What Is a Good ROAS?
And 20 Other Questions Advertisers Ask Us
By Lindsay Telles, Founder and CEO
I get asked by publishers all the time what works and why. A breakdown like this helps everyone know where they stand.
Know your advertisers. Know your publishers. Know where you can get cheaper traffic that won't deliver the same quality, and know where you can get high-end traffic that may require a higher lead cost or CPA. It's all about knowing the relationships, how to work them, and how to make them scale together instead of constantly battling over pricing, volume, or creative. It's better to work as a team with your publishers, and that's what we try to do at Sellet Media: build a bridge that gets each person what they need to succeed.
Below are the questions advertisers, publishers, and media buyers ask us most, answered with 2026 data and what we've learned running campaigns across finance, health, consumer services, and lead generation.
If there is one thread running through every answer, it is this: the platforms have never been more automated. Bidding, placements, and audiences are increasingly handled by algorithms. So the edge no longer comes from hacks. It comes from knowing your real unit economics, sending clean signals, and refusing to confuse cheap leads with good ones.
Working With a Performance Partner
What does a performance marketing agency actually do?
It plans and manages campaigns around one measurable business action: a qualified lead, an appointment, a call, or a sale. The work spans the offer and audience, media partner selection, creative and landing pages, tracking, and post-launch review.
The real value is coordination. A good partner sees the full path from first impression to revenue and can tell you exactly where it is working and where it is quietly breaking down.
How is that different from traditional digital marketing?
Digital marketing describes the channels. Performance marketing describes how a campaign is managed and judged. A digital campaign may chase reach, impressions, or engagement. A performance campaign starts with a defined action and lets that outcome steer every dollar.
That is why performance marketers look past clicks and traffic to cost per qualified lead, cost per customer, conversion rate, revenue, and return on ad spend.
How do performance agencies get paid?
Common structures include cost per lead, cost per acquisition, revenue share, fixed media placements, a percentage of spend, and monthly management fees. Many programs blend media costs with strategy, creative, or tracking fees. The right model depends on the offer, the sales cycle, performance history, and how risk is shared.
Before you sign anything, get clear answers to four questions:
What exactly counts as a billable result?
Which services are included, and which are billed separately?
How are duplicate or invalid leads handled?
Who carries the risk if quality falls short?
Which channels work best?
There is no universal winner. Paid search catches people already looking. Paid social creates demand through interest and behavior. Native places an offer beside content people are already reading. Email and SMS placements reach established audiences through trusted publishers and list owners.
The strongest channel is the one that matches how your customer discovers information and makes decisions. And audience quality matters every bit as much as the channel itself.
What should I look for in a partner?
Transparency first. Ask which publishers or media sources will run your campaign, how creative gets approved, how lead quality is reviewed, and what would make the team pause or scale a placement. If a partner cannot tell you where your traffic comes from, that tells you everything.
This is exactly why Sellet Media works through direct, managed publisher relationships rather than an open network. Advertisers see where campaigns run and which placements are producing results.
The Numbers That Actually Matter
What is a "good" ROAS?
Any ROAS above your break-even ROAS. There is no universal number, no matter how often someone quotes one on LinkedIn.
Break-even ROAS is simply 1 divided by your contribution margin. At a 25% margin you need a 4.0 ROAS just to break even. At 50% margin, 2.0 gets you there (Tracify). The popular "4:1 rule" only equals break-even at that 25% margin. It was never an empirically established benchmark.
Vendor panel data puts 2025 median ecommerce ROAS somewhere around 2 to 3, but those are not audited standards. Tinuiti makes the same point: your break-even, not a generic benchmark, should set the target before any platform optimization begins.
For lead gen, ROAS is the wrong lens until you know revenue per funded or closed lead. Define the value of a qualified lead first. Everything else follows.
What LTV to CAC ratio should we aim for?
The classic benchmark is 3:1, with 1:1 as break-even (Stackmatix). But the ratio alone can lie to you. Payback period matters just as much.
A 5:1 ratio with a 36-month payback can sink a cash-constrained business faster than a 2:1 ratio that pays back in six months (Prospeo). And payback is stretching: median SaaS CAC payback hit 18 months in 2024, up from 14 the year before, per Benchmarkit (Rework).
Whatever you measure, use fully loaded CAC (sales plus marketing) and margin-adjusted LTV (Paddle). Anything less flatters the math.
What are normal Google Ads benchmarks right now?
WordStream's 2026 study of more than 13,000 US search campaigns is the cleanest snapshot available. The headline: cost per lead fell for the first time in five years, and conversion rates rose in 87% of industries.
Segment | CTR | CPC | Conversion rate | Cost per lead |
All industries | 6.64% | $5.42 | 8.18% | $66.69 |
Finance and Insurance | 9.83% | $3.39 | 2.64% | $74.44 |
Legal is the most expensive vertical per lead at $131.63.
Look at Finance. The highest click-through rate in the set, paired with the lowest conversion rate. People are curious enough to click, then lose confidence on the page. That gap is exactly where landing page trust and pre-qualification win, and where most money in the vertical is left on the table.
Measurement in a Privacy-First World
Can I still trust my attribution after the iOS privacy changes?
Treat platform numbers as modeled estimates, not truth.
Before iOS 14.5, IDFA covered 42% of iOS attributions. As of mid-2026, only about 10% are device-level and consented. Just 21% of iOS app marketers say they are confident in their iOS attribution at all (Kochava).
Meta has tightened things further. On January 12, 2026, it stopped returning 7-day view and 28-day view windows in its Ads Insights API. It now offers an Incremental attribution model, but that is still a platform grading its own homework, not a substitute for holdouts, geo tests, or media mix modeling (Enrich Labs).
Attribution, incrementality, or MMM: which one do I need?
All three, because each does a different job (EMARKETER):
Attribution assigns credit, but it does not prove causation.
Incrementality testing tells you whether a campaign actually caused lift.
Media mix modeling gives you the cross-channel view.
According to EMARKETER, 52% of US brand and agency marketers already run incrementality tests. And the barrier to entry just collapsed. Google lowered the minimum for its incrementality experiments, so a test that once cost upwards of $100,000 can now run for about $5,000 (Google Ads Help). Google's Meridian is a free, open-source MMM that can be calibrated with geo experiments.
Smaller advertisers no longer have an excuse. A $5,000 lift test is cheaper than one month of misallocated spend.
Do I really need server-side tracking?
Yes. Browser pixels get blocked, and the algorithms can only optimize on the signal you give them. Server-side tracking sends conversion events straight from your server to the ad platforms, bypassing browser restrictions and ad blockers (Cometly). For Meta lead gen, practitioners recommend running the Pixel and Conversions API together for a reliable optimization signal (AdLibrary).
In 2026, a weak signal is not a reporting problem. It is a performance problem.
Platforms, Creative, and Scale
Is targeting dead? Should I just go broad on Meta?
Mostly broad, with clean signals and strong creative doing the targeting. Meta has expanded Advantage+ audiences across campaign types and deprecated many interest categories, making AI-driven discovery the default (Adligator). Detailed targeting inputs are now treated as suggestions, not hard limits (Hawky).
One warning for regulated verticals like finance and health: broad only works when your conversion event reflects real quality. Optimize toward any form fill and the algorithm will happily find you the cheapest form fillers on the internet.
How many creatives should I be testing?
It scales with spend, and concept diversity beats a pile of near-duplicates.
Motion's 2026 benchmark of 578,750 Meta ads found that just 6% of ads carried most of an account's budget. Brands testing 10 or more new concepts a month saw 31% lower CPA than those testing fewer than five. UGC-style creative beat polished brand video for DTC by 27% on CTR and 19% on conversion rate. And most Meta ads start losing efficiency within two to three weeks of steady spend (EchoPulse).
A rough cadence by monthly Meta spend, per Superscale:
Monthly Meta spend | New concepts |
Under $10k | About 1 strong concept a month |
$25k to $50k | About 1 a week |
$50k to $100k | 2 to 4 a week |
How do I scale spend without breaking performance?
Scale only proven winners, and do it in steps. Common guidance is to raise budgets 10% to 20% every few days while metrics hold, after three to seven days of stable results (DICloak). Large jumps can push an ad set right back into the learning phase (Ads and Scale).
Stackmatix suggests a stricter standard: about 60 days of CAC at or below target, with low weekly variance, before any major increase (Stackmatix).
Does Performance Max actually work for lead gen?
It can, if you feed it deep-funnel conversions and give it time. Google recommends one to two weeks to exit learning, and up to six weeks for complex or low-volume setups. It also advises optimizing toward actions like Qualified Lead or Converted Lead rather than raw form fills. Value-based bidding needs at least 15 conversions in the last 30 days on the goal closest to a sale (Google Ads Help).
Practitioners suggest 30 to 50 conversions a month, and running PMax alongside core Search rather than instead of it (Search Engine Land).
Lead Quality, the Question Behind Every Question
How do you measure lead quality?
By what happens after the form is submitted. The useful signals are contact rate, qualification rate, appointments, sales, revenue, refunds or cancellations, and time to convert. Review them by publisher, placement, and creative whenever possible.
A lower cost per lead does not mean a better campaign. A more expensive source can deliver fewer leads, more customers, and a lower acquisition cost. Without downstream sales data, any media team, no matter how skilled, can only optimize for volume.
Native lead forms or landing pages?
Lead forms win on CPL. Landing pages win on quality. The only fair comparison is cost per qualified lead.
One benchmark across 80 accounts makes the point perfectly. Meta Lead Ads averaged a $14.20 CPL versus $24.60 for landing pages. But lead-to-MQL rates were 38% versus 64%, and cost per MQL ended up within 3% of each other (AdRiseLab). Use landing pages when the offer needs trust-building or real qualification (Prose Media).
Cheap leads that never fund are the most expensive leads you will ever buy.
How do I stop the algorithms from optimizing for junk leads?
Send CRM outcomes back to the platforms and bid on value. Offline conversion imports feed SQLs, opportunities, and revenue back into Google Ads, so bidding chases qualified leads instead of raw volume (Prose Media). The order matters: fix the plumbing first, then assign values, then change bid strategy (SalesX).
How big is the fraud problem in affiliate and lead gen?
Big enough that every program needs validation and partner-level QA. Fraud prevention vendor Scaleo estimates that 25% of affiliate campaign leads are fake or demonstrably poor quality, and about 17% of affiliate traffic is non-human. Treat those as directional rather than gospel. Global digital ad fraud losses are projected to pass $100 billion by the end of 2026 (Scaleo).
The usual suspects are cookie stuffing, click injection, misattributed branded traffic, and fake form fills (ClickFortify). The defense is unglamorous but effective: phone verification, sub-ID tracking, publisher scorecards, and paying on qualified outcomes. It is also why we built Sellet Media on known, managed publishers instead of anonymous traffic.
How long until I know whether a campaign is working?
Technical problems and obvious mismatches show up fast. Lead quality and profitability take longer, because they depend on traffic volume, the buying decision, your follow-up process, and how long a lead normally takes to become a customer.
The fix is to agree on the test period, budget, conversion definition, and review schedule before launch. Campaigns rarely fail from a bad first week. They fail when nobody defined what a good first month looks like.
The Question Everyone Is Asking This Year
Is AI search eating my paid clicks?
On informational queries, yes. But the bleeding appears to be slowing.
In Seer Interactive's data, paid CTR on queries showing AI Overviews fell from 19.70% to 6.34% between June 2024 and September 2025 (Search Influence). Seer's April 2026 update found the decline did not continue into early 2026, and organic CTR on those
queries rebounded 85% from December 2025 to February 2026 (Seer Interactive).
The most interesting finding: paid CTR was 91% higher when a brand was cited inside the AI Overview. That does not prove causation, but it does suggest brand authority and paid performance are becoming harder to separate.
It is also a quiet argument for channels AI cannot summarize away. A trusted publisher's email lands in an inbox the reader chose to subscribe to, no search results page required.
How Does This Apply by Vertical?
The principles above hold everywhere, but each vertical bends them a little. Here is how we see it play out across the sectors we work in.
Gold and precious metals
These are high-ticket, phone-closed sales with long consideration windows, so a lead's real value shows up in funded accounts, not form fills. Cost per lead tells you almost nothing on its own.
Trusted email and newsletter audiences tend to outperform cold traffic because the reader already trusts the publisher. Compliance review of claims and urgency language should be built into the creative process from day one, not bolted on at the end.
Finance and investing
Investors respond to credibility. Placements with established finance publishers, where the offer appears alongside content readers already rely on, usually beat broad paid social.
Track results by publisher, because quality varies widely from one audience to the next. Anything tied to a securities offering also comes with disclosure requirements that shape what the creative can say.
Credit cards and lending
Meta restricts targeting for credit and other financial products, so broad audiences and strong creative carry more weight here than anywhere else.
Approval rate is the number that matters. A cheap application that gets declined is worth nothing to the issuer.
Insurance
Insurance lives and dies on consent and contact rate. Leads lose value fast, so speed to call matters as much as volume, and consent language has to hold up to scrutiny.
Measure sources by bound policies, not quotes requested.
Health and supplements
Health offers face the tightest claims review on every platform, and every claim in the copy needs substantiation. Advertorial and native formats work well because they leave room to educate before asking for the sale.
Watch refund and cancellation rates by source. Aggressive traffic usually shows up there first.
The Bottom Line
The platforms now handle bidding, placements, and audiences better than any human could. What they cannot do is decide what a good customer is worth to your business. That is still the marketer's job.
Know your break-even. Send clean signals. Test creative relentlessly. Verify lead quality before you celebrate volume. And prove lift instead of assuming it. That is how Sellet Media approaches every program, and it is why we work through direct, managed publisher relationships where advertisers always know where their money is going.
Have a question we did not cover? Reach out at selletmedia.com. Odds are we have heard it on a call this week.
Sellet Media is a direct response media company that connects advertisers with publishers and list owners across finance, health, consumer services, and lead generation.
Sources & Further Reading
Vendor-published benchmarks from AdRiseLab, Scaleo, and Superscale are directional, not industry standards.
Comments