Under the Radar

00:00:00 ◼   ► Welcome to Under the Radar, a show about independent iOS app development. I'm Marco Arment.

00:00:05 ◼   ► And I'm David Smith. Under the Radar is never longer than 30 minutes, so let's get started.

00:00:10 ◼   ► So today, we wanted to talk about, I guess, it's a bit of a hard thing to put into a box,

00:00:16 ◼   ► but I think it's the general topic of continuously changing editorial content. So if there are

00:00:23 ◼   ► things in your app that aren't at all technical features, these aren't things that you are

00:00:29 ◼   ► building in Xcode or writing in technical level, but are things that are nevertheless

00:00:35 ◼   ► changing and updating as a result of you taking action within your app to put content into

00:00:41 ◼   ► your app. And so some examples of this, and this is where this feature is sort of coming

00:00:45 ◼   ► from and why I thought it would be an interesting topic, is I recently added a new thing to

00:00:50 ◼   ► Pedometer++ called monthly challenges. So every month, there is going to be a new little

00:00:55 ◼   ► challenge badge that you can earn by doing something. So in the first month, it was you'd

00:01:02 ◼   ► reach your goal 19 times in January. In February, it's going to be have a 10-day streak. There's

00:01:08 ◼   ► examples of these, and I'm just going to continue to kind of create and manage these going forward.

00:01:13 ◼   ► Another example might be the Discover section in Overcast, where you have a whole bunch

00:01:18 ◼   ► of different links to different shows, but it's a curated list that you have to manage.

00:01:24 ◼   ► Or you have the App Store's Today tab, which is something that they're committing to update

00:01:30 ◼   ► on a regular basis. And this kind of continuously updating content has a lot of interesting

00:01:36 ◼   ► kind of benefits and challenges. And when I first sat down to decide that I was going

00:01:40 ◼   ► to add this monthly challenges thing, I think there were a bunch of aspects of it that I

00:01:44 ◼   ► just hadn't thought about that I'm now having to deal with. That's what made me think that

00:01:48 ◼   ► it would be kind of a useful thing to talk about, because it's exciting and interesting,

00:01:53 ◼   ► and I've had a really positive reaction to it. But now I'm also having to face down all

00:01:57 ◼   ► kinds of challenges and deadlines and things that I had never really had to deal with before.

00:02:02 ◼   ► - Yeah, it's the kind of thing where it can really add a lot of value to an app, to have

00:02:07 ◼   ► some kind of editorial content in there. But I suggest going into it from an indie perspective

00:02:14 ◼   ► very, very carefully. It's one thing if you have a staff, if you have a company that has

00:02:19 ◼   ► multiple people, and you can actually hire someone just to do stuff like this, that's

00:02:25 ◼   ► a very different story. When you're just one person though, committing yourself to like,

00:02:30 ◼   ► "Oh, I'm going to post a new bit of whatever into my app every week or every month or whatever,"

00:02:37 ◼   ► that adds up very quickly. And I had the extreme of this when I ran the magazine, where I had

00:02:42 ◼   ► to actually produce content once a month, because that's the thing people were paying

00:02:47 ◼   ► for with the app. It wasn't an app that happened to have some editorial content that rotated

00:02:51 ◼   ► out every month. It was like, no, literally, we are paying people, or people are paying

00:02:55 ◼   ► to get this, and I believe that for that, if I remember correctly, it was every two

00:02:59 ◼   ► weeks, which was a ton of problems. But to do something every month, say, that really

00:03:06 ◼   ► means every month. And you might think in March, "Oh, that sounds great." But then when

00:03:12 ◼   ► August rolls around and you're on vacation, all of a sudden it's like, "Wait, every month

00:03:17 ◼   ► I have to do this." When December rolls around and you're really busy with the holidays,

00:03:21 ◼   ► no, every month. When January or February roll around and nothing is going on in the

00:03:27 ◼   ► world, every month, like you still have to produce something every month. And so you

00:03:31 ◼   ► really have to be careful when you're committing to something like this, to build a feature

00:03:35 ◼   ► of your app around that people will actually use. You really have to be careful that you're

00:03:39 ◼   ► not biting off more than you can chew, that you can actually deliver on the pattern that

00:03:44 ◼   ► you're starting by building out this feature.

00:03:47 ◼   ► >> Yeah. And on the flip side, it is really cool and compelling in a very different way.

00:03:54 ◼   ► And I think there's definitely a lot of challenges, and I have a whole list of things that I've

00:04:00 ◼   ► run into that will make it hard. But I do think it's also really compelling that I think

00:04:04 ◼   ► it's interesting to add something to our apps that lets us give users a reason to come back

00:04:12 ◼   ► to the app on a regular basis, which is useful just in general, that it creates an engagement

00:04:16 ◼   ► that you just can't get with features. Adding new features to the app can sometimes be useful,

00:04:22 ◼   ► but it doesn't have that same sense of timeliness, that there's this sense that hopefully you

00:04:28 ◼   ► build a sense of trust with your user where they're going to be coming back to your app

00:04:33 ◼   ► on a regular basis because they're expecting something that they like. And if you can keep

00:04:37 ◼   ► delivering that, you've locked them into this great little virtuous cycle where they're

00:04:42 ◼   ► continuing to come back, you create content, and that's very positive.

00:04:46 ◼   ► It also, I think, is nice that it allows you to enhance your app without having to actually

00:04:52 ◼   ► add features or do code, that if you build it in a thoughtful way, you're not necessarily

00:04:59 ◼   ► going into Xcode every month or every week or whatever you're adding this kind of content.

00:05:03 ◼   ► You're potentially just publishing a P list to a website or making a change like that

00:05:09 ◼   ► that can then reflect in the app, which is also nice because then potentially you can

00:05:15 ◼   ► have people working on creating this content who aren't you, who are less technical, whether

00:05:22 ◼   ► that be someone you outsource this to or even just your spouse or a friend or something

00:05:28 ◼   ► like that that doesn't have to be technical in the way that I want to add a feature to

00:05:34 ◼   ► my app. You have to really get into that.

00:05:37 ◼   ► It's kind of fun that you can potentially highlight different features in the app that

00:05:43 ◼   ► people may not be aware of by drawing their attention to it because you have this mechanism

00:05:49 ◼   ► now where you're directly communicating with your customers in a non-just like here's

00:05:56 ◼   ► the app and here is its features. It has more of a voice to it.

00:06:01 ◼   ► For example, if I want to make sure that people realize that Pedometer++ can do floor counting

00:06:06 ◼   ► if you have a device that supports that, then I can have a challenge that includes floor

00:06:10 ◼   ► counting. If you're not really sure what that is, it's a prompt to go and do it.

00:06:16 ◼   ► Lots of challenges, lots of commitments. The reason I wanted to do the show today is because

00:06:23 ◼   ► I was sitting down and I built the feature initially for my monthly challenges and it

00:06:29 ◼   ► just had January in it. It didn't have a mechanism for adding future months or anything because

00:06:34 ◼   ► I just wanted to have it in by January 1st because that seemed like the really opportune

00:06:39 ◼   ► time to launch a feature like this. Then I got to the middle of January and I got a bit

00:06:44 ◼   ► sick and I had to take a week off working. Then I was like, "Oh, no. February 1st, this

00:06:48 ◼   ► needs to be in the App Store. I need to have this thing in the App Store by February 1st,"

00:06:53 ◼   ► which means that it needs to be submitted by the last week of January at the absolute

00:06:58 ◼   ► latest and ideally a bit before that so it can propagate out to the App Store and everyone

00:07:03 ◼   ► will have it. That pressure is not great.

00:07:07 ◼   ► - Yeah, but you're right though. It does really add a lot of potential upside. That's why

00:07:14 ◼   ► these features are so compelling because not only are feature editions by themselves harder

00:07:20 ◼   ► to get people to care about, but they're also just less effective at boosting engagement

00:07:24 ◼   ► in your app. Lots of people install apps, play with them for a little bit, and then

00:07:29 ◼   ► don't go back. This is one of the reasons why push notification spam is so prevalent

00:07:34 ◼   ► because it works and it solves a necessary problem, which is how do you get people to

00:07:40 ◼   ► come back to your app when they might have forgotten about it or gotten bored with it

00:07:44 ◼   ► or gotten tired of it? Most of the time they really just have forgotten about it. They've

00:07:49 ◼   ► moved on to other things.

00:07:51 ◼   ► There are ways to do this that are really annoying and that I really don't like and

00:07:55 ◼   ► I think are morally questionable and that definitely violate the least enforced App

00:08:01 ◼   ► Store guideline, which is the one against push notifications for marketing. There is

00:08:07 ◼   ► ultimately a lot of value in giving something that people actually want as notification

00:08:13 ◼   ► to say, "A new challenge just arrived," or something. Again, it's a really, really fine

00:08:19 ◼   ► line. If you're going to do this ethically, it's important that people know what they're

00:08:24 ◼   ► opting into when they turn on notifications for your app and that you give them very,

00:08:28 ◼   ► very good fine-grained controls to turn those off without disabling all notifications from

00:08:33 ◼   ► your app.

00:08:35 ◼   ► There are ways to do this well. I've thought before about things like with Overcast, how

00:08:40 ◼   ► maybe I could make, when you get a new recommendation for the Twitter recommendations feature, maybe

00:08:46 ◼   ► you can get a push notification for that optionally. There's all sorts of things that I could do

00:08:49 ◼   ► like that. It really is important to have some reason for people to return to your app

00:08:57 ◼   ► when they haven't been using it for a little while. There's lots of ways to do this badly

00:09:02 ◼   ► or questionably or sleazily, but if you can find a way to do it well that people don't

00:09:09 ◼   ► get turned off by and that they have fine-grained controls over disabling if they don't like

00:09:13 ◼   ► it, I don't think there's anything wrong with that. That's just not usually the way it's

00:09:17 ◼   ► done.

00:09:18 ◼   ► >> The thing about that too is there's the communication outside of the app. I think

00:09:24 ◼   ► there's even moreover the interesting things that you can do around communicating within

00:09:28 ◼   ► the app. How do you let people know that there's new things for them to see? The classic examples

00:09:35 ◼   ► of this is any kind of social media app where obviously the content isn't coming from the

00:09:41 ◼   ► app itself but it's from other users. You have little badges on the tab bars or all

00:09:46 ◼   ► kinds of indicators and things to be like, "Hey, there's something cool and interesting

00:09:50 ◼   ► over here. Come and look at it," and to create that sense of there's something new for you

00:09:56 ◼   ► to see.

00:09:58 ◼   ► That's another interesting way. Externally pulling people in because of content you're

00:10:03 ◼   ► creating, I would be very reluctant to get into that world because you're very quickly

00:10:10 ◼   ► becoming intrusive in a way that I tend to shy away from. It's possible potentially to

00:10:15 ◼   ► do it correctly. If someone has opted in for it, then maybe there's a case for it. There's

00:10:22 ◼   ► certainly a case for doing that within your app. Even to the extent for the first time

00:10:28 ◼   ► I think I've ever done it in Panometer++, I had a little pointy arrow, there's a news

00:10:35 ◼   ► bubble that popped up just one time this month to indicate to tell people that there's a

00:10:42 ◼   ► new feature in the badges section that's called monthly challenges. I wanted to let people

00:10:49 ◼   ► know because it was a time-based thing that they had to do something with that I didn't

00:10:53 ◼   ► want to have the situation where you go to the badge section only every now and then

00:10:58 ◼   ► probably because you're not earning badges every day and then you discover, "Hey, there's

00:11:03 ◼   ► only 17 days left in January and I need to hit my goal 19 times, I'm stuck." That'd be

00:11:09 ◼   ► frustrating. I did a little thing, it's like, "Hey, there's something here, take a look.

00:11:15 ◼   ► There's new monthly badges," which I usually tend to shy away from a lot of those kinds

00:11:20 ◼   ► of things and make things more organic. But I think when you get into this world where

00:11:25 ◼   ► you have this kind of editorial content that is potentially time-based, and in some ways,

00:11:32 ◼   ► the weird thing is that the more time-based it is, the more compelling or valuable it

00:11:38 ◼   ► may actually be. Sometimes the frequency or the timeliness of it is what makes it interesting,

00:11:46 ◼   ► that you have a sense of, "This is something that is useful and interesting right now."

00:11:51 ◼   ► In my case, it's like, "This is the challenge for this month and you have to earn it this

00:11:54 ◼   ► month. You can't wait and do it later." Or in another example, a content that was something

00:12:00 ◼   ► that was dealing with a live event or performance or that type of thing where there's a sense

00:12:06 ◼   ► of urgency about it is good in that it creates that engagement, but it's definitely this

00:12:12 ◼   ► funny slippery slope of making sure that you're not pushing it too strongly, but you're taking

00:12:18 ◼   ► advantage of the benefits that it can have for you.

00:12:21 ◼   ► - Right, because if you start pushing that button too often, you have to be careful that

00:12:27 ◼   ► if you're going to basically create a button that you can push that increases engagement

00:12:32 ◼   ► of your app immediately, that's a very dangerous thing to create because you basically create

00:12:38 ◼   ► a very strong incentive for yourself to push that button more and more often over time.

00:12:45 ◼   ► Notifying people of challenges or of new content or whatever else to try to get them to start

00:12:49 ◼   ► using your app again, it's a form of spending political capital. If you push that button,

00:12:57 ◼   ► you're costing yourself something. Every time you push that button, you're going to annoy

00:13:01 ◼   ► a certain percentage of those people. Certain percentage of those people are going to think

00:13:05 ◼   ► less of you, think less of your app and be more likely to disable notifications or to

00:13:08 ◼   ► delete the app entirely. It's an expensive action to do frequently. For what you're talking

00:13:16 ◼   ► about, having monthly challenge, new monthly challenge, come to the app and do this, that

00:13:20 ◼   ► I think is fine. A monthly notification or a monthly special thing in the app is not

00:13:25 ◼   ► unreasonable for almost any app that I can think of, but a lot of apps don't do that.

00:13:30 ◼   ► A lot of apps start doing it daily, and then that becomes a bigger problem. You have to

00:13:36 ◼   ► be sure that when you're creating a mechanism that you can basically push a button whenever

00:13:40 ◼   ► you'd like and benefit from that immediately, the temptation starts becoming very great

00:13:46 ◼   ► and you have to exercise a lot of self-control to not abuse that to the point where it destroys

00:13:51 ◼   ► it.

00:13:52 ◼   ► >> The thing too, it's that awkward thing where you want to make sure that you are still

00:14:00 ◼   ► keeping the core experience of your app at the forefront of what it is that you're still

00:14:08 ◼   ► presenting the user. I think it is easy also for us, the farther we get from the first

00:14:17 ◼   ► time we made the app and the initial way that you might use it, I think it's easy to start

00:14:22 ◼   ► to imagine that all of your users are as familiar with all of the features, that are as engaged

00:14:28 ◼   ► and excited and thrilled by your app as you are, when they may actually just still be

00:14:33 ◼   ► in the stage where they're getting used to the core features of it, like just the basic

00:14:37 ◼   ► utility of the application. If you start to barrage them with, "Hey, have you done this?

00:14:43 ◼   ► What about this? This thing over here, you should check this out," there's a way in which

00:14:48 ◼   ► that may just turn them off and it may be too much too soon. It's not actually prompting

00:14:54 ◼   ► engagement. If anything, it's pushing someone away because it's like, "Whoa, whoa, what

00:14:59 ◼   ► am I getting into here? I just want to count my steps, man. I just want to listen to a

00:15:05 ◼   ► podcast. I don't need all this coming at me all the time." I think it's also important

00:15:10 ◼   ► to just remember that you're not, in many ways, a lot of these types of things are maybe

00:15:17 ◼   ► starting to gear towards your more experienced users, so making sure that you're keeping

00:15:21 ◼   ► that in mind that the initial experience. You could imagine the worst version of this

00:15:27 ◼   ► would be the user installs the app and then, say, a naive implementation of monthly challenges

00:15:34 ◼   ► is like it pops up and it's like, "Hey, here's the challenge for October. Hey, here's the

00:15:38 ◼   ► challenge for September. Hey, here's the challenge for August. Hey, here's the challenge for

00:15:42 ◼   ► July." You could imagine a naive, simple implementation where it's just like the first time you load

00:15:48 ◼   ► the app up, you're like, "Hey, here's all this great stuff that happened before you

00:15:51 ◼   ► installed the app," and it's completely overwhelming and not actually useful. I think it's important

00:15:57 ◼   ► to keep that kind of thing in mind too where it still has to make sense for someone who

00:16:02 ◼   ► is just first coming to the app who isn't in the place of like if you have an editorial

00:16:07 ◼   ► section to help them discover new content. Say you're a discover section in Overcast.

00:16:12 ◼   ► That's great in terms of as a new user, it's useful to go in there and like I don't know

00:16:18 ◼   ► any podcasts. Help me find some. If it gets too complicated or constantly updating or

00:16:26 ◼   ► could potentially become just too overwhelming, then you have to be careful that you're not

00:16:30 ◼   ► crossing that line too.

00:16:32 ◼   ► We are brought to you this week by Linode. With Linode, you can instantly deploy and

00:16:36 ◼   ► manage an SSD server in the Linode cloud. You can get a server running in seconds with

00:16:41 ◼   ► your choice of Linux distro, resources, and node location. Linode serves their customers

00:16:46 ◼   ► with the help of 10 data centers around the globe so far, and they're about to add more.

00:16:51 ◼   ► Mumbai, India, and Canada will all have data centers before 2020. Linode features native

00:16:57 ◼   ► SSD storage, a 40 gigabit network, and Intel Xeon E5 processors. This is really high-end

00:17:03 ◼   ► gear so you can serve your customers even faster than before. And you don't have to

00:17:08 ◼   ► stress about overspending. Linode has designed their pricing tiers to feature hourly billing

00:17:13 ◼   ► with monthly caps on all plans and add-on services, including backups and node balancers.

00:17:18 ◼   ► So check out Linode today. They have pricing options to suit everyone. Their plans start

00:17:22 ◼   ► at one gig of RAM for just $5 a month, and they have lots of plans above that for whatever

00:17:27 ◼   ► your needs might be, including high memory plans. And Linode has a special offer for

00:17:32 ◼   ► all of us. As a listener of this show, you can go to linode.com/undertheradar. That's

00:17:37 ◼   ► under the radar with hyphens, under dash the dash radar, and use promo code radar2019 to

00:17:44 ◼   ► get $20 towards any Linode plan. On the one gig plan, that could be four free months.

00:17:50 ◼   ► And with a seven day money back guarantee, there's nothing to lose. So go to linode.com/undertheradar

00:17:57 ◼   ► and use promo code radar2019 to learn more, sign up, and make the most of that $20 credit.

00:18:03 ◼   ► Thank you to Linode for hosting all of my stuff that I host on the web, really, and

00:18:07 ◼   ► all of Relay FM.

00:18:09 ◼   ► >> All right, so another aspect I think that has emerged as I've dived into making this

00:18:14 ◼   ► kind of content that is kind of a little bit subtle, potentially, but is the nature by

00:18:19 ◼   ► which you want to make content that is broadly appealing for your whole audience, or at least

00:18:27 ◼   ► that isn't going to be turning off portions of your audience or being problematic for

00:18:31 ◼   ► them or whatever it is. And this can take a variety of different issues. But the issues

00:18:37 ◼   ► that I run into are kind of the things that are specific to a fitness based application,

00:18:43 ◼   ► where people are going to be coming to my app with a variety of different fitness goals,

00:18:49 ◼   ► a variety of different capabilities physically, and it's trying to make something that is

00:18:55 ◼   ► appealing and relevant to them. And, you know, for example, like I have users of pedometer++

00:19:01 ◼   ► who are in wheelchairs, and so having a goal that is entirely based on steps or based on

00:19:07 ◼   ► climbing stairs or things like that would become -- you know, it would be rather problematic

00:19:12 ◼   ► and not a good thing. Or even if I just make the goal be something that is just fixed at

00:19:19 ◼   ► a certain capability, you know, say, hey, I'm going to set a goal for this month that

00:19:23 ◼   ► you have three days where you walk 20,000 steps. That goal might be reasonable for a

00:19:34 ◼   ► typical able bodied person, potentially, but it's maybe something that is completely out

00:19:40 ◼   ► of reach for someone. And for them, like, getting 2 or 3,000 steps a day is the big

00:19:45 ◼   ► reach goal. Like, they've really hit it out of the park, and that's a super big accomplishment

00:19:49 ◼   ► for them. And so it's something that I've kind of had to think through and be very thoughtful

00:19:54 ◼   ► about is making sure that I'm structuring things in my content that makes it have a

00:19:59 ◼   ► broad appeal, that it isn't tied into specific people. Or if it is, if I want to make a goal

00:20:05 ◼   ► that is a bit more fixed, that I think about ways of scaling it and adjusting it to be

00:20:11 ◼   ► relevant to that user. And I mean, I have the benefit of knowing a lot about their past

00:20:14 ◼   ► history typically, so I can -- or like, are they in a wheelchair? Do they have a device

00:20:18 ◼   ► that can count floors? What is their average step count? And I can make adjustments and

00:20:24 ◼   ► things to it accordingly, but I think it's certainly something that you just have to

00:20:28 ◼   ► keep in mind that you want to be ideally broadly specific. And I mean, I imagine in the same

00:20:33 ◼   ► way, like, when you're doing your collections of podcasts and overcasts, you have to make

00:20:37 ◼   ► sure that you're not -- they're not all just tech podcasts or whatever it is that you listen

00:20:42 ◼   ► to, that you're trying to think of things that you may not listen to, you may not have

00:20:45 ◼   ► a lot of relevance for, but if a broad enough proportion of your users are interested in

00:20:53 ◼   ► that, then it should probably be represented in there as well.

00:20:56 ◼   ► Oh, yeah. I mean, my directory in Overcast has been so problematic over time. I have

00:21:01 ◼   ► not nailed it yet, but I have learned a few things on the way. I mean, the first version

00:21:07 ◼   ► of it was edited. It was hand-curated by me, and I looked into -- like, I talked to some

00:21:13 ◼   ► people about maybe helping me out with it, but ultimately, every plan I came up with

00:21:17 ◼   ► and the reality of the practice of it had the same problem, which is that when you just

00:21:21 ◼   ► have one person curating a directory of content that could include -- that is a large section

00:21:28 ◼   ► of content, like, I can't listen to every podcast episode of every podcast. I don't

00:21:33 ◼   ► -- like, you would have to hire a staff larger than, like, YouTube staff to do that. And

00:21:38 ◼   ► so there has to be some limiting factor there, and for me, my problem was twofold. Number

00:21:45 ◼   ► one was I only can understand and write English. And so I'm lucky in podcasting. Podcasting

00:21:54 ◼   ► is still extremely English-centric. There is very little podcasting demand or supply

00:22:00 ◼   ► in other languages relative to English. And so that's mostly not a problem for me, although

00:22:05 ◼   ► that is easily a problem for lots of other types of apps. But the other problem that

00:22:10 ◼   ► you run into with curation by one or a handful of people here is just a severe lack of diversity.

00:22:17 ◼   ► Like, there's no way I can listen to everything. I don't even listen to all the tech podcasts.

00:22:22 ◼   ► I don't even listen to all the big shows, like, all the big public radio shows and,

00:22:27 ◼   ► like, the -- like, I don't listen to almost any of them. I don't have time. I listen

00:22:30 ◼   ► to all my own tech stuff. So it's hard for one person to really have a broad view. And

00:22:36 ◼   ► so I switched it to an automated system a few months in. And so what it is now is literally

00:22:44 ◼   ► just based on number of recent recommendations, so that however many people hit the star button

00:22:48 ◼   ► in Overcast for a podcast in each one of those iTunes-defined categories, that's what shows

00:22:54 ◼   ► up there. Like, it's like, you know, a rolling, you know, one-week window or something like

00:22:57 ◼   ► that. And that is not -- you know, that has its own set of diversity challenges and editorial

00:23:03 ◼   ► challenges, but it's at least no longer my fault. And it's way better than what I

00:23:10 ◼   ► was doing myself before, because now I'm taking advantage of the entire user base.

00:23:15 ◼   ► So in a way, it's user-generated content. And then I also have the Twitter following

00:23:18 ◼   ► recommendation system, which can give you even closer tailored recommendations and,

00:23:23 ◼   ► like, the recommendations for you, which is based on what you subscribe to and what other

00:23:27 ◼   ► people who subscribe to those things subscribe to. So that's all, you know, well and good

00:23:34 ◼   ► and better than what I can do individually. There are a lot of challenges to user-generated

00:23:40 ◼   ► content like this that I've tried to be very careful about. I'm still not perfect

00:23:45 ◼   ► on, but I try to be very careful about. You know, it's tempting to outsource this kind

00:23:50 ◼   ► of stuff to your user base because it feels like it's free. It's like, oh, it's -- you

00:23:54 ◼   ► know, user-generated content just generates itself, and you don't have to do anything.

00:23:58 ◼   ► And that's a double-edged sword, because you have to also be very careful about the

00:24:04 ◼   ► typical pitfalls of user-generated content -- spam, illegal content, harassment, hate

00:24:10 ◼   ► speech, you know, like if there's adult filtering required, things like that. There's

00:24:14 ◼   ► all sorts of problems that can result from letting people do their own content that shows

00:24:19 ◼   ► up to other people in the app. That's why I've been very, very careful with Overcast.

00:24:24 ◼   ► The reason why there is nowhere in Overcast to rate a podcast or to leave a review, with

00:24:30 ◼   ► the sole exception of recommending individual episodes with just the on or off star, is

00:24:35 ◼   ► that if you have a way for users to enter text that becomes visible to other users,

00:24:42 ◼   ► you have a big problem of spam, harassment, illegal content, stuff like that. And so I've

00:24:47 ◼   ► been very, very careful to not let people show text to other people in the app. The

00:24:53 ◼   ► only way to do that is to make a podcast and submit it to iTunes, which themselves filter

00:24:58 ◼   ► it for a lot of that stuff, and then it'll start showing up in Overcast. And so I've

00:25:03 ◼   ► avoided having to deal with user-generated content filtering problems and abuse problems,

00:25:09 ◼   ► because I just simply don't allow it. If you can do that, that's great. That doesn't

00:25:14 ◼   ► necessarily work for every app. And I do, unfortunately, still have problems of there

00:25:20 ◼   ► not being a ton of diversity in what gets recommended. Part of that is the user base

00:25:25 ◼   ► is fault, part of that is the kind of rich get richer problem of what is being recommended

00:25:30 ◼   ► most is what's being subscribed to most, and what's being subscribed to most is what's

00:25:34 ◼   ► being recommended most. And so it kind of feeds on itself and the rich get richer, and

00:25:39 ◼   ► it's hard to break into that. So I haven't totally nailed this problem yet there, but

00:25:46 ◼   ► it's definitely way better being algorithmic the way it is now, and user-generated. It's

00:25:52 ◼   ► way better that way than it was when it was just me, one person, trying to edit an entire

00:25:58 ◼   ► podcast directory.

00:25:59 ◼   ► >> Yeah, I feel like the algorithmically-generated content thing is such this, it's great in that

00:26:09 ◼   ► if one of the biggest challenges of creating editorial content is that it is a pain to

00:26:14 ◼   ► be put on a deadline and have to constantly be able to deliver it, then turning to a machine

00:26:20 ◼   ► who can just continuously spew this out is great. That rather than you creating a list

00:26:25 ◼   ► or rather than me making monthly challenges, you can just have the computer do it. Sounds

00:26:30 ◼   ► great. But immediately there's so many problems that come into that that are so funny because

00:26:36 ◼   ► it just doesn't necessarily have the benefit of a person, a person with opinions, which

00:26:42 ◼   ► can be good and bad, or even just making judgments. And I think of an example of this is in the

00:26:48 ◼   ► activity app currently, it does monthly challenges, and the way that it does it is that Apple

00:26:54 ◼   ► has just taken the complete algorithmic approach, where it's just looking at one of your health

00:27:00 ◼   ► metrics for the past month, and then it sets a goal for you based on some increase of that

00:27:06 ◼   ► value, which in theory sort of is good, except for it can lead to these completely absurd

00:27:13 ◼   ► goals or badges that don't make sense. I've had a month where I was traveling a lot and

00:27:21 ◼   ► I did a lot of walking, and then the next month's challenge is like, "This month,

00:27:26 ◼   ► you should walk 50% more or 20% more." And it's like, "I need you to walk 250 miles

00:27:33 ◼   ► this month," or whatever it is. I know exactly what happened there. The algorithm looked

00:27:39 ◼   ► at what I did and says, "Wow, you're really doing a great job with walking. Go do 10%

00:27:44 ◼   ► more. Go do 20% more." But the reality of that is it's completely demotivating and

00:27:49 ◼   ► kind of sad, where I look at this goal and it's like, "That's impossible. There's

00:27:52 ◼   ► no way that I can do that." It makes sense that Apple doesn't want to necessarily be

00:27:58 ◼   ► creating tailored monthly goals for every person that is being done by a person, but

00:28:03 ◼   ► at the same time, as soon as you rely on an algorithm to create the content, it's going

00:28:07 ◼   ► to become problematic in different ways because it just doesn't have the benefit of being

00:28:14 ◼   ► thoughtful. No matter how carefully you design your algorithm, it's not going to be as

00:28:18 ◼   ► thoughtful or creative or relevant as doing it with someone else.

00:28:22 ◼   ► I think there's definitely this balance that we have to find between creating compelling

00:28:28 ◼   ► content that will draw people into your app that ideally is relevant to a broad range

00:28:37 ◼   ► of your users and highlights useful features of your app and is compelling in and of itself

00:28:43 ◼   ► that doesn't destroy you and make you tied to deadlines that pushes you to do too much.

00:28:49 ◼   ► But finding that balance, I think, is a really compelling tool that we have available to

00:28:53 ◼   ► us. But yes, I so want to go down the route of just algorithmically generated challenges,

00:28:59 ◼   ► but then I know it's going to be problematic, and so I don't want to do that. Instead,

00:29:02 ◼   ► I want to be on this treadmill of coming up with a new challenge every month and hopefully

00:29:06 ◼   ► doing it monthly rather than weekly or every couple of weeks is going to be not too much

00:29:11 ◼   ► that I only have to come up with 12 of them a year. My hope is just to make 12 of them

00:29:16 ◼   ► in the next few weeks, and then I can put them in the app and they'll be there. But

00:29:21 ◼   ► it's definitely a challenge, but it's a very compelling tool that I think most apps

00:29:26 ◼   ► can benefit from.

00:29:27 ◼   ► >> Thanks for listening, everyone, and we'll talk to you next week.

00:29:30 ◼   ► >> Bye.

00:29:31 ◼   ► [ Silence ]