Meta Ads Creative Strategy Guide: TOFU, MOFU, BOFU
How to split creatives by funnel stage, which test structure to pick, and how many concepts to ship per month.
Updated
Most accounts that plateau do not have a budget problem: they have ten variants of the same message. A creative strategy exists to answer three questions, in this order.
The reasoning is published below. The PDF adds the blank grids to fill in with your team.
1. Which funnel stage does this creative address?
| Stage | What the reader does not know yet | What the creative must do |
|---|---|---|
| TOFU | that the problem has a name | name the problem, not sell |
| MOFU | that your approach exists | show the mechanism, compare |
| BOFU | why you rather than someone else | remove the objection, prove |
The most common mistake: talking about the product at TOFU. Someone who does not know they have a problem is not shopping for a solution.
A reasonable starting split, to be adjusted on your own numbers afterwards: 60% TOFU, 25% MOFU, 15% BOFU in acquisition. It is not a rule, it is a starting point, because in acquisition cold audience volume dominates.
2. Which concept, and how many variants?
A concept is an idea: “the unboxing”, “the customer testimonial”, “the before/after comparison”. A variant is the same idea executed differently: another hook, another edit, another format.
The distinction is not cosmetic: it decides what you learn.
- Testing ten variants of one concept tells you which execution works.
- Testing ten concepts tells you which idea works.
An account starting out needs the second answer. Ship 3 to 5 concepts per month with 2 to 3 variants each, rather than fifteen variants of a single one.
3. Which test structure?
| Structure | What it isolates | When to pick it |
|---|---|---|
| One AdSet / One Ad | the creative, no interference | testing concepts against each other |
| One Concept / One AdSet | the concept, variants grouped | comparing ideas, not files |
| All-in-One | nothing, Meta arbitrates | when volume is there and you want speed |
| Insert in AdSet | nothing, you add to the live set | feeding an ad set that already performs |
The trade-off is always the same: the more you isolate, the more you learn, the more it costs. Every ad set has to reach enough conversions to exit learning. On a fixed budget, multiplying ad sets guarantees none of them will.
See One AdSet One Ad and All-in-One vs One Concept One AdSet for the detail of each structure.
When to cut, when to iterate
The useful signal is not one day’s ROAS. It is the two rates that separate the problems:
- Low hook rate (3s views ÷ impressions): the concept does not catch. Change concept, not edit.
- Good hook, low hold rate (15s views ÷ 3s views): the idea catches but the body loses people. Keep the concept, redo the execution.
- Frequency above 3 over 7 days: the creative is at end of life, whatever its ROAS.
What this guide does not do
It does not hand you creative angles. No grid replaces knowing your customers and their real objections. This guide organises production, it does not invent it.
It also does not tell you how much to spend. The 60/25/15 split assumes an acquisition budget that already exists; on an account spending $500 a month, the funnel question comes after the minimum volume per ad set.
Producing at that pace
Shipping 3 to 5 concepts a month with their variants, in three ratios, means thirty to fifty files to name, structure and launch. That volume is what makes manual upload untenable, and that is where Apogee comes in: the strategy stays yours, the production step is automated.