Modular Creative / 3-3-3

Modular Creative is a production and testing framework where an ad is built from independently-swappable modules — hook, body, CTA — so any one element can be replaced without re-shooting the rest. The 3-3-3 shorthand from Pilothouse defines the starting test matrix: 3 concepts × 3 variations × 3 hooks.

The framework solves two separate problems at once: it generates enough creative volume to run clean concept and variation tests, and it does so at a production cost that doesn't require a full shoot per ad.

The structure

An ad has three structural slots:

• Hook — the opening 3–5 seconds. Earns the view. Measured by thumbstop and 3-second hold rate.
• Body — the middle. Deepens the argument introduced by the hook. The creative angle lives here.
• CTA — the close. Converts intent to click.

In a modular build, each slot is produced so it can be stitched with any compatible version of the other slots. Three hooks × one body = three ads, not one. Three CTAs × three hooks × one body = nine ads, not one.

This isn't about generating volume for its own sake. It's about making each element accountable. When the only thing that changes between two ads is the hook, you know the hook is what caused the difference in performance.

The 3-3-3 matrix

The canonical Pilothouse 3-3-3 breaks down as:

• 3 concepts — three distinct angle premises: e.g., problem-solution, founder-story, unique-mechanism. Each concept has its own body.
• 3 variations per concept — three executions of the same premise: e.g., different proof points in the body, different pacing, different talent.
• 3 hooks per variation — three opening sequences to test thumbstop against each body.

At full build: 3 × 3 × 3 = 27 ads from one production sprint.

In practice most brands start smaller — 3 concepts × 1 execution × 3 hooks = 9 ads — and expand once a winning concept emerges. The principle scales up or down; the structure is the same.

How to run it

1. Script modularly from the start. The hook, body, and CTA should be written as separate units with clean in/out points. A hook that requires a specific body to make narrative sense isn't modular — it's just a different ad.

2. Shoot with swaps in mind. Capture multiple hook openings in a single shoot day: different talent lines, different on-screen text overlays, different first-frame visuals. The same setup, different opening content.

3. Run the concept layer first. Launch one execution per concept (one body, one hook) to identify the winning premise. Don't burn your hook variants before you know which body they'll be attached to. See Concept vs Variation Testing.

4. Iterate the winning concept. Once a concept body is proven, swap in your three hook variants against that body. The winner of that hook test becomes the base for CTA iteration.

5. Log every module. Tag each ad in your tracker by concept ID, body variant, and hook variant. When you're looking at performance data three weeks later you need to know which combination you're evaluating.

DTC example

A direct-to-consumer apparel brand builds a modular shoot around three angle concepts: (1) social-proof (testimonials), (2) lifestyle (aspirational use), (3) founder-story (brand origin). For each concept they produce one body sequence and three hook variants — a visual hook, a text-on-screen hook, and a direct-address hook — totaling 9 ads.

Week one: concept test. The lifestyle body outperforms significantly. Week two: hook test on the lifestyle body. The direct-address hook wins. Week three: CTA test on the winning lifestyle + direct-address combo. The urgency CTA outperforms the standard shop CTA.

Nine ads, three rounds of testing, one clear winning combination — built in a single shoot day. Without the modular structure, reaching the same conclusion would have required three separate shoots and three times the spend.

The 3-3-3 also feeds directly into Creative as Targeting: three distinct angle concepts give the algorithm three different creative signals to route to three different audience segments, all from one production sprint.

When it backfires

• Forcing modularity on narrative ads — a long-form founder story that takes 60 seconds to build to the product reveal doesn't modularize cleanly. Not every format suits the framework. Match the production approach to the angle.
• Testing hooks on a losing concept body — if the concept body is weak, no hook will save it. Sequence matters: prove the concept before you invest in hook iteration.
• Losing track of the matrix — without systematic tagging, 27 ads become noise. You can't tell which combination you're looking at, so you can't replicate or eliminate it. The production discipline is inseparable from the tracking discipline.
• Treating 3-3-3 as a launch checklist, not an iteration loop — the system is designed to compound. Run it, find winners, build the next matrix from the winning modules. Brands that run it once and declare the winning ad "the creative" miss the iterative half of the framework.

Related concepts

• Concept vs Variation Testing — the sequencing logic that makes the 3-3-3 produce clean data
• Creative Angle — what populates the "concept" layer of the matrix
• Hook / Body / CTA — the three structural slots the modular system is built around
• Creative as Targeting — why concept diversity in the matrix also expands your algorithmic audience surface