Three Tech Giants Are Racing to Run Your Ads — and the Peace of Mind You Save May Cost You Money

A while back, a friend who does app promotion vented to me: running ads just doesn't feel like a craft anymore.

Why?

Upload a few creatives, set a budget, and the platform handles everything else. TikTok says "hand it over," Meta says "I've got this," and Google just smiles knowingly on the sidelines — all three competing to make your life easier.

Sounds great, right?

But something didn't sit right with me. Less hassle and less spend are not the same thing.

Google Started This Back in 2015

What does "AI fully automated ad buying" actually mean?

It means the platform tells you: you don't need to adjust bids line by line anymore, and you don't need to hand-pick placements one by one. Give me your creatives, give me your budget, tell me your goal — I'll handle the rest.

This isn't something that popped up last year.

Google moved first. In 2015, it launched UAC（Universal App Campaigns）, later renamed Google App Campaigns. One product to run your ads across Google Search, Google Play, YouTube, and the Google Display Network.

Back then, "AI" wasn't the buzzword it is today. Google talked about machine learning — saying it could optimize for installs and in-app actions on your behalf.

By 2022, Meta couldn't sit still anymore and launched Advantage+.

Then, more recently, TikTok rolled out its own Smart+. Even Apple is rumored to be working on something similar.

All three use almost the same pitch: hand over your creatives and your money, and for everything else — just lie flat（a Chinese slang term for opting out of excessive effort, doing the bare minimum）.

But Before You Lie Flat, You'd Better Know Whose Bed You're Lying On

I dug through all three companies' offerings to figure out one thing: how many decisions are they actually making for you, and how much room are they keeping for themselves?

Let's start with creatives. All three want you to upload assets, but their appetites differ. TikTok asks for at least 6; Meta goes biggest, accepting up to 50 at a time — images, videos, even playable ads; Google sets you a cap — 5 headlines, 5 descriptions, 20 images, 20 videos, 20 playables.

Now, placements. Here's where the differences get stark.

TikTok's ads only run on TikTok's own turf — they can be Spark Ads（user-native content turned into ads）or standard ads, but either way, they never leave the app.

Meta? 23 placements spanning Facebook, Messenger, Instagram, and its Audience Network.

Google's footprint is even broader — Search, Play, YouTube, the Google Display Network, AdMob, Discover, plus a slew of unnamed search partners and third-party publishers.

See the pattern?

The more placements a platform has, the easier it is to stuff your ads into corners you'd never actively choose.

The Platform's Math Is Sharper Than Yours

Let me put it this way.

Imagine you open a restaurant and hire a floor manager. You tell him: when guests arrive, just seat them wherever makes sense.

Sounds effortless.

But this manager has a little agenda of his own: he needs to fill every table, especially the ones nobody wants — the one near the bathroom, the corner where the AC blasts directly. So whenever he gets the chance, he slides guests into those spots — after all, you never said he couldn't.

That's precisely the business ad platforms are in.

They've got a mountain of ad inventory to monetize. Some of it is prime real estate (search results pages, top of feed), and some of it is a ghost town (some forgotten partner site, a tiny notification in Messenger). The essence of AI automation is giving them a "legitimate" reason to spread your ads across all of it.

They can hit your ROAS（Return on Ad Spend）target. Hitting the target is enough. Overshooting, from their perspective, is a wasted monetization opportunity.

You need to understand this.

TikTok Is the Simplest — and the Youngest

Of the three, TikTok Smart+ is the newest and the most lightweight.

You give it creatives, budget, and goals, pick a global or language scope, and it starts running. Budget recommendations come in two tiers: for Maximum Delivery, the suggested daily budget is 10x your CPA（Cost Per Acquisition）; for Cost Cap, it's 3x your usual daily budget. On iOS, you can only run Maximum Delivery.

The learning phase lasts 7 days.

According to TikTok's own case studies, CPI（Cost Per Install）can drop by 21%. Take that number with a grain of salt — they handpicked the samples.

Meta Advantage+ asks for a bit more: beyond creatives, budget, and goals, you may also need to link a Facebook Page or Instagram account. The learning phase is considered done once roughly 50 optimization events have been collected. Meta's own 21 A/B tests claim CPA can improve by 26%. Also self-reported.

Google has the longest learning phase — up to two weeks, though higher budgets and more conversion events speed it up. Its budget recommendations are the most aggressive: for install campaigns targeting CPI, the suggested bid is 50x; for target CPA, 10x; for engagement-focused ACe（App Campaigns for Engagement）, 15x.

See — every one of them is telling you: pay enough tuition, and only then will the AI study hard.

The Real Trap Is "Effortlessness" Itself

This brings to mind two conversations.

The first was with a friend in media buying. He got into Meta Advantage+ early — and got burned.

Where did it go wrong?

The system's "easy button" was too easy to press. A few clicks and you've got an ad group set up — so smooth it lulls you into dropping your guard. Then its audiences start expanding outward, reaching people who will never convert. That's how the budget gets burned.

The lesson: without all those knobs and levers to fiddle with, what looks like less hassle is actually flying blind.

The second conversation was with a ten-year Google veteran talking about YouTube. She said something most people don't realize: YouTube — especially YouTube on iOS — is a massive chunk of Google's inventory.

But you'd better give it the right creatives.

Video formats on YouTube are wildly diverse — Shorts, long-form, pre-roll, mid-roll, bumper ads, skippable, non-skippable, in-feed sponsored cards. If you don't provide creatives sized and formatted for each, the AI can't place your ads in those premium spots even if it wants to.

So sometimes, giving a platform too many creatives is actually giving it an "excuse" to run your ads on low-quality placements. Restricting your creatives is forcing it to put your ads where they matter.

In One Sentence

These three systems genuinely make life easier for anyone running ads.

But ease comes at a price.

The price is handing over the power of fine-tuning in exchange for the comfort of "it's running itself." The peace of mind you save will likely be paid back in budget.

My own take: these tools should be used — must be used. Sit out and you'll fall behind. But when you use them, don't close your eyes. The harder the AI runs, the harder you need to watch where it's running.

As for which one to pick — that depends on where your business is, where your users are, and where your creatives are.

There's no such thing as "winning without lifting a finger" in advertising — only ad buyers who've chosen to let go of the wheel.