If you're trying to figure out how to improve ROAS on your paid campaigns, the answer is almost never "spend more." The average ecommerce ROAS dropped to 2.87:1 across the industry in 2025, while Meta CPMs climbed 19.2% in Q1 alone. Most accounts losing ground on return aren't running a bad strategy — they have three or four compounding problems that have never been fixed. This post identifies those problems precisely and gives you the fixes that actually move the number, based on what works across real accounts, not theoretical frameworks.
How to Improve ROAS: Start With What's Actually Broken
Before you test new creatives or restructure your campaigns, you need a clear picture of where the loss is happening. ROAS problems almost always originate from one of three places: bad measurement (you're not seeing the full picture), structural campaign issues (budget flowing to the wrong places), or a broken purchase path where clicks aren't converting. Treating a tracking problem as a bidding problem is the single most common mistake in paid media accounts.
The most common diagnosis I see is missing conversion data. An account will report a 1.8x ROAS because it's dropping 30–40% of actual conversions — purchases that happened via a different browser, a different device, or after a 7-day delay that the attribution window didn't capture. When the Conversions API is properly installed alongside the pixel and event data is validated, the actual ROAS often sits closer to 3.0–3.5x. The spend looked wasteful. It wasn't. The measurement was broken.
There's also a specific 2026 complication worth knowing about. Meta rolled out a major attribution update in March 2026 that changed how it reports conversions. Many accounts saw reported ROAS drop 15–30% on paper during that period — even though actual revenue hadn't changed. If your numbers declined sharply between February and April 2026, verify your backend order data against what Meta is reporting before assuming performance dropped.
Start every ROAS audit by comparing your ad platform's reported conversions against your actual backend order data for the same 30-day window. If the numbers don't match within 15%, you have a measurement problem, not a ROAS problem.
Fix Conversion Tracking Before Everything Else
Clean tracking is the foundation of every other ROAS improvement. Without it, you're running bidding algorithms on bad data, and those algorithms will optimize toward the wrong signals. Target ROAS bidding on Google and Meta's Advantage+ Shopping both require a minimum volume of conversion events to function properly — and if those events are incomplete, the system underperforms by design.
For Meta: you need both the browser pixel and Conversions API (CAPI) running in parallel. CAPI captures 20–40% more conversions than pixel alone by sending event data directly from your server rather than relying on the browser. Deduplication must be configured correctly — if it isn't, you'll get inflated event counts and the algorithm will receive conflicting signals that hurt delivery quality.
For Google: install enhanced conversions using first-party data — email, phone number, address — rather than depending solely on cookie-based matching. This matters more than it did two years ago because in-app browser tracking has degraded further post-iOS 17, and advertisers relying on cookie matching alone are operating with incomplete conversion data.
The practical test is simple: pull the Purchase event count from your ad platform for the last 30 days and compare it against orders in your ecommerce platform or CRM for the same period. The ad platform figure should be within 15% of actual orders. If it's off by more than that, stop optimizing bids and creatives until the data foundation is fixed. You're essentially steering by a broken instrument — and any optimization made on top of bad data has a high chance of making things worse, not better.
Setting Up Target ROAS Bidding Correctly
Once your tracking is clean, Target ROAS bidding becomes significantly more effective. Google recommends a minimum of 30–50 conversions per month before enabling tROAS bidding; below that threshold, the algorithm doesn't have enough signal and will either underspend or overbid. On Meta, Advantage+ Shopping campaigns need a similar volume of Purchase events to stabilize. The sequence is always: clean data first, then smart bidding, then creative and audience work on top.
- Validate pixel and CAPI are both firing and deduplication is active
- Confirm purchase event values are sending correctly in your account currency
- Check for double-counting — a common issue when both Google Tag Manager and hardcoded tags are both live
- Set conversion windows that match your actual purchase cycle (7-day click is standard; extend to 28-day for higher-consideration products)
Creative Testing: The Highest-Leverage ROAS Variable
If your tracking is clean, creative is the next place to look. The difference between a high-performing creative and a mediocre one can account for a 2–3x ROAS swing on identical targeting and budget. The data is consistent across platforms: campaigns with 10–15 active creative variations significantly outperform those running 3–5. Video ads outperform static images by around 40% in most ecommerce categories. But the real variable isn't format — it's testing velocity and creative refresh rate.
Most ecommerce brands are running 3–5 creatives at any given time. That's not enough variation for the delivery algorithm to identify winners, and it leads to creative fatigue faster than most advertisers realize. Meta's Andromeda AI update, which rolled out fully in early 2025, means that creatives saturate your available audience roughly twice as fast as they did in 2023. What used to run for 4–6 weeks before fatigue set in now often hits a wall in 2–3 weeks. If you're not refreshing creatives at that cadence, your ROAS will decline even if nothing else about the campaign changes.
A testing structure that works in practice looks like this: maintain 3 proven "always-on" controls that you know convert, run 4–6 challenger concepts in parallel, and refresh the challengers every two weeks regardless of how they're performing. When a challenger beats a control on cost per purchase over a statistically meaningful sample, it becomes the new control. Test one variable at a time — hook versus hook, offer framing versus offer framing, static versus video. If you change three elements in a creative at once, you can't isolate what moved the result.
"The best-targeted campaign in the world will stall if the creative stops earning attention. Creative fatigue is a structural problem, not a one-time fix — build your production process around continuous refresh rather than occasional intervention."
A DTC apparel brand using dynamic product ads on Meta with a structured creative refresh cycle lifted ROAS from 2.5x to 4.0x — with no changes to targeting or budget allocation. The creative system was the only variable. For a deeper look at how creative strategy interacts with campaign structure in real accounts, see the case studies where these approaches have been applied across different verticals.
Audience Strategy: Retargeting vs. Cold Traffic Ratios
Most ecommerce brands have their budget allocation backward. The strongest conversion signal in your data is always your existing customers and recent site visitors — and most accounts significantly underfund retargeting relative to the returns it delivers. Retargeting campaigns deliver 71% higher ROAS than prospecting campaigns on average. Past customers convert 50–70% better than cold prospects. Yet a significant portion of ecommerce ad budgets go primarily toward cold audience prospecting, because that's where the volume lives and where growth feels like it's happening.
A budget split that tends to work well for accounts spending $10K–$50K per month is 60–70% prospecting and 30–40% retargeting. At a 90/10 split favoring prospecting, you're leaving meaningful ROAS efficiency unrealized. The prospecting work matters — it fills the top of the funnel — but the retargeting side is where you close efficiently, and it's worth funding properly.
Inside retargeting, segment by actual user behavior rather than a single broad "past visitor" bucket. The segmentation that drives better results looks like this:
- Add-to-cart, no purchase (7 days): highest purchase intent, deserves the most budget and your most direct offer
- Product page views, no cart (14 days): strong intent, but use a different angle — social proof, reviews, or a secondary benefit
- Past purchasers (90–180 days): cross-sell or upsell based on the first purchase category
- High-CLV customers (lifetime spend more than 2x your AOV): worth spending more per click to retain — these are your most valuable accounts
First-party data is the foundation of all of this. According to Shopify's analysis of DTC brands using first-party audiences, some advertisers achieved 52% lower acquisition costs and a 5.6x ROAS by syncing customer data with Meta's Advantage+ audiences. The mechanism isn't complicated — you're just making sure the algorithm starts from a strong signal rather than cold interest targeting.
For Google, Remarketing Lists for Search Ads (RLSA) let you increase bids for users who've previously visited your site. A user who browsed your product page last week and is now actively searching your category keyword is worth substantially more than a cold query — your bid strategy should reflect that difference explicitly rather than treating both the same.
Landing Pages and AOV: Two Fixes That Compound Fast
ROAS isn't just a function of what happens inside your ad platform. It's a function of everything that happens after the click. Two variables that most media buyers underweight are landing page conversion rate and average order value. Both are within your control, and improvements in either one compound directly into ROAS without requiring any increase in ad spend.
On landing pages: your ad's click-through rate gets a visitor to the page. The page's conversion rate determines whether that click was worth paying for. A campaign generating a 2% landing page conversion rate and a 4% rate are running on identical ad spend — but the 4% rate produces twice the revenue per dollar of traffic. The ad cost is the same. The output is doubled. This matters more than most split-testing discussions suggest, because landing page CVR improvements stack on top of every targeting and bidding optimization you've already made.
Landing Page Fixes That Consistently Move CVR
The highest-impact landing page improvements for paid traffic tend to be predictable across categories. Message match — where the headline and imagery on the landing page directly mirror what the ad promised — is the most commonly broken link between ad and purchase. If your ad promotes a specific product benefit or a time-sensitive offer, the landing page should lead with exactly that, not a generic brand statement. Visitors coming from paid ads have zero obligation to search your site for what you told them you had.
- Message match: landing page headline reflects the exact promise made in the ad
- Load speed: under 2.5 seconds on mobile — every additional second costs conversion rate
- Social proof above the fold: ratings, reviews, or a trust signal visible without scrolling
- One focused CTA: send paid traffic to a page built for the specific offer, not a homepage or category page
On average order value: the higher your AOV, the better your ROAS looks at the same CPM. If you're spending $20 to acquire a customer who buys $40, you need a 2x ROAS to break even. If you can get that same customer to spend $65 through a post-add-to-cart upsell or a free-shipping threshold that's calibrated slightly above your current average order, your effective ROAS improves without touching a single campaign setting.
Free shipping thresholds set $15–$20 above your current average order value tend to drive meaningful basket-building behavior. Post-add-to-cart upsells, bundle pricing with modest discounts, and "frequently bought together" placements are the practical levers. None of these are complicated to implement, and each one raises the revenue ceiling of every click you're already paying for.
According to benchmark data from Onramp Funds' 2025 ecommerce ROAS analysis, the industry average sits at 2.87:1 — but the gap between average and top-quartile accounts (often 5x or higher) is rarely explained by media buying sophistication alone. The top performers have tighter landing pages, higher AOV from deliberate purchase-path design, and cleaner data feeding their algorithms. The combination compounds quickly.
ROAS isn't a single dial you can turn. It's a system: tracking accuracy feeds into bidding quality, which multiplies with creative performance, which gets converted efficiently by your landing page, which ultimately benefits from a higher AOV. Most accounts have at least two of these links broken. Fix the measurement first, then the creative, then the audience structure, then the post-click experience. The brands consistently running at 4x+ ROAS aren't doing anything exotic — they've just closed all the gaps that average accounts live with.
If you want to see how this plays out across real campaigns with actual spend data, the case studies on rohitmhatre.com cover accounts in DTC, F&B, and SaaS where these same principles were applied with measurable results.
Frequently Asked Questions
A good ROAS depends on your profit margins. The industry average across ecommerce is 2.87:1, and a 4:1 ROAS is generally considered strong. High-margin businesses can operate profitably at 2–2.5x, while low-margin or dropshipping models may need 5x or higher to stay in the black. The right benchmark is your own break-even ROAS, calculated from your actual cost structure — not an industry average. For a 50% gross margin, you break even at 2:1. For a 25% margin, you need at least 4:1.
ROAS = Revenue from Ads ÷ Ad Spend. If you spent $5,000 and the campaigns generated $20,000 in revenue, your ROAS is 4x. This is a gross revenue figure — it doesn't account for margins, returns, or cost of goods. For a profitability view, you need to calculate your minimum profitable ROAS based on your specific margins. A 4x ROAS on a product with 20% margins means you're still losing money after COGS.
Several factors can cause ROAS to drop in 2026: rising CPMs (Meta CPMs climbed around 19% in early 2025), creative fatigue that hits faster due to Meta's Andromeda update, or the March 2026 Meta attribution change that caused reported ROAS to fall 15–30% for many accounts even where actual revenue hadn't changed. Before diagnosing a strategy problem, compare your ad platform's reported conversions against your backend order data. If the numbers diverge, you have a measurement issue, not a performance issue.
ROAS measures revenue generated per dollar of ad spend — it's a top-line efficiency metric. ROI measures net profit after accounting for all costs including COGS, fulfillment, and overhead. A campaign can show a strong ROAS of 4x but a negative ROI if product margins are thin. ROAS is useful for comparing campaign performance and optimizing media mix. ROI tells you whether the business is actually making money from the advertising investment.
Tracking fixes can show impact within 7–14 days as the bidding algorithm relearns from better data. Creative improvements typically take 2–4 weeks to produce statistically meaningful results at normal spend levels. Structural changes — campaign consolidation, audience rebuilding, landing page overhauls — usually take a full billing cycle of around 30 days to stabilize. Most accounts see meaningful ROAS improvement within 4–8 weeks of addressing the core underlying issues in the right order.