How return on ad spend is calculated

The formula is: ROAS = Revenue from ads divided by ad spend. If your campaigns generated R52,000 in attributed revenue and your platform spend was R13,000, your ROAS is 4.0x. That is typically expressed as a ratio (4:1) or a multiple (4.0x) or a percentage (400%). All three mean the same thing: four rands of revenue returned for every rand put into the platform. For a broader look at what this metric means across different channels, see our guide on ROAS in digital marketing.

The definition of “ad spend” matters more than most business owners realise. Ad spend means the money paid directly to the advertising platform, whether Google Ads, Meta Ads, or another network. It does not include agency management fees, creative production costs, or software subscriptions. Some practitioners include additional costs in a broader “total ad investment” figure, but for measuring pure campaign efficiency and benchmarking against industry data, most published ROAS figures use platform spend alone. Mixing agency fees into your denominator will make your ROAS appear lower than industry benchmarks that exclude those costs, which makes direct comparison unreliable.

Three worked examples make this concrete. R15,000 revenue on R5,000 ad spend = 3.0x ROAS. Positive-looking number, but depending on your margin, you could be breaking even or in the red. R3,200 revenue on R4,000 ad spend = 0.8x ROAS: you are losing money before a single overhead is counted. R240,000 revenue on R50,000 ad spend = 4.8x ROAS: above the ecommerce average and, for most margin profiles, genuinely healthy. The raw number only tells part of the story until you layer in your gross margin, which the break-even section covers.

Why your attribution model changes the number

Two businesses can run identical campaigns, spend identical budgets, and generate identical revenue, yet report completely different ROAS figures. The reason is almost always attribution. Attribution determines which ad interactions receive credit for a conversion, and the model you choose changes how that credit is distributed across your touchpoints.

Last-click attribution gives 100% of the conversion credit to the final ad a customer clicked before completing a purchase. If someone first encountered your brand through a prospecting campaign, then saw a retargeting ad three days later and bought, last-click gives all the credit to the retargeting ad. That makes your closing channels look far stronger than they actually are, and your upper-funnel prospecting campaigns appear to contribute nothing even when they put the customer in market in the first place.

Data-driven attribution distributes credit across all meaningful touchpoints based on machine learning and observed conversion path data. It weights interactions by how much they actually influenced the outcome rather than applying a fixed rule. Data-driven attribution is generally more reliable than any rule-based model, but it has a minimum data requirement: at least 30 conversions per month at campaign level before Google Ads has enough signal to run the model accurately. Accounts with lower conversion volumes should use position-based or linear attribution rather than data-driven until the volume threshold is reached.

The cross-platform problem compounds this further. Google Ads typically applies a 30-day click-through attribution window. Meta Ads defaults to a 7-day click plus 1-day view window, which means Meta can credit conversions from users who saw an ad but never clicked it at all. That view-through credit inflates Meta’s reported ROAS relative to Google when you compare platform dashboards directly. A blended ROAS from both channels combined can be misleading if the attribution windows are not reconciled. The practical fix is to use a unified analytics source such as GA4 or a CRM dashboard with a consistent attribution model applied before drawing any cross-channel budget allocation conclusions. For a detailed breakdown of data-driven versus last-click attribution mechanics, see the Wizaly attribution comparison.

What ROAS benchmarks look like across industries and channels

The 2026 ecommerce median sits at approximately 2.87x overall, but that figure masks significant variation by channel and vertical. By channel within ecommerce: Google Ads ecommerce accounts average approximately 3.52x; Meta Ads ecommerce accounts average approximately 1.86x. Those figures are not a verdict on which platform is better. They reflect the different intent profiles of each platform’s audience. For a full walkthrough of what constitutes a strong number for your sector, see our dedicated post on what is a good ROAS for Google Ads.

Breaking the ecommerce benchmark down further by vertical: beauty and personal care on Meta averages approximately 1.57x versus approximately 2.80x on Google; fashion averages approximately 2.18x on Meta versus approximately 3.40x on Google; electronics averages approximately 1.92x on Meta versus approximately 3.02x on Google. These gaps exist because Google Search captures people who are actively looking to buy right now, while Meta reaches people before they are in active search mode. Google is stronger for categories where search intent is explicit. Meta is stronger for impulse categories, retargeting, and brand awareness. Source: Hawky ecommerce ROAS benchmarks.

B2B and SaaS sectors track at 1.55x to 1.70x blended across paid channels, which sounds low compared to ecommerce but is not a red flag. Longer sales cycles, multi-touch attribution across weeks or months, and high customer lifetime values mean that optimising aggressively for immediate ROAS in B2B will undercounts the true return from advertising investment. Retail media networks, by contrast, average approximately 6.1x blended because purchase intent is extremely high when someone is already inside a retail platform looking at product listings.

One critical warning when reading channel-level ROAS data: never blend Google and Meta figures before separating them. A blended 3.0x across both channels could mean Google is running at 5.0x while Meta is running at 1.5x. Those two numbers require completely different decisions. Keep them separated. Then use a unified analytics source to understand the interplay between channels before drawing budget allocation conclusions. For more context on how platform ROAS figures compare in South African accounts, see our overview at What Is ROAS? A Plain-Language Guide.

Why a strong ROAS number can still mean you are losing money

ROAS does not account for your cost of goods. A 3.0x ROAS looks positive until you factor in that your product costs 70% of revenue to make and deliver, at which point you are losing money on every sale the ads generate. The break-even ROAS formula closes this gap: break-even ROAS = 1 divided by gross margin percentage. At a 25% gross margin, your break-even ROAS is 4.0x. At a 40% gross margin, it drops to 2.5x. At a 50% gross margin, it is 2.0x. Any ROAS below your personal break-even threshold means your advertising is destroying value, regardless of what the platform dashboard reports. For the full methodology, see Triple Whale’s break-even ROAS guide.

Customer lifetime value changes the calculation for subscription and repeat-purchase businesses. A 2.0x first-order ROAS looks like a poor acquisition result until you account for the fact that the customer makes three more purchases over the next 12 months. In that scenario, the blended ROAS across all four transactions might be 6.0x or 8.0x, making the initial acquisition entirely justified. Businesses with strong retention should set their ROAS targets based on 12-month net lifetime value rather than single-transaction revenue. A practical benchmark from subscription and SaaS contexts: a 3:1 LTV-to-CAC ratio is generally considered a healthy acquisition threshold, meaning your customer’s lifetime value should be at least three times what it cost to acquire them. If your LTV:CAC is above 3:1, a lower first-order ROAS may be entirely acceptable business logic rather than a campaign problem.

ROAS vs ROI and CPA: using the right metric for each decision

ROAS measures ad efficiency: how many rands of revenue did you generate for every rand paid to the ad platform. It does not account for cost of goods, fulfilment costs, team costs, software costs, or any other operating expense. A campaign can show 5.0x ROAS while the business is running at a loss if COGS and overheads consume all the gross profit the revenue generates. ROAS is the right metric for campaign-level decisions: which ad sets to scale, which creatives to pause, which audiences to expand, which channels deserve more budget this week. See our overview of the ROAS meaning for a focused look at the definition.

ROI measures total profitability: revenue minus all costs, including COGS, staff, tools, agency fees, and overhead, expressed as a percentage of total investment. ROI is the right metric for business-level decisions: whether a channel deserves a larger share of your overall marketing budget next quarter, or whether the ads are actually growing the business after all costs are counted. Strong ROAS with weak ROI is a signal to examine costs, not just ad performance.

CPA (cost per acquisition) measures cost efficiency: how much it costs to generate one customer or one conversion. Where ROAS focuses on the revenue side of each transaction, CPA focuses on the cost side. CPA is most useful when your goal is customer acquisition volume rather than immediate revenue, for example in subscription models where payback period extends beyond the first transaction. Running all three metrics simultaneously gives you the full picture. ROAS tells you how hard each campaign dollar is working. ROI tells you whether the business is actually profitable from ads. CPA tells you what each new customer costs to acquire and whether that acquisition cost is sustainable given your LTV.

7 levers that improve return on ad spend without increasing budget

Smart bidding, audiences, and creative: the highest-impact starting points

Smart bidding with Target ROAS is one of the most powerful tools in Google Ads, but it rewards patience. When adjusting your ROAS target, move in 10 to 20% increments rather than large jumps. A sudden shift from a 3.0x target to a 5.0x target will almost certainly destabilise the campaign’s learning phase, restrict delivery, and reduce conversion volume while the algorithm recalibrates. Small, incremental adjustments let the bidding algorithm adapt without disruption. For a full breakdown of smart bidding approaches, see the Savvy Revenue smart bidding strategies guide.

Retargeting warm audiences consistently outperforms cold broad targeting on ROAS because you are reaching people who already know your brand or product. Pair this with value-based lookalike audiences built from your highest-LTV customers rather than your full customer list. A 1% lookalike of your top-spending customers will behave more like your best buyers than a lookalike of your entire customer base. Exclude recent purchasers from acquisition campaigns so you are not paying to re-acquire someone who has already converted.

Creative testing is one of the highest-use and most consistently neglected activities in paid advertising. Many accounts continue running the same creative long after performance has plateaued. Rotate formats, headlines, and calls-to-action through structured A/B tests. Test one variable per round to isolate what is actually driving the performance difference. Accounts that test creative actively tend to outperform static accounts on ROAS over time, because each winning variant raises the performance floor for the next round. See also: Improvado: how to improve your PPC ROAS.

Landing pages, LTV programmes, and budget reallocation

A conversion-optimised landing page that matches the specific promise your ad makes is not optional. If your ad promotes a 20% discount on a specific product and the landing page sends visitors to your homepage, you are paying for traffic that immediately bounces. Message match between ad headline and landing page headline is one of the most consistently under-addressed ROAS problems in South African ad accounts. For a deeper look at how landing page optimisation works alongside paid advertising, see our guide on conversion rate optimisation in South Africa.

Upsell, cross-sell, and loyalty programmes raise average order value and lifetime value, which lifts your blended ROAS over time without requiring more ad spend. A customer who spends 30% more per order makes every acquisition rand more efficient retroactively. A loyalty programme that brings a customer back for a second and third purchase transforms a marginal first-order ROAS into a strong account-level return. These are not ad account changes, but they directly determine what your ads are worth.

Budget reallocation is straightforward but requires monthly discipline. Audit your campaigns, identify the bottom 20% by ROAS, and shift that budget to your top performers. The heuristic is simple and the logic is directional rather than precise: budget follows performance. The practical caution is not to starve campaigns that are still in their learning phase or that are performing upper-funnel roles that do not show immediate ROAS but contribute to conversions downstream. Pull the plug on campaigns with sustained poor ROAS and no clear strategic reason for their existence. If you are looking for hands-on support with a Google Ads account serving the Pretoria market, see our advertising Pretoria service overview.

Frequently asked questions