Adstock • Upload CSV

Ingestion & Data Lab

Upload marketing campaign exports or explore calibrated enterprise benchmarks for Bayesian MMM optimization.

Back to Dashboard
Recommended $42.8M Spend

Enterprise Benchmark ($42.8M Spend)

730 days, 6 channels, Q4 seasonality.

2-Year Daily Search, Social, CTV, Video, Display, Retail Calibrated Saturation
↓ Download Benchmark CSV
Lightweight $800k Spend

Quick Test Sample ($800k Spend)

180 days, 3 channels, rapid iteration.

180-Day History Search, Social, Video Pre-Cleaned
↓ Download Sample CSVDownload Sample UI Dataset

Drop your CSV here or browse files

Accepts UTF-8 encoded marketing CSV exports up to 50 MB

Expected CSV Format
Auto-Prefix Channel Detection

Adstock ingests raw daily marketing performance metrics and automatically maps channels using canonical prefixes.

revenue Required

Daily total revenue or conversion volume (primary target variable for Bayesian model calibration).

spend_<channel> Required (1+)

Daily spend per channel (e.g. spend_google_search, spend_meta_social, spend_connected_tv).

date Optional

Daily timestamp (ISO YYYY-MM-DD) for longitudinal time series & lookback windowing.

impr_<channel> Optional

Daily impression volume per channel for media reach and CPM validation diagnostics.

is_holiday Optional

Binary flag (1 or 0) isolating exogenous holiday and promotional demand surges.

Model Guidance: While Adstock can train on 30+ daily records, 90+ consecutive days (ideally 1–2 years) are recommended to reliably estimate adstock decay rates and non-linear Hill saturation limits.