This program runs 80-100 parallel cells - one cell is one (business group, country, category, brand) combination, e.g. one brand of hair care in Indonesia. Every cell travels the same protocol: three phases wrapping ten stages:
| Phase | Stages | Clock |
|---|---|---|
Preparation | Design | Variable timeline; ends with the Design and Data sign-offs |
Build | EDA & ADS prep, Modeling - MMM, Modeling - PnP, Optimization, Validation, Insights & Storyboarding | Committed 8-week window, started by the two sign-offs |
Presentation | Business Presentation, Market Presentation, Final Sign Off | Scheduled around the market's calendar |
A cell's headline status moves through: Awaiting Confirmation → In Preparation → In Build → In Presentation → In Support, with On Hold as the exception state and Closed as the terminal one. One nuance trips everyone: interim results presented DURING the 8-week build do not flip the status to In Presentation - that flip happens only when all six Build stages are done and just the Presentation stages remain. Module 2.2 goes deep on this.
The deliverable arc mirrors the phases: the Data Summary (per-feed readiness plus a narrative one-liner, Preparation), then the analytical dataset out of EDA & ADS prep (Build), then the Model Read-out Deck - an analytical workbook turned client story (Presentation). Module 2.5 walks that last transformation.
Seven numbers do most of the talking in a read-out. Formulas are the program's own (Pre-Read slide 8):
| KPI | Formula | The decision it answers |
|---|---|---|
ROI | Incremental Revenue / Spend | Did past spend on this channel pay back? (Backward-looking) |
mROI | Change in Revenue / Change in Spend | Should the NEXT dollar go here? (Forward-looking, from the response curve) |
CPM | Spend × 1000 / Support | Am I buying impressions efficiently? |
Effectiveness | (Incremental Revenue × 10) / Support | How hard does each impression work, independent of its price? |
ASSR | Total advertising spend / Sales revenue (%) | Is my overall media weight sane for a brand this size? |
Contribution % | Driver volume / Total volume | What share of my sales does each driver produce? |
Due-to % | (Driver volume P2 - P1) / Total volume P1 | What explains my growth vs last period? |
ROI vs mROI: the investment rule
The pre-read's worked example is worth internalising verbatim. ROI says what a channel returned on average; mROI says what the next unit of spend would return, read off the response curve. The rule: mROI > 1 - invest more; mROI = 1 - hold; mROI < 1 - reduce and reallocate. A channel can carry a splendid ROI of 3.5 and a mROI of 0.6 at the same time - it paid back handsomely on average and is now saturated at the margin. Reading those two numbers together is the whole art; Track 3.3's interactive optimizer lets you feel it.
Contribution % vs Due-to %: two different client questions
These are the most commonly confused pair in client conversations (Pre-Read slide 10 exists precisely because of it):
- Contribution % answers "what share of my volume does TV drive?" In the worked example, TV drives 57M of 201M total volume: 57 ÷ 201 × 100 = 28%.
- Due-to % answers "how much of my GROWTH is due to TV?" TV-driven volume moved from 27.2M to 29.8M against a period-1 total of 95M: (29.8 - 27.2) ÷ 95 = 2.7% of the total 11.6% growth.
A channel can be a big contributor and a negative due-to at once (it drives a lot of volume, but less than last year - the pre-read's Organic Influencer row does exactly this: 32% contribution, -1.3% due-to). When a client asks "what is driving my growth", the answer is the due-to table, not the contribution one.
UL_Rapid ROI_Pre-Read_Document 1.pptxslide 8 (KPI glossary with formulas), slide 9 (mROI worked scenarios), slide 10 (Contribution vs Due-to worked example), slide 11 (Effectiveness/CPM interpretation callouts)- Root
MASTER_SCHEMA.md:tbl_Stages,tbl_Phases,tbl_CellStatus(Control tables, admin-owned and authoritative), Delivery sheet section (In Build vs In Presentation semantics)