Orbit
UA Budget & Goal Control
Last sync
—
Formulas —
CALCULATIONS_MAP.md

Loading…

Looker Looks —
ID Title Table Rows Cohort window Used as
Spend & Engagement —
window: matured cohort (install 7–14d ago, D7 final)
Campaign Spend Installs CPI D7 events Event %
CPE Projection
windows: matured [today-14…today-8] · recent [today-7…today-1]
Campaign D0|matured D7|matured D0|recent Ratio (P) CPE
ROAS Goal Budget —
The spend ceiling implied by each game's ROAS goal — organic revenue covers the gap between the D7 goal and predicted media ROAS. Compare Max Total Spend against actual Spend Last 30 Days. Shows ROAS-goal games (set per game on the Games tab).
Game D90 / D60 Unified Goal (P) Reta Profit D90 / D60 · 30d (P) Org Rev D7 · 30d (p) Paid ROAS D7 UA ROAS D90 / D60 Goal ROAS D7 Goal Max Total Spend Spend Last 30 Days Max Daily Spend Yesterday Spend Budget Variance Budget Difference
Fixed Monthly Budget —
Total monthly budget —
Games are split into groups. Each group has its own monthly budget and its own horizon, and its budget is split across its games so they all land at the same projected net ROAS at that horizon (net = gross × (1 − withdrawals − Ext costs − GC cost), Apollox's rates per game, listed on each group's card). Per-game caps are derived, paced against month-to-date spend. Pick a game's group in the Group column; shows fixed-budget games (set per game on the Games tab).
Game Group (P) Org Rev D7 · month (p) Paid ROAS D7 Max Total Spend (p) Net ROAS @ group (p) ROAS D7 (p) ROAS D90 (p) ROAS D180 (p) ROAS D360 Spend This Month Max Daily Spend Yesterday Spend Budget Difference
Gross Net Gross Net Gross Net
Paid ROAS —
Predicted D7 media ROAS — the number every budget solve pivots on. Smart (live since 2026-08-13) completes each of the last 14 cohort days from the revenue it has already matured — the cohort n days old is read at rung D(n−1) and completed with that rung's matured multiple, so days 8–14 are actual D7 with no extrapolation at all. Classic is the old method: re-extrapolate the whole live window from its D0 (the three columns to its left). Smart is what feeds (p) Paid ROAS D7 in the budget tables above; classic is shown beside it during the changeover so it's clear what moved.
Game Live (L14D) Matured (L30D, +7d) (p) D7 · classic (p) D7 · smart
D0 D0 D7
Predicted Organic D7 Rev —
Predicted organic D7 revenue per day — the organic analog of Paid ROAS. The live value is (p) D7 /day · smart: the mean completed D7 of the last 14 cohort days (actual once a cohort is 8+ days old, revenue so far × the matured D7÷Dk multiple before that — Looks 157/158). The classic figure beside it — a recent day's organic D0 revenue (Look 141) scaled by a matured cohort's D7÷D0 decay (Look 140) — is the retired method, kept for reference and as the fallback. Multiply by the horizon — 30d for ROAS Goal Budget, month-days for Fixed Monthly — to feed the budget tables. Click a row for the day-by-day picture: actual organic D7 revenue per cohort day (solid) vs the smart cohort-completion prediction for cohorts not yet matured (dotted) — Looks 157/158.
Game Live (L14D) Matured (L30D, +7d) (p) D7 /day · classic (p) D7 /day · smart
D0 /day D0 /day D7 /day
Multiples —
Chained cohort multiples per revenue stream — paid (Has Cost = Yes, Looks 163/161/159) and organic (Has Cost = No, Looks 164/162/160). Same method as the Monthly tab (since 2026-10-25): six short steps D7→D14→D30→D45→D60→D75→D90, each from the cohorts that reached it in the last 30 days (Looks 165/166), so a change in the revenue curve shows up within weeks. The columns group them into D7→D30, D30→D60 and D60→D90, then D90→D180 (since 2026-11-02: the same walk over the tail rungs 105…180, Looks 168/169, cohorts older than 181 days — no fallback, so a young game reads —). A game with too few matured cohorts falls back to the old wide windows (31–149 / 61–209 / 91–269 days, Looks 163–159) — hover a value to see which. The (p) columns multiply the legs left to right; at cap is the blend of the two streams at the game's cap — its Fixed Monthly cap, or the solved Max Total Spend for a ROAS Goal game.
Game D7→D30 · paid D7→D30 · organic D30→D60 · paid D30→D60 · organic D60→D90 · paid D60→D90 · organic D90→D180 · paid D90→D180 · organic (p) D7→D60 · paid (p) D7→D60 · organic (p) D7→D90 · paid (p) D7→D90 · organic (p) D7→D180 · paid (p) D7→D180 · organic D7→Dh · at cap
Long-term Multiples —
Long-horizon growth — D90→D360 multiples per game and per channel (IAP / IAA), plus the IAP/IAA D30 revenue split that re-weights them into the (p) D90→D360 blend feeding (p) ROAS D360.
Game IAP / IAA D30 (L60) IAP / IAA D30 (L420) D90→D360 IAP (L420) D90→D360 IAA (L420) D90→D360 (L420) (p) D90→D360
Change Logs —
Time User Area Change
Goal vs Reality Factor —
Campaign (BQ code) Game Type Goal now D7 Goal (window) Real D7 Factor Days Applied
Factor = Real D7 ÷ D7 Goal (window) — both sides anchored to the same cohort dates and weighted by the same per-day spend. Real D7 is the delivered rev D7 ÷ cost of each cohort day in the window shown above: the older week is matured (measured revenue, every cohort past D7), and the newest 7 cohort days — marked S in the detail — are smart-completed from the revenue they have already matured, with the same rung multipliers as (P) D7S. (Rows marked agg have no per-day coverage yet and fall back to Look 44's matured-window aggregate.) D7 Goal (window) is the goal that was actually in effect on each of those same days (point-in-time, from the goal-change log), put on the D7 scale by type — D7-like: the Goal itself, D28: Goal ÷ D7→D30 — and spend-weighted by that day's cost so it lines up with Look 44's own spend-weighted ratio. Goal factors are deliberately ignored in that denominator, so an applied factor can't drag its own ratio toward 1. Below 1 = the campaign delivered less than the bar it was given (punish candidate); above 1 = more (promote candidate). Selecting Reality (auto) on the row in Mapping applies exactly this factor: D7 Goal = raw D7 basis × factor. Click a row for the day-by-day goal and spend behind the window.
Game D0 Health —
Paid · D0 ROAS
Game D0 Last 7 Days (-Yesterday) D0 Yesterday Δ vs 7D
Campaign D0 Health —
—
Campaign Game Spend D0 ROAS
Yesterday Avg/day · prior 7d Δ Yesterday Prior 7d Δ
Today vs Yesterday —
—
Today —
(P) Today full day —
L7D pace —
Yday full day —
Cumulative spend by hour, UTC — today (solid) vs yesterday (dashed). Only fully-reported hours are compared: the solid line stops at the last hour every selected network has completely reported (AppLovin runs ~45m behind, Unity ~95m, Mintegral ~2.5h), and the gap up to the "now" marker is withheld partial data — not missing spend. Narrowing the network filter to a faster-reporting network extends the window. Adjoe has no hourly reporting: Orbit samples its "cost so far today" hourly (and on each refresh) and accumulates the deltas into hours itself — its curve is exact at every sampled hour, with at most an hour of shape smoothing in between.
Games —
Order Name Store Link Horizon Enabled Budget Type Manual Unified Goal %
Add game
The Name should match how this game's row appears in Look 49 (case is ignored). Disable a game to hide it from every tab (including its Mapping columns) while preserving its campaign data on disk. To delete a game, first clear its campaign codes from every row in the Mapping tab — re-enable it first if it's disabled, so its columns show.
Campaign Map —
Label Type
Conversion Factors — Confidence: ≥2K 500–2K <500
Campaign Chain (p) D7→D60 (p) D7→D90 IAA / IAP (D30)
D7→D30 (L60) D30→D60 (L90) D60→D90 (L150) D90→D180 (L240) D180→D360 (L420)
Campaigns
— tracked
Alerts
— active alerts
Goal —
D0 drop —
Bidder —
Data —
Max Daily Spend
— usd / day
Yesterday · vs cap
— usd — off cap
AppLovin · % of L7D
— of spend