Developing Perspective

#177: Something in Mind.


00:00:00   Hello, and welcome to Developing Perspective. Developing Perspective is a podcast discussing

00:00:04   news of note in iOS development, Apple and the like. I'm your host, David Smith. I'm

00:00:09   an independent iOS developer based in her new Virginia. This is show number 177. And

00:00:14   today is Thursday, March 13th. Developing Perspective is never longer than 15 minutes.

00:00:18   So let's get started. All right, so I'm going to be continuing on my in my series about

00:00:25   how towards a better app store things and thoughts and ideas that I have about things

00:00:28   that could actually practically make the App Store a better place.

00:00:31   And I want to first off say thank you to all the people who've been giving me feedback

00:00:34   on this.

00:00:35   I've been getting a lot of great emails and responses to that.

00:00:38   And so I really appreciate kind of those thoughts.

00:00:40   It's helpful for me to hear either things that I agree with, things that I disagree

00:00:46   with, whatever it is, just things that help me to kind of formulate these thoughts I've

00:00:50   been developing over the last little while.

00:00:51   And so if you have other thoughts, either responses to things I've said or things that

00:00:55   you've always thought, if you had this great idea that you could apply to the App Store

00:00:58   that would actually make it better. This is a great time to reach out to me and I really

00:01:01   appreciate that. Another side note is that next week I will be at NS Conference in Leicester

00:01:08   in England. If you are a listener to Developing Perspective and will be there and hear this

00:01:12   before then, please come out and make sure you let me know. I'd love to meet you. I always

00:01:16   love to hear from people who enjoy the show. It's always kind of fun to hear, you know,

00:01:20   how you've, the most interesting to me usually is where you picked it up, how you found out

00:01:24   about the show and kind of what you think about it. So if you listen to the show and

00:01:27   and will be at NS Conference this next week.

00:01:30   Make sure you say hi.

00:01:31   I'd really appreciate it.

00:01:32   All right.

00:01:32   So otherwise, I'm going to dive into the topic of today's show,

00:01:35   and that is going to be search.

00:01:37   And search is a pretty important thing,

00:01:39   because if you consider the fact,

00:01:41   like I was building on last week's show

00:01:43   where I was talking about how there's just so many, so many,

00:01:46   so many apps in the App Store, if it wasn't for search,

00:01:51   the vast majority of them would be completely irrelevant,

00:01:53   because they'd only be available unless they

00:01:55   were-- if they were featured or if they were ringing incredibly

00:01:57   highly, which is a pretty small subset of the store.

00:02:00   So search is the important thing that allows you to find

00:02:04   an application that isn't one of those two things, isn't

00:02:06   being currently featured, or isn't currently

00:02:09   ranking very highly.

00:02:10   So isn't doing high volume or doing high grossing.

00:02:13   And so that probably is, by definition, most apps.

00:02:18   And so the way that you implement search is vitally

00:02:20   important for helping customers find the specific

00:02:23   app that they're looking for.

00:02:26   And at the same time, search is also really one of the most

00:02:32   gameable, in some ways, attribute of the App Store,

00:02:35   which is a little bit tragic.

00:02:36   In some ways, you could argue maybe that Grossing or the

00:02:39   Chopped Charts are a bit manipulative and sensitive.

00:02:42   But the reality of that is it takes a lot of work to

00:02:44   maintain kind of scamming or working on the App Store

00:02:47   rankings, because you're having to actually have

00:02:49   downloads to sort of put to that.

00:02:51   But search is a bit more subtle, because it's the

00:02:54   algorithm that's determining the ranks in searches is something that's probably a bit

00:02:58   slower moving and is a bit easier to kind of predict or work with. And as a result,

00:03:04   search is something that I think has honestly been very, you know, has been attacked by

00:03:07   developers good and bad for years. I mean, back in the old days, you even used to be

00:03:12   able to, you know, completely keyword spam your app or description and it would improve

00:03:16   your search ranking. And the reality is, the reason people do that is because search is

00:03:21   so important. There's, you know, for any number of ways, I would say my gut says, you know,

00:03:26   at least a third of app downloads are coming from searches, not just top charts and featured.

00:03:31   And that's just a gut, but I know in my own experience that's often the case because search

00:03:35   is where you go when you know what you're looking for. Search is where you go when you

00:03:39   have something in particular in mind. The top charts and the featured lists are things

00:03:44   that you browse, whereas search is somewhere you go to actively find something. It's you've

00:03:49   You've heard a name of an app.

00:03:50   You've heard about an app called Podometer++, say.

00:03:54   And you go and say, well, let's go find that app.

00:03:57   The unfortunate thing right now is if you go into the App Store

00:03:59   and search Podometer++ in the store,

00:04:02   the fourth record is the app you're looking for.

00:04:04   Or sorry, not the fourth.

00:04:05   The 14th record is the app you're looking for.

00:04:08   The 13 before that are not the app you're looking for.

00:04:10   There are a variety of podometers, many of which

00:04:12   are great apps.

00:04:13   But unfortunately, it's not exactly the one

00:04:16   that you're looking for.

00:04:17   And so that's where search can get kind of complicated.

00:04:19   Because search is one of the few things

00:04:20   that the user actually has something in mind

00:04:22   that they're trying to drive towards.

00:04:25   So what can we do about that?

00:04:28   How can we make search better?

00:04:30   And there's various recommendations

00:04:32   that I have that fall into a variety of categories.

00:04:34   But the first one is that the interface that the App Store

00:04:38   has-- and this came in iOS 7-- is kind of terrible.

00:04:41   And I'm talking about the cards, where

00:04:43   So you have these massive, giant screenshot-oriented search

00:04:49   results, where it shows one app at a time.

00:04:51   And I understand kind of what they're doing there.

00:04:53   They're making sure that the screenshots

00:04:55   are shown full and bright and obvious to the user.

00:04:59   But at the same time, it makes it

00:05:01   very hard if your app isn't the first search

00:05:04   result for something, or the first one or two search results,

00:05:07   to actually ever be seen.

00:05:08   In the old system, when you'd search,

00:05:10   you'd get a list of apps with the icons on the left

00:05:12   the names on the right, and maybe the rating I think was shown. What that does is that

00:05:16   gives you the ability to much more quickly kind of browse through a variety of choices

00:05:21   rather than just having to go one by one. It's kind of like a slideshow. And so I think

00:05:26   I have a lot, but I would perp, as maybe the broad start off on saying is, I think it'd

00:05:31   be great if Apple really, really loves that card's approach and is completely wedded to

00:05:35   it, is to also have the option that you could toggle a button on the top or something to

00:05:39   to have list view or grid view, essentially.

00:05:41   Like you often see on a lot of websites.

00:05:43   Or you could still have that list view available,

00:05:45   if that's what the customer wanted,

00:05:47   because they're trying to find something very specific.

00:05:50   Next, it's kind of-- the next recommendation

00:05:52   is a little bit squishier, but I think

00:05:54   it's still pretty practical.

00:05:55   In that, one of the biggest problems

00:05:58   I think I have with search now is

00:05:59   that it's very odd to try and determine

00:06:03   what is defining the relevance or the ranking of the search

00:06:06   results.

00:06:07   And often, the results you get back

00:06:09   aren't things that are necessarily

00:06:12   the best apps for the customer.

00:06:13   They're the apps that are often things

00:06:15   like, for example, the oldest app-- that meets that criteria--

00:06:18   the app with the most downloads, which in some ways

00:06:21   is a good thing.

00:06:22   But there doesn't seem to be a lot of, I guess you could call

00:06:24   it, punishment or reward for things that are objective,

00:06:30   would objectively make an app better.

00:06:31   It's pretty tricky.

00:06:32   And while I'd love to see some types of manual curation, which

00:06:35   is sort of the next thing that I'll talk about,

00:06:38   One thing that I'd like to see is even just

00:06:40   for just the objective things, the things that are very clear

00:06:43   and obvious.

00:06:44   For example, if an app was last updated two years ago,

00:06:48   that should probably not be the first search result.

00:06:50   If there is an app that meets that search criteria,

00:06:52   that was updated six months ago or six weeks ago.

00:06:55   There are objective measures that we can look for that.

00:06:58   There's things that you could base on, not just downloads,

00:07:01   but returns, number of crashes even.

00:07:05   I was realizing recently that iOS and iTunes Connect

00:07:08   has access to all the crash logs.

00:07:10   And so they can probably see which apps crash the most.

00:07:12   There's a lot of these objective things

00:07:14   that you could measure and weight

00:07:15   and that I think would be far better in terms

00:07:17   of improving the relevance of that ranking

00:07:20   to try and find what's the best app for this particular search

00:07:24   criteria.

00:07:25   And then piggybacking on that-- and this

00:07:27   is what I mentioned a moment ago--

00:07:28   is I'd love to see a bit more actual curation about this.

00:07:31   And curation is a bit complicated.

00:07:33   But my suspicion is that there's a very steep drop-off in terms of the number of search

00:07:40   terms that are used a lot, and there are a number of search terms that are used very

00:07:44   infrequently. The number of people who are searching, even for argument's sake, who

00:07:50   are searching for Facebook is probably much, much higher than people who are searching

00:07:55   for, I don't even know, apps to map tunnels. I don't even know where that came from. But

00:08:02   But you know what I mean.

00:08:03   Like I'm trying to understand that I'm sure there are probably maybe a dozen, two dozen,

00:08:09   five dozen, whatever it is, maybe less, probably maybe a hundred.

00:08:12   Let's call it a hundred.

00:08:13   There's probably, if you took the top 100 search terms that people look for in the App

00:08:16   Store and you look at them, the search results for those terms seems like they should have

00:08:22   some amount of curation applied to it.

00:08:25   And the canonical example for that is to do something like if you search for Twitter right

00:08:29   now, or even maybe take it a little stronger, you search for Twitter client. Right now,

00:08:34   you don't get the things that you would sort of expect to get back for that. You'll get

00:08:37   all kinds of junk and other stuff around that. And that has to do with a lot of like Tweetbot

00:08:41   and Tweet and Twitterrific, for example, two great Twitter clients, which have done non-paid

00:08:47   updates. And so they kind of have new had to restart all of their search mojo, which

00:08:51   is a whole other thing. But it doesn't really make sense. And maybe that goes back to the

00:08:55   previous point about relevance. But it's one of those areas where like there should be

00:08:59   be a little bit of manual curation.

00:09:01   At least for those very popular search terms,

00:09:03   someone should go through and make sure

00:09:05   that the search results you're getting back make sense.

00:09:08   And if they're not, either adjust your algorithm

00:09:10   so that that's not happening in terms

00:09:12   of it doesn't have to be directly curated

00:09:13   like a featured list.

00:09:14   But I would hope that there's a lot of iteration happening

00:09:17   on there, especially monitoring the most popular terms to try

00:09:21   and make sure that they're good.

00:09:22   And the last thing I wanted to talk about

00:09:25   is a kind of a more general point about search

00:09:27   as it is now just doesn't scale.

00:09:31   And I'm reminded and thinking a little bit about something

00:09:34   like Amazon.com.

00:09:35   And in the last week, I was complaining about how big

00:09:37   the App Store is, how it had millions and millions of items

00:09:40   and how that is causing such trouble for the App Store

00:09:43   for developers and for customers.

00:09:45   And I think of a store like Amazon.com,

00:09:47   which has many million items in its store.

00:09:51   It is a massive marketplace.

00:09:53   You can buy anything you could possibly imagine, but I don't find searching it to be nearly

00:09:58   as complicated.

00:09:59   As an example, this morning I needed to buy a new toaster.

00:10:05   And I know toasters and podcasts, what are you going to do?

00:10:07   But I was going to go buy a toaster and I just wanted to buy a fairly inexpensive slot

00:10:12   toaster that was stainless steel to match the type of appliances I have in my kitchen

00:10:19   and that had four slots.

00:10:21   That was what I wanted.

00:10:22   the toaster I had in mind. And even though I'm starting from a desire to search through,

00:10:27   you know, whatever it is, many million items, it didn't take long to find the toaster I

00:10:30   wanted because you have things called filters. And this is the last thing that I want to

00:10:35   kind of talk about is that I think the App Store is old enough and big enough and mature

00:10:39   enough now that it really does need kind of a little bit of more power search capability,

00:10:44   the ability to add filters onto things, the ability to build more complicated search queries,

00:10:49   that's the experience that I think you people are used to on websites where in this case

00:10:53   with my toaster I type in toaster and then I go to the side and I get a whole bunch of

00:10:57   results who knows even it was probably 10,000 results for toaster then I click I want slot

00:11:01   toasters I want sold by it be a prime eligible I want four slots and I want stainless steel

00:11:09   and then I sort by price look at the top two or three find the one I want and buy it and

00:11:14   And that was a very straightforward operation to search through, you know, thousands of

00:11:18   toasters.

00:11:19   But it's straightforward because I knew what I wanted.

00:11:22   And so giving the users the capability to do those types of filters, to do those kinds

00:11:26   of search is, I think, would be a massive improvement on the state of the App Store.

00:11:30   And it's a little bit tricky, though, and I'll give them that.

00:11:33   Because to do that, you require excellent metadata.

00:11:36   But honestly, that doesn't really bother me too much either, because just demanded from

00:11:40   developers.

00:11:42   Think of criteria that are useful.

00:11:44   of criteria that you could apply to an application, many of them you could just derive directly

00:11:47   from the application. You know, it has been updated in so long, it supports this device,

00:11:52   is useful for front and back camera or whatever it is. Like there's, whenever those criteria

00:11:57   that you could kind of imagine, I mean even the number of years the developer has been

00:12:03   around, average star rating, I mean you could imagine a lot of criteria that you could kind

00:12:07   of apply to it that would help the user to narrow in on what they want. And some of that

00:12:12   metadata you ask users for, some of it you just derive, but either way having those kinds

00:12:16   of things would be really cool. And doing that kind of a thing also gets you to the

00:12:20   last thing I was going to talk about, which is a better or multiple forms of sorting.

00:12:25   Well, like I said before, the current relevance-based sorting that they apply across the board works

00:12:31   all right if you get the algorithm right. But what if for me as a customer I'm looking

00:12:35   for something that's different, I'm searching with a different criteria in mind. Maybe I

00:12:39   Maybe I only want to see free apps.

00:12:41   Maybe I want to see only paid apps, honestly.

00:12:43   Maybe I want to sort from highest price to lowest price.

00:12:46   You hear all the time about people who are, for example, doing things like they're business

00:12:50   customers and they're looking for something that doesn't have in-app purchase.

00:12:53   Maybe that's something that an educational customer doesn't want.

00:12:56   And so being able to say that and say, you know, "Doesn't offer in-app purchase," or "Sort

00:12:59   by number of in-app purchases," or whatever.

00:13:02   There's all of these things that you could apply to your search that I think would make

00:13:05   the App Store experience so much better.

00:13:08   So that's it for today.

00:13:09   is kind of what I'm talking about about search. I think it's an area that they could do some

00:13:13   of the biggest impact. And it would be, in some ways, the smallest amount of work or

00:13:17   the smallest changes from the existing policies and the existing rules and just make search

00:13:22   better. And I know that's easy to say, but a little bit hard to do. But hopefully some

00:13:26   of those things that I just listed out would be practical examples of ways that they could

00:13:29   do that.

00:13:30   And that's it for today's show. The schedule for the next week might be a little bit off

00:13:34   just because, like I said, I'll be traveling to NS Conference in England. And so I'm not

00:13:38   exactly sure when I'll be back at the microphone again.

00:13:41   But like I said, if you're going to be there, please, please

00:13:44   let me know.

00:13:44   I'd love to meet you.

00:13:45   It'd be awesome.

00:13:46   And otherwise, I hope you have a great week.

00:13:49   Happy coding, and safe travels if you're coming to Enniskov.

00:13:52   Bye.