Your ROAS looks fine. Your profit keeps falling.
ROAS counts the revenue the platform can see. It’s blind to COGS, discount depth and returns — so budget scales the ad sets that win on ROAS and lose on margin. Contribution flattens while every dashboard still shows green.
Read-only access · one month of data · no credit card
ROAS counts revenue. Your bank counts what’s left.
On one live account a hero SKU reported 3.1× ROAS and looked worth scaling. Run the same revenue down through COGS, discounts and returns, and the ad spend behind it cleared a loss. ROAS graded it a winner. The P&L never agreed.
Three gaps between a healthy ROAS and the money you actually keep.
01
ROAS stops at revenue
It counts the sale, not what the sale cost you. COGS, discount depth and returns all land after the number the platform grades itself on — so a 3× ad set can still sell below cost.
02
Returns and discounts arrive late
The refund and the promo code hit your margin days after ROAS has already called the ad set a winner and told you to put more budget behind it.
03
A healthy blend hides the losers
Blended ROAS or MER is an average. A few ad sets selling below cost disappear inside a 2.8× headline until contribution — not revenue — is the number you rank on.
Price every ad set on the margin that actually cleared.
Niti Lift resolves one customer across store and marketplaces, folds in real COGS, discount depth and returns, and reprices each ad set on contribution instead of platform ROAS. The call that looked like “scale” becomes “kill,” and you get a short daily list of what actually pays — with the reasoning shown on every call.
Platform ROAS 3.1× ✓
What Meta reports — looks like a winner
True contribution −6% · sold below cost
AOV $19 · COGS $14.10 · discount 18% · returns 7%
The same ad set. Priced on margin, “scale” was the wrong call.
An average can’t show you the ad sets losing money inside it.
Blended ROAS and MER net your winners and losers into one healthy-looking number. Niti grades every ad set on its own contribution, so the few selling below cost stop hiding behind the ones that aren’t.
All three are losing money inside the 2.8× average.
Every call recorded. Every outcome graded.
Kept ad set live — 0.42× → 2.06×
Meta overclaimed the loss ~4.9×. Scaled, not cut.
Cut prospecting 15%
Raise bids · 12 keywords
Retire fatigued creative
Shift 3% across channels
Straight answers.
- Isn't this just ROAS with COGS subtracted?
- No. Niti resolves one customer across store and marketplaces, credits cross-channel revenue the platform can't see, and factors real discount depth and returns — then grades whether the call it recommended actually worked.
- We already keep COGS in Shopify — isn't that enough?
- COGS is the start. Margin also moves with discount depth, returns, and the cross-channel revenue a single platform can't attribute. Niti prices on all of it, the same way on every call, so one ad set is comparable to the next.
- Does Niti touch my ad accounts?
- No. It's read-only to start and surfaces ranked recommendations with the reasoning behind each one. Nothing changes in your accounts without your one-tap approval.
- What does the audit cost?
- Nothing. Bring one month of spend and sales; we reconcile it against your real COGS and show which ad sets and SKUs sold volume at a loss. No integration, no credit card.
Bring one month of spend. We’ll show you what it actually earned.
The free margin audit runs on your own numbers — which ad sets and SKUs sold below cost, and the calls that would put contribution back.
Read-only · no integration · no credit card · 45-day outcome guarantee