Current answers to the marketing questions consumer businesses ask. Sign in free to read them all.1,898 sources, at least 3 behind every answer.

Show your work

Monday was math day in the Answer Book: two break-even formulas got fixed, subscription targets got a new floor, and two operators showed why that matters. Google kept shipping, and there are two new answers, including whether to test creative in a cheaper country.

By Scott Stroud, Editor

Hand-drawn curve that rises to a peak with a red flag on top, coin stacks growing under it and faint past the flag

Our answer on the metric that governs total spend changed its break-even math on Monday. Break-even is now 1 divided by your contribution margin before ad costs. The old version divided by a margin that already had ad spend taken out, which counted the ads twice and set the bar too high. The call around it holds: govern spend by contribution margin in dollars and keep adding spend only while that total rises. If you built a ROAS target from the old formula, it's worth ten minutes with a calculator.

The same fix reached Google. Our answer on setting a target ROAS now works out the floor for a first target from your margin after all variable costs and before ad spend, not gross margin, which puts the floor too low when shipping and fees are big. Two answers, one mistake, both corrected on the same day, which is the closest a database gets to an apology.

The subscription question

Mike Beckham of Operators told his podcast audience that a Meta campaign running at a 1.0 ROAS in a consumable category won't cover costs unless you have 100% margins (Operators). Then the show's account posted the exception: scaling at a first-order ROAS around 0.7 can make sense, but only if reorders repay the acquisition cost before cash runs out (Operators).

That lines up with Monday's change to our answer on setting a blended efficiency target. For subscription businesses, new-customer targets now start at first-order break-even and go lower, as Zach Stuck does, only when recurring revenue and cash can carry it. Going below break-even is no longer the starting point. Beckham's 0.7 sits in that lower zone, so the cash question is the whole game.

Ash Melwani posted a subscription-offer test at Obvi: conversion up about 10%, a subscription take rate around 40%, order value down 23% and profit per session down 12%, with payback on renewals not yet proven (Melwani). That post ran as a Triple Whale sponsorship, and Triple Whale sells ecommerce analytics. Our answer on the business metric that governs spend says the same thing his numbers do: judge by contribution margin per cohort and prove payback before you fund it.

Meta notes

  • Meta's on-device conversion counting starts November 12, and Pilothouse warns that any attribution test running across that date may shift for reasons that have nothing to do with the setting you're testing (Pilothouse). Pilothouse sells Meta ads management.
  • Our answer on keeping Meta's on-device conversion setting already says to switch it off on any pixel where a Black Friday test depends on Meta-reported conversions. Two days left to decide.
  • Ben Heath says turning ads off usually makes results worse, so leave low-spend ads running and add new ones (Heath). Barry Hott adds a caveat: Meta learns from small spend but, with little data, it favors attention-getting ads over efficient sellers (Hott).
  • Our answer on pausing ads or using automated rules makes the same call, and Hott's caveat is a good reason to check what the starved ads are actually selling.
  • Foxwell Digital, an agency, published its Black Friday guidepromo creative can go into existing campaigns, evergreen campaigns keep running, and backup campaigns get built early because Meta's ad approvals slow down at peak (Foxwell Digital). It also says not to cut budget over the slow Saturday and Sunday between Black Friday and Cyber Monday.
  • Foxwell Digital, an agency, published its Black Friday guideOur answer on budgets for seasonal peaks agrees on the prebuilt copy.
  • Northbeam, which sells attribution, says Meta's click-through rate across placements has climbed more than 40% year over year since January 2025, and reads it as Meta optimizing for clicks (Northbeam). If your CTR looks heroic, our answer on judging a creative test already keeps CTR out of the verdict.
  • Instagram's AI age estimation now runs worldwide, moving suspected teens with adult birthdays into Teen Accounts (PPC Land). Meta said nothing about ad delivery or audience sizes, so this is one to watch, not act on.

Google

Google is moving the Demand Gen opt-out from the Display Network down to the ad group level. Campaigns that opted out get rewritten automatically, Display-only ad groups inside them get paused, and from the week of December 1 the API rejects the old campaign-level setting (PPC Land). Google says serving won't change. If you run Demand Gen remarketing, check your ad groups before the holiday freeze, not after.

New in the book

Two new answers went up Monday. The featured one: testing creative in a cheaper market says do it only when that market's buyers convert like your main market's, treat its winners as candidates, move them over as the same post and confirm them on cost per purchase before scaling. Confidence is low, since the case rests on one operator's client. The other one is measuring channels you can't hold out, for national TV or a sitewide sale.

Four more calls changed besides the ones above:

Meta's on-device counting starts November 12, in 2 days. Black Friday is 17 days out.

Every fact links to its source. Opinions are the writer’s, built from the sources linked beside them.

Every issue