The exact files we work from.

Two CSV files, six or more months of daily history, one currency. You get back an independent ROAS read for every campaign and a recommended budget allocation for the next period, as a report plus the full result tables.

What to provide

One media campaign CSV

Daily campaign delivery using all of the campaign columns below. One row per campaign per day.

Meta Ads Manager · Google Ads · TikTok · YouTube · DV360 · TTD · creator platform exports

One collective outcome CSV

Daily business results by outcome channel using all of the outcome columns below. This is the revenue side the campaigns are measured against.

Shopify export · CRM summary · retail sales file · finance report

Advertiser or brand name

Used on the report cover. Nothing else is needed.

Future planning budget, optional

The budget the allocation should plan against. If omitted, total historical spend is used.

Campaign CSV · required columns

Column	Required content
date	Daily date; YYYY-MM-DD recommended.
campaign_id	Stable unique campaign identifier.
campaign_name	Human-readable campaign name.
platform	Examples: Google Ads, Meta, TikTok, YouTube, DV360, TTD.
funnel_stage	Stable stage such as awareness, consideration, or conversion.
modality	Paid search, paid social, video, display, CTV, creator video, etc.
spend	Daily advertising or creator campaign spend.
impressions	Daily delivered impressions.
clicks	Daily clicks.
conversions	Daily attributed conversions; context, not revenue truth.
currency	One currency code, matching the outcome file.

Measuring influencer campaigns? Add three columns: influencer_id, creator_handle, and creator_tier (nano, micro, mid, macro, mega). Everything else stays the same, and you get modeled ROAS per creator plus a creator-level allocation.

Outcome CSV · required columns

Column	Required content
date	Daily date.
outcome_channel	Examples: DTC, Online Retail, Brick & Mortar, Social Commerce.
revenue	Daily channel revenue; primary model target.
sales	Daily units or completed sales for context.
orders	Daily order count for context.
currency	Must match campaign-delivery currency.

Samples

Input campaigns-sample.csv 30 days, synthetic data, exact shape

Input outcomes-sample.csv 30 days, synthetic data, exact shape

Output What you get back: sample report Rendered from a real model run on the sample data