Fractional CDAO
Stands up or resets a data, analytics, and AI function: the hires, the platform, the governance, and the operating cadence with the executive team.
You hold the function, its platform, and its cadence, running without Called Shot.
Fractional CDAO · Measurement · Audience models · AI systems
Every engagement opens with a one-page scope: the result, how it will be read, the date of the read that settles it, and the price. Then the work.
About the one result you need, by when, and how you measure it today.
It names the result, the measure, the date, and the price. If the data you have cannot support a called shot, the scope says what would make one possible instead.
On a fixed cadence against a holdout, the group the program leaves untouched so the result has a baseline. Each read is reviewed with your team before it travels.
You hold what was built: the IP, the pipelines, and the results.
Stands up or resets a data, analytics, and AI function: the hires, the platform, the governance, and the operating cadence with the executive team.
You hold the function, its platform, and its cadence, running without Called Shot.
Builds or audits the marketing mix model, the MMM that reads what each channel returned, and the attribution program behind the media plan: the data standards, the refresh pipeline, the holdout design, and the readout the executive team acts on.
You hold the model, the pipeline, and the readout.
Builds the models that say who to reach and when: days to next purchase, likelihood to churn, likelihood to become a high-value customer. Scores every customer on a schedule and pushes the audiences into the places the money is spent, the CRM program, digital media, CTV, and social, timed to each customer's window.
You hold the model, the scoring pipeline, the audiences in your own platforms, and the holdout that shows what they moved.
Designs the AI features and automated workflows that run in production: LLM features grounded in your own data, multi-step pipelines that act on a schedule, each reviewable by the people accountable for it and built to a cost per user set in advance.
You hold the system, the tests it is judged against, and its cost per user, in writing.
A survival model for Belle Tire that predicts each customer's next purchase and brings it forward. In production since November 2024, read against a holdout the whole way.
Read the caseA hierarchical Bayesian marketing mix platform serving multiple clients, with every model release versioned and the reads used directly by planning teams.
Read the caseFour grounded editions a week for 10,000 ballclubs, after a 50% cost reduction, with every draft checked before it publishes.
Read the caseMarketing and analytics leaders at brands and agencies who need a senior operator for a defined stretch rather than a permanent hire. Executives who want a second opinion on a measurement or AI program before committing to it. Teams with a number to move next year and no one yet willing to name it.
Called Shot is founded by Courtney Perigo, a data, analytics, and AI executive in Chicago: twenty years building the models and the teams behind calls like these, on the agency side and the client side. The case studies, the writing, and the CV are at courtneyperigo.com.
A short conversation about the one result you need, by when, and how you measure it today. If the data cannot support a called shot, Called Shot says so, and says what would.