When we audit a Meta Ads account that is not working, the same handful of issues turn up almost every time. None of them are exotic, and none of them require a bigger budget to fix — which is usually the first thing an owner assumes is the problem.
Here is what actually accounts for most of the wasted spend we see in Australian accounts.
Targeting that is really just a guess
“Everyone in Sydney aged 18 to 65” is not an audience. It is an absence of a decision. Every campaign should start from a specific problem a specific person has, and the targeting should follow from that — even in an era where Meta's delivery does more of the work than manual interest stacking used to.
That does not mean narrow for the sake of it. Broad targeting genuinely works well now, but only when the creative does the qualifying. If the ad speaks to nobody in particular, broad delivery will find you the cheapest, least interested attention available.
One creative, run until it dies
Running a single ad creative until it gets tired is the biggest source of rising costs we see. Creative fatigue is real and consistently underestimated: the same ad shown to the same audience for weeks stops earning attention, and cost per result climbs even though nothing else changed.
A rough working rule — if an ad has been running more than three or four weeks without a refresh, test a new version before you assume the audience or the offer is the problem. We generally run three to four genuinely different angles per campaign and let the data decide, rather than the loudest opinion in the room about which one looks better.
Boosting posts instead of running campaigns
A boosted post optimises for engagement. A campaign built properly in Ads Manager optimises for the outcome you actually want — a lead, a call, a purchase. They look similar and they are not the same tool.
This one is worth checking even in accounts run by an agency. Related mismatch: the campaign objective versus the business goal. A campaign optimised for traffic will happily bring cheap clicks from people who never intended to buy anything. It did exactly what it was told to do.
Bidding against yourself
Audience overlap is a quieter leak. Five ad sets targeting roughly the same people means those ad sets compete in the same auction — you are bidding against yourself and paying a premium for the privilege.
Checking overlap before launching, rather than after costs climb, is a ten-minute job that saves real money. Consolidating ad sets often lowers cost per result without touching anything else.
Conversion tracking that measures the wrong moment
If the pixel fires on a page view rather than a completed enquiry, the algorithm optimises for people who land and leave. Add the Conversions API alongside the pixel, define the event as the thing that actually matters to your business, and attach a value to it where you can.
Once that is right, everything else gets easier, because the platform is finally optimising toward the same outcome you are being judged on.
Fix these five and most accounts we rebuild see cost per lead drop within the first couple of weeks — not because of a secret setting, but because the money stops going to the wrong places.
The short version
- Start targeting from a specific problem, not a demographic bracket.
- Run three to four creative angles and refresh before fatigue sets in.
- Build campaigns in Ads Manager; leave boosting for brand posts.
- Check audience overlap before launch, not after costs rise.
- Track completed enquiries with values, not page views.