Meta, Google, TikTok Are All Taking the Steering Wheel: Where Should Cross-Border Sellers Bet Their Ad Budget?

A while back, a friend of mine who runs a cross-border e-commerce business invited me to dinner. Halfway through the meal, he pushed his phone across the table to show me his ads dashboard.

"Look, these days I only do two things: enter a URL and set a budget. Everything else, the system runs by itself."

I asked: so what about your keywords? Your audience segments? Your bidding strategies?

He said: gone. The platforms took them all.

After that dinner, I went through all three platforms' AI ad products: Meta's Advantage+, Google's AI Max, TikTok's Symphony and GMV Max. The more I read, the clearer it got: this year's shift is not as simple as "a platform rolled out a new feature."

The advertising business is moving from "AI-assisted" to "AI-driven."

What does AI-assisted mean? You're driving, and AI hands you a water bottle and reads out the traffic. What does AI-driven mean? AI drives. You sit in the passenger seat and buckle up. That's it.

The direction has changed. Let's take them one by one.

Meta: A Fully Managed Butler — But You'd Better Keep It Fed

This year, Meta's Advantage+ has gone all the way from "optional" to "default."

You give it a landing page URL and a budget. Who to target, where ads run, how creatives get optimized, how bids get adjusted — the system handles all of it.

Sounds great. But there's a catch.

In Q1 this year, Meta released a set of numbers: e-commerce advertisers using the full Advantage+ suite saw average customer acquisition costs drop 32%.

My God — 32%. That's a lot.

But — notice this "but": the catch is that you have to feed the system 20 or more new creatives every week. Its ranking model eliminates weak creatives, and it eliminates them fast. If you don't feed it, it goes hungry. When it goes hungry, your costs go up.

For cross-border sellers, there's something even more practical: AI localization. You upload creatives in Chinese, and the system automatically generates the English version, the Spanish version, the Arabic version. Translation, voiceover, the small cultural tweaks — all done for you.

Test results show that AI-localized creatives get a click-through rate 18% higher than human translation.

It used to be: find translators, find voice actors, adapt creatives market by market — slow and expensive. Now, upload one file.

That's not ad spend you're putting out. That's feed for the machine.

Google: The Keyword Era Is Officially Over

Where to begin? Let's just start with the word "keyword."

After all these years of running ads, we've all gotten used to one thing: buying keywords. What words people search, what you bid, which ads match.

This April, Google opened AI Max to all advertisers. That move put a period at the end of the keyword era.

Now you don't touch keywords anymore. You provide a landing page and product descriptions, and the AI works out for itself what the person on the other side of the screen actually wants to buy.

Even more interesting is AI Mode. Right in the search results, users can see AI-generated product comparisons and recommendations.

Think about it: what does this mean for a small brand nobody has ever heard of?

It means opportunity. When the AI ranks results, it looks at the quality of your landing page and your user reviews — not your fame. For the first time, small brands can stand on the same starting line as the big names and compete on content.

On the creative side, Gemini can take a single product photo and generate more than 50 combinations of headlines and descriptions.

Impressive!

But I have to warn you: AI-written copy sometimes exaggerates, overstating what the product can do. Before anything goes live, a human has to check it.

No matter how fast the machine runs, there has to be a clear-headed person sitting next to the wheel.

TikTok: A Content Factory Up Front, a Cash Register in the Back

TikTok's play is different from the other two. It's betting on content production plus a closed conversion loop.

Symphony's job is making content: writing scripts, auto-editing, turning out different versions in batches. By the official numbers, it saves 70% of production time.

Where does the saved time go? Into shooting more creatives.

The real killer is GMV Max. It distributes ads the social way: prioritizing people who've seen similar content, liked it, left a comment.

Their conversion rate runs 40% higher than traditional targeting.

Put simply: the people who watched, liked, and commented were already tempted. TikTok is just helping you push them one more step forward.

Also, it launched two new placements this year: Logo Takeover and Prime Time. Good for brand-building, but expensive — don't expect them to drive conversions directly. For conversions, GMV Max is still what you rely on.

So How Should the Budget Be Split? Do the Math by Stage

All three platforms say they're great. You can hear it — that's the line all three sales teams would give. Let's ignore the pitch and do the math.

First: monthly revenue under $10K.

All in on Google.

Why? Search traffic comes with clear intent, a short conversion path, and a gentle learning curve. And the bar is low: Google lets you start at $20 a day, while Meta and TikTok both need $50.

Small business. Every dollar has to be spent where intent is clearest.

Second: monthly revenue between $10K and $100K.

Meta and Google, 60/40.

Meta handles scaling volume and retargeting; Google catches the people who come searching on their own. Category matters too: standardized products — tools, accessories — go to Google; visual, non-standardized products — apparel, home goods — go to Meta. One wins by being found; the other by being seen.

Third: monthly revenue over $100K.

Run all three — but split the jobs clearly: TikTok for volume, Meta for efficiency, Google for intent.

Volume, efficiency, intent. Remember those three words and those three companies — that's all you need.

One Last Reminder: Consolidate, Then Consolidate Again

On one thing this year, all three algorithms are strikingly consistent: they love big campaigns and hate a pile of small ones.

Why? Meta's Advantage+ needs enough data volume to finish learning. Google's AI Max needs enough conversion signals to optimize intent matching. TikTok's GMV Max needs enough engagement data to recognize high-value users.

AI is like a person: if it doesn't eat enough, it can't do good work.

So take your 10 small-budget campaigns and merge them into 2 to 3 big ones — give the AI room to optimize.

The first week after merging, the numbers will wobble. Don't panic. That's normal. It has to relearn.

The friend from the beginning of this article asked me after dinner: so who exactly should I bet on?

I said: if budget is tight, master one platform first. All three of this year's AI systems run on accumulated data. If you keep switching back and forth, every system stays stuck in cold start, and none of them ever learns.

Pick one, feed it till it's full, then talk about the next.

And here's wishing you: buckle up early, ride in this AI-driven car — and still sleep soundly at night.