Build a 30-city × 5-section spreadsheet, hand-fill four of the five columns with verified local data, and let Claude only write the intro paragraph per row — no spintax, no template fills, no hallucinated landmarks.
Most marketing attribution is a $40K/year vendor contract answering a question you can solve in a Google Sheet. Here are the three models I have built for clients since 2014 — last-click, first-click, and time-decay — with the exact formulas, the column layout, and the one trap that catches every team the first time.
Most LTV models are either too simple (AOV x N) or need a data team and a Python notebook. This is the 5-input spreadsheet model I have used across DTC, SaaS, and subscription businesses since 2019 — with the exact ChatGPT prompt that builds a custom version for your business in under two minutes.
