Value, Cost And Banked Revenue
The short version of how to track marketing ROI: decide what one conversion is worth, add up every cost of producing it, then match those leads back to money that actually landed in the business. Tags, dashboards and attribution models exist only to make those three numbers trustworthy.
Purge Digital has worked in this industry for over 15 years as a Google Partner, Meta Business Partner and HubSpot partner, so we see the ad platform layer, the analytics layer and the CRM layer of the same month side by side. They rarely agree, and reconciling them is the actual job.
The Formula And How ROAS Differs
Marketing ROI is (revenue minus marketing cost) divided by marketing cost, expressed as a percentage. ROAS divides revenue by ad spend alone and ignores fees, tools and labour. Cost per acquisition tells you what a conversion cost but nothing about what it was worth. You need all three, and you need to know which one you are quoting.
A worked example. A quarter of activity costs $10,000 all in: $7,500 of media, with management fees and tool subscriptions making up the rest. It produces 40 sales worth $40,000 in revenue. ROI is $30,000 divided by $10,000, so 300%. ROAS on media alone is 5.3 times. Cost per acquisition is $250.
Notice how much better the ROAS figure looks than the ROI figure on identical results. That gap is one reason to understand what hiring a Google Ads agency covers before you compare anyone’s reported numbers.
Count Every Cost Before You Divide
ROI figures we are shown by prospective clients are often wrong on the denominator, not the numerator. A defensible cost input includes media spend, agency or management fees, software and tool subscriptions, creative and production, and the internal hours your own staff spend briefing, approving and following up leads.
Two Australian distortions sit on the revenue side and both flatter the result:
- GST mismatch: some systems report revenue inclusive of GST while ad platforms report spend exclusive of it. Pick one basis for both sides of the division, or the answer moves by the whole GST component.
- Revenue instead of margin: top-line revenue ignores cost of goods, delivery and payment fees. For a fair read, back those out and calculate ROI on gross profit. Campaigns that look strong on revenue sometimes break even on margin.
Conversion Goals, Values And Tracking Tags
Measurement follows definition. Before any tag goes on, write down which actions count as conversions, what each one is worth on average, and how a conversion is marked as won or lost. A form fill with no value attached to it cannot produce an ROI figure, no matter how well it is tracked.
- Step 1, define and value: list the conversions that matter, assign a value to each, and agree who updates those values as prices change.
- Step 2, tag traffic consistently: one UTM convention across every campaign, source and medium, applied the same way by everyone. IAB Australia’s Future of Measurement review, presented in September 2026, identified eight priority areas and called for a common campaign data taxonomy serving mix modelling, attribution and experimentation from one structure.
- Step 3, install the platform plumbing: key events in Google Analytics 4, the Meta Pixel and Conversions API where social is running, and conversion imports back into the ad platforms.
- Step 4, capture what happens offline: phone enquiries, quotes closed by email and walk-ins need call tracking and a CRM stage, or your best channels will look like your worst.
- Step 5, consolidate into one view: four platform logins produce four stories. We break the numbers down in plain language rather than exporting a dashboard and calling it a report.
Privacy Law Sets The Tracking Limits
There is no Australian Standard and no licence for marketing measurement. What binds tracking is privacy and anti-spam law, because matching leads to revenue means handling personal information. Three rules change what you can build.
Under Australian Privacy Principle 7 of the Privacy Act 1988, the Office of the Australian Information Commissioner states that an APP entity must not use or disclose personal information it holds for direct marketing unless an exception applies, and where one does apply you must still provide a simple opt-out and honour it. That covers audience lists, remarketing segments and conversion matching, as set out in the OAIC’s APP Guidelines Chapter 7.
For email and SMS attribution, the Spam Act 2003 requires consent before any commercial electronic message, including a message asking for consent. ACMA states you need consent from each recipient and must be able to produce evidence of it on request, including where an agency or platform sends on your behalf, so consent records belong in your tracking stack. ACMA’s guidance on avoiding sending spam is the reference, and its unsubscribe fact sheet cites five business days for actioning a request.
Why Your Three Reports Never Agree
Anyone working out how to track marketing ROI hits this wall inside a month: Google Ads claims one number of conversions, the analytics property reports another, and the CRM shows a third. All three can be right. They are measuring different things with different credit rules.
Our approach is to treat the ad platforms as optimisation tools and the CRM as the arbiter of revenue. Platforms each claim the sales they touched, so a buyer who clicked a search ad and a Meta ad appears in both. Last-click reporting hands the whole sale to the final step, data-driven models spread it, and some conversions are modelled rather than observed. None of that is wrong, but you cannot add them together.
The reconciliation we run monthly is deliberately boring: deduplicate overlapping platform claims, put revenue on one GST basis, convert it to gross profit, then compare against closed-won records. One defensible figure goes to the business owner, with the platform numbers shown separately as what each channel believes it did.
Low Lead Volume Changes The Method
If you generate a dozen or two leads a month, attribution data cannot answer the ROI question. The sample is too small for any model to allocate credit reliably. Switch methods rather than buying more reporting: run clean before and after comparisons on one change at a time, and use holdout tests where you withhold or pause a channel in one region and compare it against the rest.
That also addresses incrementality, the question of whether the revenue would have arrived anyway. Brand search in particular tends to absorb demand created elsewhere, so pausing it briefly often teaches you more than a year of dashboards.
On benchmarks, our own Meta figures are the honest answer to what good looks like. Across our client Meta ad accounts we have managed $1,988,241 in ad spend across 216 ad accounts and generated 96,069 leads, and cost per lead has ranged from $1.02 in one vertical to $56.36 in another. Those are Meta Ads results only, and they are the reason we will not quote a cost per lead before seeing an account.
Channel Metrics And Review Cadence
Each channel needs its own metric set feeding the single ROI figure. On a Google Ads rebuild for Sunshine Coast 24hr Towing we brought cost per click down 47% and the new landing page converted at 45.09%, decisions that came from campaign-level tracking rather than an account average.
- Paid search: cost per click, conversion rate and cost per conversion by campaign, read against the thresholds that matter for your own margins.
- Meta and social: cost per lead and results at creative level, since creative fatigue moves the number faster than targeting does.
- Email: consented list growth, revenue per send and unsubscribe rate.
- SEO and content: no cost per click exists, so amortise the retainer across the months it works and track non-branded organic entrances through to closed revenue. That is how we report on SEO for small business accounts.
Cadence matters more than tooling. Check tactical metrics weekly, and review ROI properly once a month with a longer strategic look each quarter. On the Meta accounts we manage, that monthly rhythm is a plain-language report and a call with the named person on your account.
Auditing The Agency That Spends It
Whoever spends the budget should not be the only party who can see the data. Insist on admin access in your own name to the ad accounts, the analytics property, the business manager and the CRM, and ask which reported conversions are modelled. Our list of questions to ask an SEO agency applies equally to paid media reporting.
This suits Australian business owners and marketing managers already investing in search, social, email or web, and we run accounts remotely nationwide from our Eagle Farm office in Brisbane, including clients in Sydney, Melbourne, Perth and Adelaide. If you want your current numbers pulled apart, take the free half-hour discovery session with our Brisbane-based team, no obligation. GET IN TOUCH at hello@purgedigital.com.au, or call 07 3053 5322 or 0483 968 501.
Frequently Asked Questions
How Long Before A Campaign Shows True ROI?
Long enough to cover your sales cycle. If quotes typically close three months after the first enquiry, a monthly report compares this month’s spend against revenue earned from older enquiries and will call a profitable campaign a loser. Set the lookback window to match your real close time and report lead cohorts, not calendar months.
Which Tools Do We Actually Need?
Four jobs, not four brands. Analytics to see site behaviour and key events, a CRM to carry each lead through to closed-won revenue, platform tracking such as the Meta Pixel and Conversions API to feed the ad systems, and one consolidated report. Call tracking is the usual fifth for service businesses that convert by phone.
Why Do Tracked Conversions Undercount Results?
Consent choices, device privacy settings and shorter cookie lifetimes all break the link between a click and a sale, so tracked conversions usually sit below what the CRM records. Server-side conversion feeds close part of the gap. Treat platform conversion counts as directional and your own sales records as the truth.
How Do We Prove SEO Is Paying Off?
Measure movement in non-branded organic enquiries against the amortised retainer, and watch the technical health signals that precede it. Recovery work shows up in impressions before it shows up in revenue, which is why we report both rather than one.
Can We Just Ask Customers How They Found Us?
Ask, but never calculate ROI from the answers. Dropdown surveys and mention-this-ad codes capture the last thing a customer remembers, not the sequence that moved them. They are useful colour next to tracked data and unreliable as a substitute for it.
