What sets Google Ads App campaigns apart from other ad types is this: you barely choose any keywords, websites or audiences yourself. You simply hand Google a few lines of text, some images and a video or two, then tell it one thing. Do you want more installs, or the kind of user who actually pays? Google then runs the ads automatically across Search, Google Play, YouTube, the Display Network and Discover, finding the people most likely to download and most likely to spend. For a small Vancouver business, the real value is plain. Even without a dedicated ad team, you can let the system put your app in front of the right people, and optimise for either "installs" or "paying users" so you waste less budget on those who install and delete within a day. A local fitness app, for instance, is better off telling Google "I want people who'll buy a monthly subscription" than simply chasing download counts. The system then shows the ad more to people whose behaviour suggests they'll subscribe. This article picks up where the App ads introduction left off: how App campaigns run across channels automatically, and how they optimise for installs versus paying users.
Who is this article for? ❓
If you have a mobile app, such as local food delivery, fitness, bookings or a utility, and want more people to download it but aren't sure you can manage the ads single-handedly, this is worth a few minutes. It won't walk you through the dashboard step by step; instead it shows you what this ad type does automatically on your behalf, and which two decisions are still yours to make. If you don't have an app, and you run an ordinary local service or online shop, then Search, Display or Shopping ads suit you better, so feel free to skip this one.
First, separate two layers: it's a campaign type, but what you steer is the bid
Google Ads is built in nested layers. The outermost is the campaign type, and App campaigns sit right here, on the same level as Search, Display, Video and Shopping. But there's one fundamental difference from the others. With Search ads you pick the keywords yourself, and with Display ads you pick the websites and audiences yourself, whereas App campaigns hand almost the entire targeting layer over to the system: you don't specify keywords, placements or audiences one by one. What you genuinely still control, and what you most need to think through, are two other layers: the assets (the text, images and video you supply) and the bidding (whether you pay for installs or for a paying action). So the focus of this article isn't how to target, but what instructions to give the system when it does the running for you.
What "running across channels automatically" actually means
According to Google's official documentation, you provide a set of ingredients: a group of text ideas, several images, one or two videos, HTML5 assets if relevant, plus your app's name, category and rough audience. The system then tests different combinations of these assets and automatically serves the best-performing ads in these places:
- Google Search (including search partners): your app appears in the results when someone searches a relevant term.
- Google Play: shown in store search results, related apps and the homepage. People here are already looking for apps, so the intent is the most direct.
- YouTube: appears as video or image within relevant content.
- The Display Network: surfaces in Gmail, across mobile websites and inside other apps.
- Discover: in the feed of recommendations on the Google homepage.
In other words, the work that once meant launching several separate ad types and picking placements for each is bundled into this single campaign type. That's precisely why it suits a small business that's short-handed: you don't need to know how to run Search, Display and Video separately, because the system judges for itself which ad performs best where.
The decision that's really yours: bid for "installs" or for "paying"
This is the most important layer of this ad type, and the one most easily overlooked. The three main optimisation goals within App campaigns come down, put plainly, to which kind of result you want:
- Installs (tCPI): the system works to get you as many downloads as it can. On the audience side this targets "All users", aiming to maximise potential install volume. It suits the stage when you've just launched and need to build up a user base first.
- In-app actions (tCPA): you care not just about the install, but whether people actually use the app afterwards. You pick a conversion action that matters, such as completing a sign-up or placing a first order, and the system biases delivery towards people whose behaviour suggests they'll take it.
- In-app value (tROAS, i.e. Target ROAS / return on ad spend): what you value is how much revenue each installed user brings back, which suits apps with subscriptions or in-app purchases. This is the "find people who'll pay" setting.
Back to that fitness app. Choosing "Installs" alone will deliver a long list of downloads, but plenty of those people try it for two days and delete it. Switch the goal to in-app value (tROAS) instead, set "buying a monthly subscription" as the action to optimise for, and the system learns from your existing payment data, showing the ad more to people whose behaviour resembles those who subscribe. The download figure may look less impressive, but a higher share of them actually pay. The budget goes to people who'll renew, rather than to padding the install count.
What the user on the other side actually experiences
On your side, all you configure is the assets plus a bidding goal. But the person on the other side of the screen meets your app in several different settings: they might see it while searching on Google, come across it browsing the Play Store, run into it on YouTube, get a nudge while using another app or in Gmail, or spot it in the Discover feed on the homepage. In most cases a single tap takes them straight to the store to install, a very short path. More to the point, when you switch the goal to paying users, what the system adjusts isn't only the bid but who gets to see it. It steers the ad more towards people whose behaviour predicts they'll subscribe, rather than spraying it at everyone.
Where this sits in your business
App campaigns sit in the "acquiring new users" part of marketing and customer acquisition. Their job is to bring more of the right people into your app. Whether those people then stay, and whether they're willing to pay, comes down to your app's own experience and how your store listing presents it. So it works hand in hand with your product: the ad brings in people who look like they'll pay, and the app catches them and gets them to genuinely stay and subscribe. Polish the assets and the store listing, and the campaign's optimisation has better raw material to learn from.
What happens if you don't get this right
A common mistake is fixating on the download count alone. Installs are cheap and the number looks good, but if most of those who install simply wander off, the money is effectively wasted. The trouble usually lies in one of three places: leaving the goal stuck on installs without moving it up, not setting up the payment data the system needs (which actions count as "valuable"), or assets so thin that the system has little to test and learn from. The result is a pile of downloads and barely any paying users, which is why some people conclude "app ads only bring zombie users". The tool isn't the problem; it simply wasn't told what you actually want.
When it's worth a serious look
A few common moments: you've just launched the app and urgently need a first batch of users to build the base; you already have some downloads but want to shift from "padding the numbers" to "finding people who'll pay"; you run a subscription or in-app-purchase model and care a great deal about the revenue each user brings; or you simply have nobody dedicated to running ads and need a "let the system run it automatically" approach. If any one of these applies, App campaigns are worth serious consideration.
How 5U Website helps
App campaigns look very automatic, but automatic doesn't mean hands-off. What they really test is whether you've made it clear what you want. Here is our honest take. First, if you run a local service or a shop and don't actually have an app, this isn't for you, and a search or local campaign will spend your money better. Second, for the businesses that do have an app, we usually recommend starting on installs (tCPI) rather than jumping straight to a value goal. Chase tROAS on day one and the system has almost no payment history to learn from, so it can't tell who's likely to pay. The sensible path is to run installs first, let real subscription and purchase data build up, and only then switch the goal to in-app value so the system has something solid to optimise against. What we do for clients is make that call on timing, decide which in-app actions to set as "valuable" signals, prepare assets rich enough for the system to test and learn from, then keep watching install quality and return and adjusting, so the budget goes, as far as possible, to people who'll stay and pay. If you're planning to promote an app, or you're getting plenty of downloads now but seeing no revenue, take a look at our website design and digital marketing services, or send us an email describing your situation. We usually reply within one to two business days.
To get the whole picture of how Google Ads fits together first, start with the Google Ads overview; to work out which campaign types exist and which one suits you, see the comprehensive guide to campaign types.
Last updated:
