From messy ad data to statistical proof.
Four steps. Under two weeks.

Every engagement follows the same process. No surprises. No scope creep.

01
Days 1–3

Data Audit & Unification

Share your ad platform exports, revenue data (Shopify, NetSuite, WooCommerce), email performance from Klaviyo, and any offline data. I audit for tracking gaps and quality issues, then unify everything into a single clean dataset.

What you get:

Clean unified dataset + data quality report documenting every gap, fix, and assumption.

Google AdsMetaTikTokAmazonMicrosoftKlaviyoShopifyNetSuiteBigQuery
"You don’t need clean data to start. I’ve built models from Google Sheets exports and CSV dumps. If you have 2+ years of weekly data, I can build a model."
02
Days 3–8

Model Build

Regression MMM calibrated to your specific channel mix, promotional calendar, and seasonality patterns. Not a template — trained on your actual revenue history.

The model accounts for:
  • Adstock decay — how long each channel’s impact persists after spend stops
  • Email lag — 3–14 day delay between send and revenue response
  • Diminishing returns — saturation curves per channel via Hill functions
  • Seasonality & external factors — decomposed from genuine channel impact
  • Cross-channel effects — amplification and cannibalization
What you get:

Trained model with documented assumptions, variable coefficients, and R² validation.

03
Days 8–12

Dashboard & Budget Optimizer

The model becomes a live interactive tool. Budget sliders, marginal ROAS curves, scenario planning, MER tracking. Built in vanilla JS/HTML — no dependencies.

What you get:

A dashboard your team actually uses. Same style as the live demo on the home page, trained on your real data. Hosted free on Cloudflare Pages. $0/month.

04
Day 12+

Handoff & 30-Day Support

Full code handoff. Walk through model assumptions, coefficient interpretation, how to rerun with new data. Complete independence — no ongoing vendor relationship required.

Rerun cadence:

Typically quarterly. Available as standalone reruns ($500–$1,500) or included in the Ongoing tier ($3K–$5K/year).

FAQ

Common questions, straight answers.

How is this different from Triple Whale or Northbeam?

They’re attribution platforms using pixel tracking and template models with monthly subscriptions. McFly Ads builds a custom statistical model on your actual revenue history using regression on aggregate data. No user-level tracking. Privacy-safe. You own the code. No monthly fees. Full comparison →

How many channels can the model handle?

Up to 14 in a single regression model. Most tools struggle past 6–8 before multicollinearity becomes unmanageable. Regularized regression handles correlated predictors, which is why it works where ordinary regression breaks down.

How much historical data do I need?

Minimum 18 months of weekly data. 2–3 years ideal. Consistent spend per channel during this period is important for reliable coefficients.

What if my data is a mess?

Expected. That’s Step 1. Data audit and cleanup is included in every engagement. If you can export spend and revenue by week, we can work with it.

How long does this actually take?

Under 2 weeks from kickoff to dashboard handoff. Enterprise firms quote 3–6 months. I have a proven process that gets applied to new data — the methodology doesn’t need to be invented from scratch.

Do I need a data science team?

No. The dashboard is built for marketers and founders. Self-explanatory sliders, plain-language labels. If you can use Google Analytics, you can use this.

What happens after the 30-day support period?

You have two options: run independently with full code ownership (most clients choose this), or continue with the Ongoing tier that includes quarterly reruns and strategy calls. Either way, you’re never locked in.

What’s the difference between MMM and multi-touch attribution?

MTA tracks individual users and breaks with iOS updates. MMM uses aggregate weekly data — no user-level tracking, privacy-safe, measures everything including offline. Full explainer →

Ready to see what’s actually driving your revenue?

Book a Free Consult →