Country-level profitability: finding the geos quietly earning — or draining

WHAT YOU'LL WALK AWAY WITH

Why every aggregate ROI number is a geo blend — and what the blend hides
How to build a country P&L: revenue side and spend side, per geo
Four geo archetypes and the move each one calls for
The three surgical actions: shift, cap, exclude — and a 20-minute monthly pass

01Every aggregate number is a blend

A campaign reporting 3.1x isn't running at 3.1x anywhere. It's running at 6x in one country, 2x in three others, and 0.7x somewhere you've stopped looking. The aggregate is an average — and as the scale/hold/kill framework puts it, averages are hiding places.

An "average" campaign is a winner and a leak wearing the same name.

Geography is the single biggest source of that spread, because both sides of the P&L move by country at the same time: eCPMs vary by 10x or more between premium and emerging ad markets, and install costs vary just as violently in the other direction. Two campaigns with identical creative and identical audiences can have opposite economics purely on geo mix.

10x+

TYPICAL eCPM SPREAD · PREMIUM VS EMERGING GEOS

02Build the country P&L

The unit of analysis is one row per country: revenue (all streams, net-to-you), spend, net, True ROAS. Sort by net, and the map appears:

NET BY COUNTRY · TOP 5 · DEMO PORTFOLIO

US$6,667

IN$4,889

ES$4,000

DE$3,556

BR$3,111

TOP 5 = 78% OF NET · 42 COUNTRIES TRACKED

78%

OF NET FROM 5 COUNTRIES · CONCENTRATION IS NORMAL — UNMANAGED CONCENTRATION ISN'T

Concentration like this is typical and fine — if you know it's there. The risk isn't that five geos carry the portfolio; it's a policy change, an eCPM slump, or an FX move in one of them arriving as a surprise.

03Reading the map: four geo archetypes

ARCHETYPE	PROFILE	THE MOVE
Premium engines	High eCPM, high CPI, deep ad demand (US, DE type)	Defend and feed. These fund everything — monitor daily, scale carefully.
Volume plays	Low eCPM, low CPI, huge audiences (IN, BR type)	Win on efficiency. Profitable at scale only with disciplined bids and strong fill.
Breakouts	Revenue trending up faster than spend (the ES pattern)	Move budget while the window is open — breakouts don't wait for quarterly reviews.
Leaks	Persistent sub-1x return inside otherwise healthy campaigns	Cap or exclude — surgically, without touching the campaign around them.

HONEST CAVEAT

Small geos produce noisy ratios. Set a minimum data bar — meaningful spend and installs per window — before a country is allowed to earn an archetype. Below the bar, it's "unclassified," not "a leak." Killing a geo on forty dollars of evidence is how breakouts get strangled early.

04The three moves

Shift. Reallocate budget from hold-tier geos to breakout geos. This is the highest-leverage UA action that requires no new creative, no new campaign, no new anything — just noticing in time. When ES runs +22% against baseline for two straight weeks, the shift is the decision.
Cap. A geo that underperforms but shows promise gets a bid cap: stay in the auction, stop overpaying for it.
Exclude. A persistent leak gets removed from targeting. The campaign around it instantly reads truer — often revealing that the "mediocre" campaign was a good campaign carrying a bad passenger.

Notice what's not on the list: killing the whole campaign. Geo problems get geo solutions; the scalpel beats the axe.

05The 20-minute monthly geo pass

TRY THIS · MONTHLY

Pull the country P&L — one row per geo, all streams, net-to-you, trailing 30 days
Confirm the top-5 concentration number; know it cold
Assign archetypes to every geo above your data bar
Execute one shift, review every cap, and re-test one exclusion from last quarter
Flag any top-5 geo whose trend broke — that's next week's investigation, not today's panic

Daily anomaly monitoring catches the ES-style breakouts between passes — that's the 4-minute morning's job. The monthly pass is where the map itself gets redrawn. Twenty minutes, because the ledger already did the assembling; the whole discipline collapses if each row takes an hour of CSV surgery first.