Four Meta Ads defaults that quietly spend your money

The easiest Meta Ads leaks are rarely exotic. They are settings an account inherits when campaigns are created quickly and nobody goes back: missing outcome tracking, click optimisation, broad placements, and no second creative to rotate toward.

Meta’s defaults are tuned for Meta’s goals: spend the budget, maximise the metric you asked for, use every surface available. None of that is malicious. But if the metric you asked for isn’t the thing you actually want, the machine optimises very efficiently toward the wrong destination.

1. No pixel. So Meta optimises the only thing it can see

Without a working pixel and an outcome event, Meta has no idea what happens after the click. It can’t optimise toward sign-ups, bookings or purchases, so it optimises toward clicks, because clicks are the only signal it has.

The subtle cost: Meta’s delivery system learns to find the people most likely to click, and the population that clicks most is not the population that buys. The fix is one pixel and one standard event on the page that means money: a lead form, a booking confirmation, a checkout. Standard event names (Lead, Schedule, Purchase) work immediately; no custom conversion setup needed to start.

2. Optimising for link clicks, and losing them on the way

An ad set optimised for LINK_CLICKS rewards the tap, not the visit. Compare link clicks with landing-page views to see how many people actually loaded the destination.

A gap can come from slow connections, back buttons, or accidental taps on mobile placements. It can grow when optimisation is set to link clicks, because the cheapest clicks are the most accidental ones, and the system dutifully finds more of them. The first rung of the fix ladder is free: optimise for landing-page views instead, so Meta only counts a click that actually arrived. The second rung is the pixel event from default #1.

3. Audience Network on by default, and eating the budget

Advantage+ placements are the default, and they include Audience Network: Meta’s ad slots inside third-party apps. Use the placement report to compare its share of spend with its share of confirmed outcomes before leaving it enabled.

Audience Network isn’t automatically bad, but it should be a decision, not a drift. If your results come from Facebook and Instagram feeds, check what share of spend the breakdown report assigns to Audience Network, and if it’s buying clicks that never land, restrict placements and let the feeds have the budget.

4. One ad per ad set: nothing to test, nothing to fatigue into

A single-ad ad set has no creative test running. Meta has nothing to rotate toward when the one ad wears out, and ads do wear out. Frequency creeps up, CTR sags, and the system can only serve the tired creative harder. Two to three ads per ad set is enough to give the delivery system a real choice.

Check your own account in five minutes

Why defaults, specifically

None of these four required a mistake. They’re what an account looks like when campaigns were created in a hurry, which is to say, they’re what most small accounts look like. The platform won’t flag them because the platform is doing exactly what the settings say. The audit that catches them has to come from your side of the table.