What to budget
for SaaS marketing.

The most-quoted numbers for private software companies come from SaaS Capital. Its 2026 survey, completed in March 2026, found the median private SaaS company spent 15% of annual recurring revenue on sales and 8% on marketing. Equity-backed companies spent about 70% more on sales and twice as much on marketing as bootstrapped ones, and faster-growing bootstrapped companies spent more on both.

Use that as a sanity check, not a plan. The survey isn't Australian, and the percentages cover all of marketing (people, tools, content and events, not only ads). A bootstrapped Australian company selling a $49-a-month product has very different maths from a funded one selling five-figure annual contracts. Two companies at 8% can be in completely different health.

Monthly recurring revenue (MRR) is a bucket with a hole in it. Every month some customers leave. New sales have to cover that before the bucket gets any fuller. Put in your numbers to see the split. The starting figures are made up.

Run the calculator with 3% monthly churn and then with 1.5%, and the new MRR you need to hit the same target drops sharply. That's the uncomfortable truth in a lot of SaaS budgets: a chunk of the ad spend is paying to replace customers the product, onboarding or pricing lost.

Churn also compounds. 2% a month isn't 24% a year; it's about 22%, because each month's loss comes off a smaller base. The point stands either way: losing a fifth of your customers every year means a fifth of next year's marketing is spent getting back to where you started.

Budget surprises cut both ways. In Capterra's survey, budget overrun hit 36% of disappointed Australian software buyers, and a customer who feels the bill crept up is a customer who leaves. So before raising the ad budget, look at why people leave. If the top reason is 'it didn't connect to our accounting software' or 'too hard to set up', the cheapest growth is fixing that, and the marketing money goes further afterwards. The how software buyers choose page covers the reasons Australian buyers regret a purchase.

The cleanest way to set a ceiling on paid marketing is payback: how many months of gross profit from a new customer it takes to earn back what you spent winning them. If a customer pays $120 a month at an 80% margin, they bring in $96 of gross profit a month. At a 12-month payback, you can spend up to about $1,150 to win them.

Pick the payback period from your cash, not from a blog post. A bootstrapped company paying its own bills might need six months. A funded company planning for scale might accept 18 or more. Then multiply that ceiling by your trial-to-paid rate to get the most you can pay per trial. Use your own rate: the 2026 Conversion Report found a median free-to-paid conversion of 8% across 200 products, but the spread is wide. At 8%, a $1,150 ceiling per customer means about $92 per trial. If your ads can't beat it, more budget won't help; the funnel needs work first.

This is also where the Spend Curve matters. The first dollars on the best searches usually pay back fastest. Each extra dollar buys less, and at some point it stops paying back inside your window.

For most early-stage software companies selling to small businesses, the order that tends to work is this. Search first (people already looking for the job you do), then review sites and comparison content (where buyers check you), then social and retargeting (to people who've shown interest). Content and SEO take months to pay back but keep working after you stop paying.

Budget for the measurement too. Tracking paid customers back to the ad that found them, and keeping it working, is what makes every other line on the budget make sense. It also sets a floor on ad budgets: Google suggests a bid goal with at least 15 conversions a month. A budget too small to buy that many paid customers or activations will struggle to let automated bidding learn. The SaaS Google Ads page explains the set-up.

Australian software companies often claim the R&D Tax Incentive for product development. Marketing isn't part of it. business.gov.au lists market research, market testing or market development, or sales promotion among the activities excluded from core R&D. A survey used to design a core experiment may count as a supporting activity; one used to plan marketing doesn't.

That matters for the budget conversation with your accountant. Don't plan marketing spend assuming a refund on it, and keep marketing work clearly separate from development work in your records. Ask your R&D adviser where the line sits for your own product; this is a marketer's summary, not tax advice.