably make m

        • by AmiMoJo (196126)

          It's different to Android.

          • Not at all. To develop for Android, you still have to own/buy a computer, spend your time developing the software, and pay someone a commission to sell it. And you have no more guarantee that the program will sell. It's the same basic model and risk either way.

      • Re: (Score:2, Informative)

        by spire3661 (1038968)
        Mac mini $599 ($519 refurb), Ipad 2 $399 ($350 refurb), $99/year dev license x 2 (mac and iOS). By my count its $1000 up front and $200/year. That is an astonishingly low price for a pretty feature rich dev system. Add $120 for 16 GB RAM upgrade (DIY).
      • To develop and market a $0.99 or $1.99 app, the developer needs to drop many thousands of dollars on various pieces of Apple hardware, and then there's the annual fee that needs to be paid, too.

        You need a computer anyway, so that cost is zero as far as factoring in how much it affects the amount you do/do not take in. And a used mac, at under $500 would do fine for iOS development.

        The developer fee is just $100/year...

        So all around you have no idea what you are talking about.

        • by tepples (727027)

          You need a computer anyway

          You say this as if one doesn't already own a computer. The computer that's already on your desk and paid for won't work.

          so that cost is zero

          No, it's $200 for a second operating system on top of the cost of a Mac if you want to run the software that you already have.

          The developer fee is just $100/year

          "Just"?

          • You say this as if one doesn't already own a computer. The computer that's already on your desk and paid for won't work.

            For most developers it will. Been to any technical conferences or colleges lately?

            No, it's $200 for a second operating system on top of the cost of a Mac if you want to run the software that you already have.

            Now you are making up $200 out of thin air.

            "Just"?

            Yep. That's a pretty small price for the resources you get, including all WWDC videos. It's low enough it blocks no-one that can a

            • Been to any technical conferences or colleges lately?

              Not lately, but when I attended Rose-Hulman, I don't think I saw more than six Macs in dorms. Every student had the school-issued laptop, and it ran Windows.

              Now you are making up $200 out of thin air.

              MSRP for Windows 7 retail. (The OEM version isn't for Macs [microsoft.com].)

              • Not lately, but when I attended Rose-HulmanMSRP for Windows 7 retail. (The OEM version isn't for Macs.)

                You had to go full retard...

                Why would you buy that?

                That would assume you have Windows applications already you'd like to run. Which means YOU ALREADY OWN A WINDOWS BOX YOU CAN KEEP USING!

                Retard.

                The secondary level of mental damage you exhibit, the one that has you going FULL retard, is knowing you CAN run OEM Windows 7 copies under virtualization but insisting people know or care what the license says.

                • Which means YOU ALREADY OWN A WINDOWS BOX YOU CAN KEEP USING!

                  Then please explain what you meant by the sentence "You need a computer anyway" in your previous comment [slashdot.org]. If we take for granted that one already owns a Windows box and is not selling it to afford the Mac, then the Mac is bought to run one application. Such an expense is perfectly justifiable for a day job, I'll grant, but tougher to explain to one's SO for a hobby-turned-business built with sweat equity in one's spare time. I will further grant that there is one situation where already owning a Mac is lik

            • by blacklint (985235)

              Let's say I want to buy a copy of Windows, because it doesn't come with my Mac. I'm not even going to splurge for Ultimate, and settle for Windows 7 Home Premium. Newegg [newegg.com] has it for "$189.99 was: $199.99". Yes, you can get OEM versions for half that, but saying $200 for a Windows license is certainly not out of thin air.

              • Let's say I want to buy a copy of Windows, because it doesn't come with my Mac.

                At this point there are very few people indeed that require a Windows license. There is almost no software you cannot have now on the Mac, and as noted at this point most students get macs as first computers anyway so migration is really moot.

          • by sribe (304414)

            No, it's $200 for a second operating system on top of the cost of a Mac if you want to run the software that you already have.

            No it's not. It's at most $50 for VMWare, and then you use VMWare Converter to create a VM image from your current Windows OS install. Or, if you're an actual serious Windows developer, you're already running virtual machines anyway to control your environment, and you just copy them to your Mac--even better if they're on external drives ;-)

            • It's at most $50 for VMWare, and then you use VMWare Converter to create a VM image from your current Windows OS install.

              Provided that your existing Windows OS install is retail, not OEM. If it's OEM, the license is not transferable to your Mac's motherboard.

  • In the two sample runs they show, the Innovator does well in one and the "Milker" with multiple redundant apps does well in the other. The "Optimizer" who improves their best app comes in second in both, and I'd wonder if that holds over a larger set of simulations.

    I suspect that what might be more interesting is the standard deviation of ending positions over many runs.
    • by ceoyoyo (59147)

      A graph without error bars is meaningless.

      • by Fencepost (107992)
        In this case they're showing graphs of values from two separate runs. This is fine, there's no error information to be shown, it's not aggregated numbers as in a poll, etc. I'm just suggesting that if their results are that variable (even assuming the two graphs shown are extremes) then aggregated data (with standard deviations shown) might be more useful.
        • by ceoyoyo (59147)

          As I said, a graph without error bars is meaningless. From two runs they get different results so that tells us that there is variation but doesn't give us enough information to properly quantify it (to make error bars). So their graph is meaningless. Do the copy cats really do better, or was it just on one of the runs and they almost never survive? We don't know. Their results, as presented, are meaningless.

          They obviously did run multiple simulation runs, and their paper may well quantify the variatio

"In matters of principle, stand like a rock; in matters of taste, swim with the current." -- Thomas Jefferson

Working...