How to Know If Your Ads Are Actually Working (Beyond ROAS)
If you’ve ever felt like your ad platform says things are going great while your business feels tense, you’re not imagining it.
This usually happens because the question you’re asking is different from the question your ad platform is answering.
Ad platforms are very good at answering:
“Which ads got credit for purchases?”
“Which campaign looks efficient inside this system?”
They are much worse at answering the question founders actually care about:
Are ads increasing profit and sustainable growth for my business?
This post is about answering that question — without becoming an analyst, building a warehouse, or trusting gut feel alone.
Step 1: Decide what “working” actually means
Before looking at metrics, define success. Ads can be “working” in one sense and failing in another.
Common definitions of “working”:
Profit-first: Ads are working if contribution profit increases.
Cash-flow-first: Ads are working if they pay back fast enough to not strain inventory or cash.
Growth-first: Ads are working if they acquire new customers at a sustainable cost.
Launch/awareness: Ads are working if they reliably generate qualified demand, even if profit comes later.
If you don’t choose the definition, ROAS chooses it for you — and ROAS usually defaults to credited revenue, not business outcomes.
Step 2: Understand why ROAS misleads (even when it’s “accurate”)
ROAS isn’t fake. It’s just incomplete.
ROAS answers:
“How much revenue did this platform claim credit for, relative to spend?”
That’s different from:
“Did ads improve the business?”
ROAS gets distorted by several structural issues:
Attribution ≠ causality
Attribution models assign credit, not cause. They can’t tell you whether a purchase would have happened anyway.
Returning customers
As brands grow, more purchases come from people already primed to buy. Platforms often take credit for these.
Discounts, returns, and margins
A 3.0 ROAS can be excellent or terrible depending on margin, refund rate, and discount depth.
Demand pull-forward
Some campaigns cause people to buy sooner, not more. Short-term ROAS looks great; long-term lift is flat.
Creative fatigue masked as “stable ROAS”
Efficiency can quietly deteriorate (higher CPMs, higher frequency) while ROAS stays flat.
ROAS is a diagnostic metric. It is not a verdict.
Step 3: Use a 5-signal framework that’s harder to fool
Instead of asking one metric to tell the whole story, look for alignment across signals.
Signal 1: Blended efficiency (business reality)
Pick one blended metric and track it consistently:
MER (Marketing Efficiency Ratio) = Total revenue ÷ Total marketing spend
Blended ROAS = Total revenue ÷ Paid spend (define scope once)
If platform ROAS is up but blended efficiency is down, ads are likely reallocating credit, not creating value.
Signal 2: New customer volume
Track:
New customers per week/month
New customer share of orders
Direction of new customer CAC (even roughly)
Ads that “work” but don’t bring in new customers eventually stall growth.
Signal 3: Contribution margin after marketing
You don’t need perfect accounting. Directional clarity is enough.
At minimum, approximate:
Revenue
– COGS
– shipping & fulfillment
– payment fees
– refunds/returns
– marketing spend
If contribution margin trends down while ROAS trends up, something is wrong.
Signal 4: Payback speed
Ask:
How quickly does paid spend pay back?
Is payback getting faster or slower as spend increases?
Slow payback isn’t always bad — but it must match your cash and inventory reality.
Signal 5: Incrementality (directional, not perfect)
Perfect incrementality testing is rare. Directional checks still help:
When spend increases, does total revenue move?
When spend pulls back, does revenue soften?
Do geo or channel shifts create noticeable differences?
If ads never move total outcomes over time, be skeptical.
Step 4: A simple weekly workflow (30 minutes)
You don’t need a dashboard marathon. Use this order:
Start with business reality
Total revenue
Total marketing spend
Blended efficiency (MER or blended ROAS)
New customers
Then check platform diagnostics
CPM trend
Frequency trend
Spend stability
CPA trend
Cross-check on-site behavior
Paid traffic conversion rate
Mobile vs desktop conversion
AOV and discount rate
End with one decision
Scale
Hold
Cut
Change creative or offer
If you finish with “we’ll watch it,” the system isn’t helping yet.
Respecting the platforms (they’re not the enemy)
Meta, Google, TikTok, and other platforms are excellent at:
controlling spend
diagnosing delivery issues
evaluating creative performance
managing auctions
They are not designed to answer profit-level, cross-tool business questions.
That’s not a flaw — it’s just scope.
The mistake is asking them to be your source of business truth.
A simple decision rule you can actually use
Here’s a clean heuristic that works surprisingly well:
Ads are working when blended efficiency and new customer volume are stable or improving — without contribution margin collapsing.
Everything else is supporting detail.
A clean next step
If your current process involves six tabs and a gut feeling, try asking your business one clear question instead:
“Are ads increasing profit, or just taking credit?”
Nurii is built for exactly this kind of question — connecting ad performance to the rest of your business so you can decide with confidence, not guesswork.
