CaddieOpen in CaddieDownload for Mac

How to Get Started with Facebook Automated App Ads

Shared by Caddie Library · Updated
App Masters13:09Open on YouTube
Chapters show their main points

The bottom line

Facebook Automated App Ads is the fastest, cheapest way to test a new app, so start broad with a few creatives and small budgets. Scale only once tracking through an MMP confirms a good return.

Key takeaways

  • Automated App Ads is the easiest and cheapest way to test a new product. 3:29
    • It tests creatives and geos fast, and shows whether users come in cheap enough to hit a target ROI or CAC.
    • The caveat is less visibility: no placement data and little demographic data, so geo and creative are the main levers.
  • A launch needs only 10 to 20 creatives, and static images can beat video. 5:26
    • Up to 50 assets are allowed, but holding some back matters because fast results burn through creatives.
    • Refresh creatives at least every two weeks, ideally every week.
  • Start with small budgets and raise them only when return is good. 6:37
    • Run a couple hundred dollars a day per campaign, per platform, across one or two geos or US versus worldwide.
    • Increase spend only after you see good ROAS, because you can always spend more later.
  • Facebook's system finds buyers on its own once it has data. 8:17
    • New campaigns usually start by optimizing for installs, then move to purchase, with no target CPA to set.
    • Value optimization with a minimum ROAS target needs more data than a new campaign has.
  • Set up an MMP and your data pipeline before you spend. 10:31
    • Facebook's own numbers always look somewhat better than MMP data, because it is a self-attributing network.
    • Real-time postbacks and a data visualization tool such as Looker or Tableau let you match retention and new users back to each campaign, and setup can take weeks without an engineer.

I never advocate testing, like huge budgets in the beginning because you really just don't know.

Gina Kwong

Chapter summaries

0:00 Planning acquisition for a new app

A new app launch starts with choosing geos, platforms and a goal, then testing broadly to find what converts.

  • The launch plan depends on the product, the markets and the goal. 1:35
    • A new product is approached differently from an older one.
    • Pick target geos and decide on iOS, Android or both.
    • The goal may be enough installs to check retention and engagement.
    • Or the goal may be bringing users in to monetize them.
  • Broad testing first, then drilling down, suits mid-size marketers. 2:29
    • Automated App Ads and Google UAC are built for this approach.
    • A broad test shows which audience clusters hit your KPIs.
    • Relaunch campaigns focused only on the clusters that worked.
    • Group countries using the market research you already have.

4:42 Getting started as a small team

A small team can launch with simple static images and does not need a large budget.

  • Static images are enough to get going. 5:03
    • Playable ads cannot be tested on Automated App Ads.
    • Images should show clearly what you want customers to do.
    • Facebook users behave differently from users on other platforms.
    • Static images appeal more for some specific products.
    • Creative testing on Automated App Ads is a separate topic.

6:35 Budgets and campaign settings

Small daily budgets plus a simple optimization event let Facebook learn before you spend more.

  • Start each campaign at a couple hundred dollars a day and raise it only after good results. 6:37
    • Set the budget per campaign and per platform, across one or two geos or a worldwide test.
    • Gina Kwong calls herself a very conservative tester who wants to understand results first.
    • Her reason against big early budgets: you really just don't know.
    • The budget can always go up later, so a slow start costs little.
  • A new campaign optimizes for installs, then moves to purchase, with no target CPA. 8:17
    • Value optimization takes a minimum ROAS target, such as 10% return in the Facebook UI.
    • A new campaign lacks the data for that target.
    • Which event to start with depends on how fast the system learns.
    • Facebook serves the ad to anyone likely to buy and moves on if they don't convert after a few shows.
    • The system blocks publishing until certain problems are fixed, so little gets forgotten.

9:35 Tracking and data setup before spending

A new app needs an MMP and a working data pipeline first, because Facebook's own numbers flatter results.

  • An MMP is required, because Facebook's own data always looks better. 10:31
    • Facebook is a self-attributing network, as are others.
    • Kwong has helped several startups move to an MMP and connect their data.
    • She won't name the size of the gap, only that it always exists.
    • MMPs are fairly affordable for a smaller product.
  • Real-time postbacks and a visualization tool tie every user back to a campaign. 11:17
    • Postbacks run from the internal system to the MMP and then to Facebook.
    • Looker, Tableau or internal queries show new users, revenue and retention per campaign.
    • With a dedicated engineer this goes faster. Without one it can take a few weeks.
    • Run a small test campaign first to confirm the MMP and Facebook talk to each other.
    • Some teams use only the Facebook UI. Kwong warns them, but works within the limits given.

In the growth library

Gina Kwong, Mobile Growth lead at Electronic Arts (EA). Explains how to launch a new app on Facebook Automated App Ads (AAA): how much to spend, how many creatives to use, which optimization events to pick, and what tracking must be in place first. The payoff is a cheap, fast way to find the geos and creatives that work before committing a larger budget.

More from App Masters · The whole library, and ask it anything

Transcript259 lines
0:00

What is up app nation. It is Steve P. Young founder of Appmasters.com. The place you go when you want action packed content related to growing your app downloads, and more importantly, those revenues. And today I've got normal guests. I tracked her down and I was like, you gotta come onto the podcast, please. I begged her to come on to share with us all that. She knows about UA, cause she's been doing it for a long time and I heard on a different podcast.

0:22

And that's when I said, you know, Gina. You got to comedically. Come back on. We're gonna talk all about Facebook. AAA. So that's automated app installs and then Google UAC. And then how do you really get started with just acquisition in general? So without further ado, let me introduce a guest. Her name is Gina Kwan. She is growth at a company called electronic arts EA, as you guys might know it, but Gina, welcome to the show.

0:45

Thanks, Steve. Uh, very happy to be here. Hey, Gina. I want to start with this. So I've gotten clients and some have decent sized budget. Some have no budgets, but for the decent sized budgets, the thing that I hear from them often is, Hey, Steve, we've got like 10, 20 grand, maybe that's their whole marketing budget. Right. And let's go big. Let's go big on launch. What would you say and how do you working at a big company?

1:07

Like EA how do you start proach approaching this? The start of any UAA campaign? Sure. Yeah. I think that, um, with any campaign or any product, uh, it really is dependent on. Whether it's a new product, or if it's like a product that has been around for a little while. So that, that changes kind of how you are. I would approach that, um, what's your product. Let's try it with the new product, because most of the people are launching a new app.

1:34

That's fair. Um, so when a new product, I think that you still need to know, you understand kind of like where. Which geos that you believe that you will be able to, um, be successful in or, you know, geos that you want to target, um, whether it's iOS and Android, both platforms, or do you, is it, you know, you want to focus on one, um, And like, what is the goal really? The initial launch, right? Is it to get enough installs to see whether the, um, retention is and engagement within the app is what you expect it to be, or are you really just kind of launching the, um, these campaigns to bring users in and monetize off of them?

2:20

Right. So like, You having that information is important to figure out how to approach the strategy. And what's really interesting is that, you know, marketers are actually, and marketers are actually have, um, really exciting tools to help them do this. Like the midsize like marketers because, um, AAA in UAC really is built for that. It's like the. Broad the strategy to test broadly so that you can figure out your, the clusters of audience that, you know, K will convert for you, whatever metrics you're looking for, KPIs.

2:58

Um, and then you can drill down and relaunch campaigns that are just focused on those areas. So. Um, yeah, that's how I would approach it. Like really clustering countries, figuring out like what works for your, you know, market research that you've done before, what you think that, um, product will work in. Um, would you still start with a Facebook AAA campaign as like, Hey, let's, let's start with this. Let's see where we get.

3:28

Yes, absolutely. I definitely recommend testing on there because with, um, AAA, you can just really test really fast. You can test out what kind of creatives will work for you. What kind of, um, which geo you will, um, be most successful in. And then really seeing like how you're bringing in, um, Users at a much cheaper rate and that you can actually have a chance to getting a certain ROI or, you know, cost per purchase or Caitlyn CAC, um, depending on what you're looking for.

4:02

So. I think it's, it's the easiest and fastest and likely the cheapest way to test a new product. Now there are caveats to AAA, right? Like you don't get to see placement anymore. You don't really get to see a whole lot of demographic. Um, so those are definitely the caveats, but we're looking at, you know, what, AAA, you're kind of really looking at the broad strokes here with like geo, um, you know, um, yeah, like creative, because those are the big levers that, um, You need to kind of guide how you want to test, like in UAC or other, you know, ad networks, DSP, things like that.

4:42

Okay. So my audience, not, maybe not as huge, sophisticated UAA, people who have humongous budgets, they're more like solo, small business Indies, you know, maybe a team of five to 10 tip people. When you're starting with AAA, what are the assets you need? Is it just the creatives? What do we need just to get going? Yeah. Um, so you can have 50 creative assets that are, um, static or static images or videos. You can not test, play doubles on AAA.

5:15

So, um, yeah, you all you really need is like static images that really convey what you're trying to get your customers to do. And video is also engaging as well. Um, but you don't, I mean, I have definitely seen static images, outperform video, um, You know, like Facebook is, is an interesting platform. Like people just behave a little differently than a lot of other platforms on Facebook because, you know, static images kind of.

5:47

I don't know, sometimes it appeals more for some specific products. Um, yeah. And like, you can have 50, but I really don't think you need 50. Sometimes I think that, um, I, my personal recommendation is like you launch with maybe 10 or 15, maybe 20. Uh, depending on how many you have, because you don't want to like use it. Yeah. Everything you have, right. Because you also do need to refresh the fact that you test it and it's like fast, cheaper results.

6:14

And you get a whole bunch of installs means that it's going to burn through your creatives pretty quickly. So you're going to have to refresh at least every two weeks. Um, ideally every week, uh, And you, you know, I have always kind of advocated creative testing on AAA, but that's a separate story. So you don't really need a massive budget to really test on triple. I mean, I would recommend like, you know, if you're testing one or two GOs or if you want to do like us and then worldwide to see which geos do well, I would.

6:46

Basically kind of set them to, you know, a couple hundred bucks a day for each campaign per platform and then go from there. And if you see good results then, and you get good ROAS, then increase it. Um, I never advocate testing, like huge budgets in the beginning because you really just don't know. And I'm a very conservative UAE, uh, tester, you know, like I'm there, like let's just go off of like do this, you know?

7:11

And I do that. I do actually, I used to have that with a former colleague of mine. Um, She introduced me to that emoji. And I was like, Oh my gosh, that is us. But, uh, I'm very frugal with that because, you know, I'm a big fan of obviously meeting all of my goals. Um, that's really important to me and like making sure that we understand what's going on before increasing budgets, because you can always spend more, um, yeah.

7:38

Complete new question, is it, do we say like, here's our target CPA? Here's our, what do we, what are we controlling within Facebook? Just say, Hey, I've got a couple of hundred bucks a day to spend and you determine everything and go, can we, are there some parameters that we can limit Facebook to? Sure. Yeah, no, that's a very legitimate question, uh, because it's very confusing and Facebook, I think with, uh, yeah, no, it took me a little mint.

8:02

Uh, it took me a minute to like write it all down and be like, okay, this is what it means. So I can explain it to people because I, I do it. Right. And then I'm like, wait, how do I explain this? Uh, yes. So you definitely have to decide whether you want to optimize towards installs a purchase, which is very common, right. Or if you want to optimize towards Rojas, Or like value optimization that's called in, um, on Facebook.

8:25

So you don't really set an amount other than on like the row app, which is called men, relax. Uh, it's a target. So like, let's say you want to target 10% return based on the Facebook UI. Right. Then you said there is a mineral loss. Now, when you're just starting a AAA campaign, you don't have enough data to do that. So you'll likely have to run towards optimizing towards purchase. Um, Or install, actually, depending on how much, uh, how fast the system learns.

8:53

Usually you kind of start with install campaigns and then you move to a , which is the event. And you do not set a target CPA. Basically Facebook is going to look for everyone and anyone that will convert to a purchase for you. So if they serve it to, you know, an ad to somebody and they don't convert after a couple of times, they're just going to move on to somebody else. Um, Yeah. That's like, that's kind of the beauty of this, right?

9:20

It's not set it and forget it, but at least it's not like super difficult that most likely you won't forget anything because there's a lot of times they won't let you, the system won't let you publish, um, until you do like fix certain problems and whatever. And Gina. So let's say you got hired for a brand new startup that is mobile, and they want you to head there, your UA, and they've got nothing, right? The new app is coming out.

9:45

What is like the tech stack. And I know you work for bigger companies. So like, I don't want you to reveal anything that you don't want to reveal, but like, what is the recommended tech stack? You're like, Hey, before you start thinking about UA, here's some critical SDKs or whatever it is you need to have in place. Because I think with, you know, I deal with a lot of the startups who are just starting to use it and I'm like, Look, let me just make sure everything is talking to each other first and foremost, before we start spending lots of big monies.

10:08

Absolutely. And what's really funny. I didn't really use to work for big companies. I actually have more experience with startups. So I'm very familiar with this process because I have helped several companies to really like, um, move MNPS, you know, and like set up kind of, um, data connections and be able to access the data. So. Yeah. I mean, in talking to a comp, like a startup that has a brand new product, definitely because of the app, you have to have an MMP just because like you with any sort of testing that is pretty important because I actually meant to mention it earlier.

10:48

Like Facebook UI data will always look a little bit better than MMP data. I'm not going to talk about certain percentages, but if you, you can always expect that. And that's because. Facebook is a sounds network, right? So self-distributing network. So as you can imagine that that happens, um, and same for, you know, any other self attributing network. So I would have that integrated and also connecting the internal data to be able to pass back like, um, Post backs to them and B and then to Facebook, um, I would love for it to be in real time.

11:29

So then there's callbacks to an internal system and then back down and wimpy and then to Facebook. So that should be like a smooth process process, which takes a minute. If you have a dedicated engineer, It's a lot more helpful, but if you don't, it could take a few weeks, you know, depending on resources and then you need to have, um, a data visualization tool, whether it's like Looker or Tableau or any sort of internal, um, I guess if you need to, you know, pull via query so that anytime you run a campaign, you know exactly what's going on internally, like exactly how much you're getting back, right?

12:08

Like, are these new users, are they not new users? Um, et cetera, like how there, and then, like you were saying, like retention, because that you need to be able to match it back to the campaign on any, on all the levels after they enter the app. Yeah. Yeah. Yeah. So that's, that's what I was telling them too. I was like, Hey, let's just set up a small campaign just to make sure the MMP is talking to Facebook. Everything is probably talking to each other.

12:35

And then before we start really driving it more. Absolutely a little bit. Yeah. That's like the most ideal case. Right. But a lot of times it's like, Hey, we're just going to go off the Facebook UI. And I'm like, I'm going to warn you. This is what's going to happen. But Hey, sometimes, you know, they're okay with it. And we all work within the limitations of what we're kind of given and, uh, you know, especially a small company, they don't have all these resources, even though MMPS are fairly affordable.

13:03

From what I remember, especially for a smaller. Uh, title or smaller product.