Online store: organic flat 6 months — ads running, catalog full, GA4 flatlines

Quarterly review: the flat line that silenced the room

The head of e-commerce opened GA4 on the projector. Organic traffic chart from March to August: nearly a flat line. Next slide: Google Ads spend up 22% this quarter. ROAS? Slightly down.

“We have 1,200 SKUs. Ads are running. Content team published 8 blog posts last month. Why is organic not moving?”

That question had been asked in three consecutive meetings. Answers were always the same: try more content, raise budget, wait for season. Nobody had concrete data.

Two battlefields, both stuck

Buyers in this category research on Google (specs, comparisons, reviews) and increasingly in AI answers (“compare running shoe A vs B for marathon”). This store had full product range and active ads — but organic visibility was stagnant on both fields.

Internal team prompt: “Best online store for [category] in Indonesia”. Competitors with tidy comparison pages and structured FAQ? Named. This brand? Never.

What they had already tried (and why it was not enough)

Ad budgets raised each quarter — paid traffic up, organic flat. New SKUs added 40/month with no category strategy. Freelancers wrote generic blog posts not linked to converting product pages. Popular SEO plugin installed — green dashboard, traffic unmoved.

The problem was not lack of products or ads. The problem was hundreds of pages blocking the important ones, and nobody measuring whether AI was already sending traffic.

Audit findings: 312 pages eating crawl budget

Once we audited, the numbers were clear:

847 indexed pages — 312 zombies. Under 100 words, zero clicks in 90 days. Search engines spent crawl budget on pages that never brought buyers.

Cannibalization. Three URL pairs fighting the same query — “men’s running shoes” vs “running shoes men” vs duplicate category filters. Google could not pick a winner; all ranked mediocre.

LCP 4.2s on best-sellers. PageSpeed score 62 — not red, but slow enough to lower crawl priority on the pages that mattered most.

GA4 did not separate AI-referral. Traffic from perplexity.ai mixed with direct. Team had no idea if the new field was already signaling.

The controversial decision: archive 280+ zombie pages first, do not produce new content. Marketing initially pushed back — “those are already indexed.” Data won: dead pages were blocking pages with conversion potential.

What we ran (Oura Growth + Atlas)

Growth at Oura starts with diagnosis, not volume:

1. Clean structure — archive zombies, merge cannibalization, focus crawl on ~560 valuable pages.
2. Fix CWV on 12 priority product pages — LCP target under 2.5s.
3. Rebuild internal linking — category hubs → converting products → supporting answer-first content.
4. 8 answer-first content clusters for real buyer questions: specs, comparisons, shipping, warranty.
5. OAVF — product entity, FAQ schema, AI-referral segment in GA4.
6. Monthly Atlas loop — every sprint backed by data, dry-run before production, per-cluster impact verification.

How to read the numbers below

Patterns we see in similar Growth engagements:

• Organic starts moving 4–8 weeks after zombies archived and cannibalization resolved — not after new content.
• PageSpeed climbing to 90+ usually impacts priority pages first, then spreads to categories.
• AI-referral often appears after structured comparison pages and FAQ — separate from organic in reports.

Once momentum returns, next iteration focuses on highest-margin product clusters — not chasing every keyword at once.