How the number is made, and how a result is measured.

1. How the score is built.

   • Forecast customer by customer from your store’s own history: how often they buy, whether they have lapsed, what they spend, and how much of it comes back.
   • Money expected later counts for less than money expected sooner. That is what “valued in today’s dollars” means.
   • A forecast can be checked against what customers went on to do. Tacet publishes that check.

2. What it counts, and what it leaves out.

   • In: your orders, customers, products and product cost, and refunds, over one read-only connection.
   • Out: payment fees, fulfilment and ad spend. Forward CLV is a product-margin figure, not profit. The customer P&L is where the fully costed number lives.

3. Where the number is weaker.

   • No product cost on an item: the figure is shown as a labelled range, not a single number.
   • Thin history: where a store has too few repeat purchases, Tacet says so.
   • One-time buyers: with no repeat history to learn from, their return propensity is left blank, not guessed.

4. How a playbook is measured.

   • About 20% of each playbook’s audience is held back before the first email, and the same customers stay out for the whole read.
   • The read date is fixed before anything sends, so a result cannot chase a flattering week.
   • The outcome is profit after discounts and refunds, against the held-back group. Never attributed revenue.
   • Every result carries a range. Inconclusive is never shown as a win, and a playbook that found nothing is retired.

The results, each with its range, are on the results page. The terms are defined in the glossary.