What actually happened when your ads didn't work?
Talk to enough trade businesses about marketing and you hear the same story so often you could finish it for them. They ran ads for three months, maybe through an agency, maybe themselves. Leads came in. Almost none of them turned into work. They stopped, told themselves ads don't work for their industry, and went back to word of mouth.
I want to take that story seriously rather than talk anyone out of it, because the frustration in it's completely earned. The money was real and the jobs didn't arrive. But in almost every one of these accounts, when you actually open them up, the ads did what ads do. They put the business in front of people and produced clicks and enquiries at a plausible cost. What failed was everything on either side of them. That matches what I found auditing 31 of these businesses by hand: two had a tracking problem, while 13 were running conflicting details and 14 couldn't be found in search at all, all of it written up in 
What we found auditing 31 Australian trade businesses.
What was missing?
Three things, usually all at once.
Nothing captured the lead properly. The clicks went to a homepage, or into a Messenger thread, or through a lead form left on its most frictionless default settings. So the enquiries were thin: no reliable phone number, no email, nothing you could act on twice.
Nothing followed up. One call, maybe two, and the lead was written off as dead. The research on this is consistent and a little confronting: reaching and converting a lead takes somewhere around six to eight touches on average, and warm enquiries often need five to twelve. How many touchpoints The lead who books on the fourth contact isn't a miracle. They're normal. And the speed of the first response matters as much as the count, because the enquirer filled in two other forms the same evening.
The short life of online sales leads
And nothing recorded anything. No tracking on the site, no record of which lead came from which ad, no link between the campaign and the jobs that did close. Which means there was no way to tell the good campaign from the bad one, let alone improve either. The whole thing ran blind for three months and was then judged on a feeling.
Why does it feel like the ads failed?
Because the ads are the only part you can see. The invoice says advertising, the dashboard says clicks, and when the jobs don't arrive, the visible thing gets the blame. It's a house with no foundation. You move in, the walls crack, and you blame the bricks.
The platforms make this worse by being easy to start. Boosting a post takes two minutes, so it feels like the system has been tried. Meanwhile the parts that decide the outcome, the landing page, the follow-up sequence, the tracking, don't announce themselves. Nobody ever got a notification saying your leads went unanswered again today.
Is this a solo operator problem?
Mostly, yes, and it's arithmetic rather than a character flaw. If you're on the tools all day, out on a 40 degree day or in someone's roof cavity, the enquiry that arrived at 10:15 gets seen at 6pm, and the fourth follow-up call simply never happens because there's no version of your week where it can. The businesses that make ads work either have someone whose job includes the follow-up, or a system doing the first touches automatically. There's no third option where the operator does it manually and it lasts.
That's the honest reason a system has to exist before the ads do. Not because the software is impressive. Because the alternative is you, at 6pm, choosing between dinner and a lead who has already gone cold.
What would make ads work the second time?
In order: somewhere proper for the click to land, a follow-up sequence that runs whether or not you're in a ceiling, and tracking that records which ad every job came from, all in place before the campaign restarts. With those three in place, the same ads that didn't work tend to read very differently, and more usefully, you can finally see which half of them genuinely didn't work and turn those off.
This article is one piece of a longer chain. The whole method, from the click through to the tracking that feeds job values back into the bidding, is written up in full, and you're welcome to take it and do every step yourself.
How do I know if my past campaign actually failed?
Mostly you can't, and that's the finding. If there's no record of which enquiries came from which ads and what happened to each one, the campaign was unmeasurable, which is different from unsuccessful. Whatever you do next, set it up so that this question has an answer next time.
My agency ran everything and it still didn't work. Same story?
Often, yes. Ask what the destination was, what the follow-up sequence looked like, and whether anything was feeding closed jobs back into the platform. Some agencies do all of this well. Plenty deliver ads alone, and ads alone is the exact setup that produces this story. It isn't a question of anyone lying to you. It's a question of what was actually in scope.