If your customers reorder, subscriber count is the business and monthly ROAS is a distraction. An established subscriber can be worth several multiples of a one-time buyer — on one account we manage, about six times — which means the price you can afford to pay for that customer is nothing like what a single month of ad efficiency would suggest.
The number that goes first in the report
Active subscribers. Then subscriber growth. Then revenue. ROAS moves down the page, because optimising it directly is how accounts end up buying cheap branded-search orders and calling it performance.
A True Portfolio CAC calculation separates what it genuinely costs to win a customer from those cheap orders. A ROAS floor and a stop-loss if CAC approaches LTV keep the trade honest in the other direction.
Retention is usually not the problem
It is worth checking before you spend anything. On one supplement account we audited, 30-day retention sat at 89.3% and 90-day at 70.6% — well ahead of the 50–60% typical for the category. The subscriber base was flat anyway, because acquisition was only matching attrition. That is an entirely different problem with an entirely different fix.
Making the second order easier than the first
Market-basket analysis every month, then combination offers built out of what people already buy together. Brand-Tailored Promotions aimed at cart abandoners and lapsed buyers. And a supply plan that moves subscriptions to a sister SKU before a stockout can cancel them, because nothing kills a subscriber base faster than an out-of-stock notice.
Full method and numbers: the supplement case study.