Social ads
for software.

The question isn't which platform is better. It's who reaches for the credit card. If the buyer is a cafe owner in Gosford trialling rostering software, or a sole-trader bookkeeper looking for a practice tool, they're on Facebook and Instagram most evenings and they decide for themselves. Meta's reach and lower costs usually make sense there.

If the buyer is a head of operations at a 200-person logistics firm in Parramatta, you need to find them by job title, industry and company size. That's what LinkedIn is for, and it charges for it. Many software companies with a self-serve plan and a sales-led enterprise plan end up running both: Meta for the trials, LinkedIn for the demos.

Plenty of products don't need either at the start. If people already search for what you do, Google Ads and SEO catch demand that exists. Social ads create demand, which is slower and costs more to prove. Start there only when search is tapped out, or when nobody searches for your category yet.

A software company knows more about its prospects than almost any other business: who signed up, who got stuck, who looked at pricing, who left. Put in rough numbers for each group to see what you can use.

Meta's pixel has standard events made for subscriptions. StartTrial is when a person starts a free trial of a product or service you offer, and Subscribe is when someone applies to start a paid subscription. Both can carry a value and a predicted lifetime value. Most SaaS set-ups send the first and forget the second, so Meta learns to find triallers.

Send Subscribe from your billing system when the first payment clears, through Meta's Conversions API if you can, so it doesn't depend on the person still having your site open. Then run a campaign that optimises for it once you have enough volume, and compare cost per paid subscription, not cost per trial.

Meta also lets you see results split by audience segment. Jon Loomer's explainer describes existing customers as people who have bought or signed up for your services, and engaged audiences as people who've interacted without buying. Set them up even if you never target them, so you can see how much of your spend lands on people who already know you.

Software ads tend to fail in one of two ways: a stock photo of a smiling team, or a dashboard so busy nobody can read it on a phone. What works more often is one screen, one job, a few seconds long. The invoice going out. The roster filling itself in. The before (a spreadsheet, a whiteboard, three apps) and the after.

Write to the job, not the person. Meta's ad standards say ads must not assert or imply personal attributes, including vulnerable financial status. 'Stop chasing unpaid invoices' is fine for accounting software. 'Are you drowning in debt?' isn't, and it would put people off anyway.

On LinkedIn, the same rule applies with a different tone. A short video from a founder or product lead explaining the problem they built the product for often does better than a polished brand ad, because LinkedIn is where people expect to hear from people.

The most valuable audience you have is people who signed up and stalled. They wanted it enough to hand over an email address. Something stopped them. A discount rarely fixes that. A 20-second video showing the step they're stuck on often does.

Pricing page visitors are the second-best audience. They were checking cost. Show them a worked total for a team their size and how cancelling works, because budget overrun was a problem for 36% of disappointed Australian buyers in Capterra's survey. An ad that answers the money question lowers the risk of signing up.

Cancelled customers are worth targeting only if the reason they left has changed. If they left because the integration with their accounting software was missing and you've now built it, say exactly that.

Uploading a customer list to Meta or LinkedIn means handing over personal information, even when it's hashed. The OAIC says a person can be reasonably identifiable from data like a hashed email address when it can be matched with other information, and that targeted ads based on personal information need a simple opt-out. Say in your privacy policy that you share hashed data with ad platforms, configure pixels to collect the minimum, and keep pixels off any page that would reveal sensitive information.

Small businesses under $3 million turnover are often exempt from the Privacy Act, but many software companies are covered or choose to comply because their customers ask. The SaaS advertising rules page has the detail.