Scaling a DTC apparel brand from stalled ad spend to a profitable retargeting funnel

Situation

The brand had been running paid social for 14 months with a plateaued blended ROAS of 1.4 — below the breakeven threshold once COGS, fulfillment, and a 12% return rate were factored in. Revenue was flat quarter over quarter despite a 22% increase in ad spend over the same period, meaning efficiency was degrading, not just scale.

The founding team’s working hypothesis was creative fatigue — the internal response had been to brief new ad creative every 2–3 weeks. Six creative refreshes in, ROAS hadn’t moved. This was the signal that the problem sat upstream of creative.

Diagnosis

A full-funnel audit (traffic source → landing page → cart → checkout → post-purchase) surfaced three compounding structural gaps, none of which creative iteration could fix:

Strategy

Rebuilt the funnel bottom-up rather than iterating creative on top of a broken structure. Sequence mattered: fixing the catalog feed first meant every dollar spent afterward — including on the same “fatigued” creative — became more efficient by default, before a single new ad was made.

Phase 1 — Weeks 1–2: Infrastructure

• Rebuilt the product catalog feed with full attribute mapping (price, availability, variant-level images) and connected it to dynamic product ads.
• Verified every ad now deep-links to the exact SKU and variant shown in the creative, not a collection page.
  
  

Phase 2 — Weeks 3–4: Audience architecture

• Split traffic into three tiers: cold (lookalike + interest), warm (site engagers, non-purchasers, 30-day window), and retargeting (cart/checkout abandoners, 7-day window).
• Assigned distinct creative and offer logic per tier — cold sold the brand story, retargeting led with the specific abandoned product plus a time-boxed incentive.
  
  

Phase 3 — Weeks 5–8: Conversion rate optimization

• Cut checkout from 5 steps to 2, added guest checkout, compressed mobile load time from 4.1s to 1.8s.
• Added trust and urgency signals at the product-page level (stock count, verified review snippets near the CTA).
  
  

Phase 4 — Weeks 9–12: Scale and prune

• Reallocated budget weekly based on cohort-level ROAS rather than platform-reported averages, which were masking the retargeting tier’s true efficiency.
• Cut the two lowest-performing ad sets entirely rather than “optimizing” them — reinvested that spend into the top-performing retargeting segment.
  
  

Why it worked — the transferable principle

The structural fix (catalog feed + audience segmentation) was the lever, not creative optimization. Creative changes made after Phase 1 performed better simply because they were now reaching the right audience with the right product link — the same underlying creative quality, redirected through a fixed funnel, outperformed six rounds of “fresh” creative run through a broken one.