If you're trying to figure out how to improve ROAS on Meta ads while your CPMs keep climbing, you're dealing with a structural problem that needs a structural solution. Tweaking bids and budgets won't fix it. The only lasting path to better ROAS in a higher-CPM environment is to improve the efficiency of every other part of the conversion equation — and in some cases, change the metric you're optimising for entirely.
I've managed Meta ad accounts across DTC brands spending $15K–$100K/month through the CPM increases of 2024–2025. Here's what actually moved the needle — not what sounds good in theory, but what showed up as improvement in the numbers.
ROAS vs MER: Why ROAS Alone Is Lying to You
Before addressing how to improve ROAS on Meta ads, I need to challenge whether ROAS is even the right metric to optimise. In a post-iOS world, your Meta ROAS figure is based on incomplete attribution. The platform is showing you the revenue it can see — not the revenue it's actually driving.
ROAS as reported by Meta: Revenue attributed to Meta clicks and views, based on a 7-day click / 1-day view attribution window. Misses iOS conversions, multi-device journeys, and any purchase that happens outside the attribution window.
MER (Marketing Efficiency Ratio): Total revenue ÷ total ad spend. Doesn't depend on attribution accuracy. Tells you the actual return on your entire marketing investment. Harder to manipulate, harder to inflate, and a more honest picture of business performance.
The practical implication: Stop reporting ROAS as your only success metric. Report MER alongside ROAS. If your MER is healthy (typically 3–5x for DTC with 50–65% margins) but your Meta ROAS looks low, the issue may be attribution rather than actual performance. If both MER and ROAS are declining, you have a real problem that needs the levers below.
The Meta CPM Trend in 2026 — Context That Changes Your Strategy
Average Meta CPMs for eCommerce and DTC brands rose 20–35% through 2024–2025 and have stabilised at a higher base in 2026. The primary drivers: more advertisers competing for the same inventory, iOS reducing the efficiency of Meta's targeting (so the platform charges more to compensate), and the overall shift toward video formats which typically carry higher CPMs.
What this means practically: a DTC brand that was profitable at a 2.5x ROAS when CPMs were $15 may need a 3.2x ROAS to maintain the same margin at $24 CPMs — assuming everything else is equal. This is why ROAS targets need to be periodically recalibrated against current CPM levels and your current margins, not left as a fixed target set in 2022.
| DTC Category | 2022 Avg CPM | 2026 Avg CPM | Impact on Break-Even ROAS |
|---|---|---|---|
| Fashion / Apparel | $12–$18 | $22–$35 | +30–45% higher ROAS needed |
| Beauty / Skincare | $14–$20 | $24–$38 | +25–40% higher ROAS needed |
| Home & Kitchen | $10–$16 | $18–$30 | +25–35% higher ROAS needed |
| Health / Supplements | $15–$22 | $26–$42 | +30–45% higher ROAS needed |
Lever 1: Increase Creative Velocity
Meta's research shows that creative quality accounts for approximately 56% of ad performance outcomes. In a rising CPM environment this number matters more — high-relevance creative earns a lower effective CPM than low-relevance creative, even in the same auction. Good creative is, in effect, a CPM discount.
Creative velocity means producing and testing more creative assets more frequently. The objective: always have fresh creative in active testing so you're never in a situation where all your spend is on fatigued ads. The rhythm I use: minimum 3 new creatives per week on accounts spending $15K+/month, minimum 1–2 per week on accounts spending $5K–$15K/month.
The stacked creative testing framework: Test one variable at a time. Week 1: 5 different hooks, same product angle. Week 2: take the winning hook, test 3 different product angles. Week 3: take the winning combination, test 3 different CTAs or offers. This stacked approach compounds over 6–8 weeks into creative validated at every element — rather than a single ad that "seemed good" at launch.
Lever 2: Expand Your Audience
One of the counterintuitive but well-documented effects of the 2026 Meta algorithm: broader audiences often deliver lower CPMs and better ROAS than narrow, interest-stacked audiences. A large audience gives Meta's AI room to find efficient buyers — the algorithm seeks low-competition pockets of inventory within the pool. A narrow audience forces Meta to buy whatever inventory exists for those specific people, at whatever price they command.
If your prospecting campaigns are targeting interest-stacked audiences of under 2 million people, test the same creative against a completely broad audience (no interests, no demographic filters beyond age and location). In the majority of DTC accounts I've tested, the broad audience delivers equal or better ROAS at a lower CPM than the interest-stacked version.
Also consider geographic expansion: if you've been running US-only, test Canada and Australia — both have lower CPMs and strong DTC purchasing behaviour. English-language creatives transfer without modification and can dramatically improve your blended account efficiency.
Book a free 30-minute DTC Meta Ads review. I'll go through your account live, identify the ROAS drivers and drags, and give you a prioritised action plan you can execute the same day.
Book a Free DTC Meta Ads Review See DTC Case StudiesLever 3: Test New Offers and Angles
When CPMs rise, the pressure shifts to conversion rate — you need more of the people who see your ad to buy. The most direct lever on conversion rate is the offer and angle, not the creative format. A new hook on the same offer moves the needle moderately. A genuinely better offer moves it significantly.
Offer tests that consistently outperform in rising-CPM environments: (1) Bundles — a bundle offer at the same or higher total price often converts at a higher rate than individual products because it increases perceived value and AOV simultaneously. (2) Threshold incentives — "spend $75, get free shipping" increases AOV and makes the unit economics work at higher CPAs. (3) Limited-time framing — not fake urgency, but genuine seasonal or event-based windows that legitimately move purchase timing forward.
Lever 4: Optimise the Post-Click Landing Page Experience
Rising CPMs make every click more expensive. If your landing page CVR is 2.5% and you improve it to 3.5%, your effective CPA drops by 29% — with no change to ad spend, CPM, or bidding. In a high-CPM environment, post-click optimisation has proportionally higher value than when clicks were cheap.
The highest-leverage landing page improvements for DTC: page load speed (every 1-second delay on mobile reduces conversion by ~7%), social proof above the fold (showing reviews or customer count in the first scroll zone), and reducing the path from ad click to add-to-cart. For single-product brands, test sending traffic directly to the product page rather than the homepage — the product page almost always wins. Also test removing navigation from the landing page: a page where the only action is "buy" converts higher than a page with 8 exit points.
Lever 5: Switch Bidding Strategy
In rising CPM environments, default Lowest Cost bidding can sometimes outperform Cost Cap because it gives Meta's algorithm full flexibility to find efficient inventory. However, if your account has enough conversion data (50+ purchases per ad set per week), switching to a Cost Cap set 20–30% above your average CPA can protect profitability during CPM spikes.
The specific scenario where switching helps: if your CPMs are volatile — spiking during certain days or peak hours — a Cost Cap acts as a ceiling that prevents the algorithm from bidding at peak auction prices. This can reduce ROAS volatility even if it slightly reduces total delivery volume. Pair this with Lever 7 (dayparting) for maximum effect during high-CPM windows.
Lever 6: Use Advantage+ Shopping Campaigns
Advantage+ Shopping (ASC) is Meta's fully automated campaign type for eCommerce. When it works — with good creative, adequate pixel data, and a product catalogue — ASC consistently outperforms manually structured campaigns in ROAS and CPM efficiency. The algorithm has more flexibility to optimise placement, audience, creative delivery, and bidding simultaneously, giving it more tools to find efficient inventory in a high-CPM environment.
The conditions for ASC to outperform: minimum 1,000 purchases in your pixel in the last 180 days, at least 5 creative assets loaded, and a product catalogue connected with accurate pricing and availability. If these conditions are met, test ASC against your best-performing manual campaign with a split budget over 14 days. In roughly 60% of the accounts I've tested this in, ASC outperforms manual structure on ROAS within the first 30 days.
Lever 7: Apply Dayparting to Reduce CPM Exposure
Meta's auction prices are not uniform throughout the day. CPMs are typically 15–25% lower during overnight hours in your target market and 15–30% higher during peak evening hours (7–11 PM in the target timezone) when consumer attention and advertiser competition peak simultaneously. Dayparting lets you reduce spend during high-CPM windows and concentrate it during efficient ones.
How to implement: pull your hourly performance data from Ads Manager (Breakdown → Time → Hour of Day). Look for hours where CPM is consistently above your account average and CVR is below average — these are your efficiency drains. Schedule campaigns to reduce or pause delivery during the worst 4–6 hours per day. The CPM savings from not competing at peak auction prices can be significant for accounts where margin is tight.
Important caveat: Dayparting only works meaningfully at scale — typically accounts spending $300+/day. Below that threshold, the delivery reduction from dayparting may prevent campaigns from exiting the learning phase or cause too-limited delivery. Test dayparting on your highest-spend campaigns first before applying account-wide.
Lever 8: Reduce Creative Fatigue Systematically
Creative fatigue is one of the most direct causes of ROAS decline in established Meta accounts — and one of the most commonly ignored. When an audience has seen the same ad 4+ times, three things happen: CTR drops, CPM rises (lower relevance score), and conversion rate drops. All three are ROAS killers operating simultaneously.
The systematic solution: build a creative refresh calendar. Every 3–4 weeks, retire the 2–3 lowest-CTR ads from your prospecting campaigns and replace them with freshly tested creative. Never let a single ad run for more than 6 weeks to a cold audience without a test variant running alongside it.
Also audit for the difference between visual fatigue and conceptual fatigue. Changing the background colour on the same product shot is not a refresh — it's the same ad with different wallpaper. A genuine refresh means a new hook, a new product angle, or a completely different creative format (UGC vs studio vs demo vs testimonial). The format shift alone, even with the same product and offer, can reset audience engagement significantly.
The 30-Day ROAS Recovery Plan
If your Meta ROAS is under pressure right now, here's the sequence to run over the next 30 days:
| Week | Priority Actions | Expected Impact |
|---|---|---|
| Week 1 | Set up MER tracking alongside ROAS. Pull hourly CPM data and identify high-CPM dayparts to reduce. Audit creative age — retire anything over 6 weeks old from cold prospecting. | Immediate CPM reduction (5–15%) from dayparting. Honest view of actual business performance via MER. |
| Week 2 | Launch 5 new creative hooks against your best-performing product. Test broad audience vs existing interest-stacked audience on a 50/50 budget split. | Identify winning hook within 7–10 days. Audience test reveals whether broad or interest targeting is more efficient at current CPMs. |
| Week 3 | Optimise landing page CVR: test headline, remove navigation from key product pages, test CTA copy. Scale the winning creative from Week 2. | CVR improvement of 15–30% from headline/CTA test. Winning creative at scale driving lower CPM via improved relevance score. |
| Week 4 | Test a new offer (bundle, threshold incentive, or limited window). If pixel data supports it (1,000+ purchases last 180 days), launch an Advantage+ Shopping test campaign. | Offer test results visible in 7–10 days. ASC learning phase complete, performance data available for evaluation. |
Running all 8 levers over 30 days won't show results on every single one — but 3–4 will move significantly, and the compound effect typically produces a 20–40% ROAS improvement over the period. The key is executing systematically and measuring each change cleanly, rather than changing multiple things at once and being unable to attribute what drove the improvement.
I work with DTC brands spending $5K–$100K/month on Meta. I'll review your account, benchmark your CPM and ROAS against current market rates, and give you the specific levers to pull — in priority order.
Book a Free DTC Meta Ads ReviewFrequently Asked Questions
A good ROAS for Meta ads depends entirely on your product margins. For DTC brands with 50–60% gross margins, a 2–4x ROAS is typically the breakeven-to-profitable range. Brands with 70%+ margins can be profitable at 1.5–2x ROAS. Brands with 40% margins need 4x+ to be profitable. Calculate your minimum profitable ROAS as 1 ÷ (gross margin %) and use that as your floor — not an industry benchmark.
MER stands for Marketing Efficiency Ratio, calculated as total revenue divided by total marketing spend. Unlike ROAS, which measures revenue attributed to a specific ad or campaign, MER measures the overall efficiency of your entire marketing investment. MER is more reliable than ROAS in a post-iOS world because it doesn't depend on attribution accuracy — you're dividing actual revenue by actual spend. A healthy MER for a DTC brand with 50–65% margins is typically 3–5x.
If CPM rises while your CTR and CVR stay flat, your cost per purchase rises proportionally and ROAS falls. At $20 CPM with 2% CTR and 3% CVR, your CPA is approximately $33. At $30 CPM with identical rates, CPA rises to $50 — a 52% increase and a corresponding ROAS decline. The only ways to compensate are to improve CTR (through better creative hooks), improve CVR (through landing page optimisation), or increase average order value (through bundles or threshold incentives).
Cost Cap tells Meta to keep your average cost per result at or below a specified amount while still spending budget. Bid Cap places a hard ceiling on individual bid amounts, giving more control but typically lower delivery volume. Use Cost Cap when you have a clear CPA target and want to maintain volume with cost protection. Use Bid Cap when you need strict per-impression cost control. In rising CPM environments, set Cost Cap at 20–30% above your historical average CPA — not at your aspirational target.
The clearest signals: declining CTR week-over-week on the same ad, rising frequency on cold audiences above 4, CPM rising without audience changes (tired creative gets lower relevance scores and therefore higher auction CPMs), and a widening gap between impression volume and conversion volume. The rule I use: any ad running for more than 30 days to a cold prospecting audience needs to be tested against new creative regardless of whether performance looks stable — decline is often gradual before it becomes sudden.