Creative Testing Pipeline & Naming Convention (2026)

Most DTC brands don't have a creative problem, they have a measurement problem. They ship dozens of ads a month, yet can't say which hook, angle or format drives sales, because nothing is labelled consistently. For Canadian e-commerce teams competing against US giants on the same Meta and TikTok auctions, that blind spot is expensive. This tutorial shows you how to build a repeatable creative testing pipeline and a naming convention that turns every ad you launch into usable data.

Step 1 — Design a naming convention you'll actually keep

A good convention is short, human-readable and machine-parseable. Encode only the variables you'll want to filter or pivot on later: launch date, funnel stage, concept, hook, format and iteration. Separate fields with a single delimiter (an underscore) and never reuse it inside a field. Example: 2026-08_TOF_UGC-testimonial_hook-price_9x16_v03 tells you at a glance the date, the top-of-funnel stage, the concept, the hook, the aspect ratio and the version.

• Date: the YYYY-MM of launch, so you can sort chronologically.
• Funnel stage: TOF, MOF or BOF, to compare prospecting versus retargeting.
• Concept: the big idea (UGC testimonial, founder story, product demo).
• Hook: the angle of the first three seconds (price, problem, social proof).
• Format and ratio: static, video, carousel, 9x16 or 1x1.
• Version: v01, v02 and so on, so iterations never overwrite their parent.

Step 2 — Run the four-stage testing pipeline

• Brief: write a one-line hypothesis per creative, such as 'if we lead with price, TOF CTR beats our 1.2% baseline'.
• Launch: group three to five variants in one testing campaign with a fixed daily budget and an identical audience.
• Read: wait for statistical signal (50+ link clicks or three purchases per variant) before judging; ignore day-one noise.
• Scale: move winners to a separate scaling campaign, iterate one variable on near-winners, and archive losers with a note.