Most Google Ads and Meta accounts running in Malaysia today are already on some form of automated bidding, and a growing share of the ad itself — headlines, creative combinations, even who sees which version — is being assembled by the platform’s own AI rather than chosen manually. That shift didn’t arrive as one announcement. It built up over several product changes until keyword-by-keyword, placement-by-placement control became the exception rather than the default. The practical question for advertisers isn’t really whether to use it, since most no longer have a full manual alternative. It’s what still requires a human decision, and what quietly goes wrong when nobody checks.

What’s actually being automated now

On Google, Performance Max runs a single campaign across Search, Display, YouTube, Discover, Gmail and Maps, with the system deciding budget allocation and placement mix based on the assets and conversion goals provided. Smart Bidding strategies such as Target CPA, Maximize Conversions and Maximize Conversion Value set an individual bid for every auction using signals — device, location, time of day, audience history — that no one could realistically evaluate by hand at that speed.

Meta’s Advantage+ campaigns work on the same principle: audience selection, placement and creative combinations are handled by Meta’s system rather than set manually, and Meta has publicly stated an ambition to move toward largely AI-generated ad creative and targeting by the end of 2026. Whatever the exact timeline, the direction is clear. The parts of paid media that used to be manual craft — bid adjustments, audience layering, ad variant testing — are steadily becoming inputs into a system, rather than actions a media buyer performs directly.

Why this isn’t “set and forget”

The system optimises toward whatever you tell it to

Automated bidding chases whatever action is marked as a conversion, at whatever volume and cost the algorithm judges will produce more of it. That’s the useful part and the risky part. If the conversion signal is clean — a genuinely qualified enquiry, a completed purchase — the system gets efficient at finding more of exactly that. If the signal is noisy — a form that fires twice, spam submissions counted as leads, WhatsApp enquiries that never make it into tracking at all — the algorithm optimises just as hard toward more of that noise, because it has no way of knowing the difference. We’ve written before about why website conversion tracking often breaks and why Google Ads leads don’t convert even when cost per lead looks fine — both problems get more expensive under automated bidding, not less, because the system now has a direct mandate to spend harder toward whatever it’s being told counts as success.

Budget and account structure still shape the outcome

Smart Bidding and Advantage+ need enough conversion volume flowing through a campaign to find a stable pattern. Splitting a modest budget across several overlapping campaigns, changing budgets or conversion goals frequently, or restructuring an account mid-month all reset or dilute that learning process. None of this is visible as an error. The dashboard keeps reporting, the campaign keeps spending, and the results are simply less stable than they should be for the spend involved.

What Malaysian advertisers should still control

Automation narrows the list of manual levers, but it doesn’t remove the need for judgement. The parts that still matter:

  • The conversion actions the system optimises toward, and whether they represent genuine commercial value rather than form-fill volume
  • The creative assets and messaging fed into the system, since automation can only recombine what it’s given, not invent a stronger offer
  • Account and campaign structure, so learning data isn’t split thinly across too many overlapping campaigns competing for the same auctions
  • Feedback from CRM and lead-qualification data back into the ad account, since click-level tracking alone can’t tell the algorithm which leads actually turned into paying customers
  • A regular review cadence against real business outcomes — sales, qualified pipeline, revenue — rather than platform-reported conversions alone, since those two numbers can drift apart quietly

The risk for smaller Malaysian accounts

This is where the trend bites hardest for SMEs. Automated bidding performs best with a steady, reasonably high volume of conversions to learn from. Many smaller Malaysian advertisers run monthly budgets that produce too few tracked conversions for the system to find a genuinely stable pattern, which shows up as bidding that swings between efficient weeks and expensive ones without an obvious cause. The usual fixes are structural rather than creative: consolidating fragmented campaigns instead of running several small ones, broadening the conversion action to something that happens more often earlier in the funnel while still meaning something commercially, or accepting a longer manual or semi-automated phase before handing full control to the algorithm. None of these are the platform’s default setup, which is built around advertisers who already have volume.

Where this is heading

As more of the bidding and creative decision-making moves inside the platform, the advertiser’s real influence increasingly comes down to two things: the quality of the data feeding the system, and the clarity of what “success” is defined as. Manual bid tweaking is becoming a smaller part of the job; making sure the numbers the algorithm sees actually reflect the business is becoming a bigger one. This is part of why MRVS looks at paid media alongside website tracking, CRM data and lead quality as one connected system rather than optimising the ad account in isolation — an automated bidding strategy can only be as accurate as the conversion data it’s given, wherever that data comes from.

Frequently asked questions

Does automated bidding still need conversion tracking set up correctly?

Yes, more than ever. Automated bidding removes manual control over individual bids, which means it relies entirely on the conversion signals it’s given. Broken or inflated tracking doesn’t just misreport results under automation — it actively directs spend toward the wrong outcome.

Is Performance Max or Advantage+ suitable for a small monthly budget in Malaysia?

It depends more on conversion volume than the budget figure itself. An account generating too few tracked conversions per month typically sees unstable results from fully automated bidding, regardless of spend, and often performs more predictably with a narrower, more manual setup until volume builds up.