24LED

The client

24LED.ro sells LED strips, aluminium profiles, power supplies, floodlights, industrial fixtures and chandeliers, with a showroom in Bucharest and a shop on the Gomag platform. Its customers are installers, architects and owners doing their own project, and the decision is made on specifications: watts, lumens, voltage, IP rating. It is technical lighting, bought by people who know what they are looking for.

The average order is around 357 lei, and gross margin is 33%. The Romanian LED lighting market grows by roughly 9.7% a year, but the field is crowded: iluminat-ieftin.ro, luminiterasa.ro, eMAG and Temu bid on the same searches. Over 1,100 products are listed in Google Shopping.

The project is ongoing, since November 2024. This page is updated as measured results come in.

The challenge

The client’s target for Google Ads was a ROAS of 4 to 5. The first question we asked was whether 4 is enough. In a shop with a 33% gross margin, where 0.787 lei of every leu reported by Ads remain after VAT and shipping, the answer was not obvious, and the account was running at 3.21 across all 7 Performance Max campaigns: 42,570 lei in cost and 136,491 lei in value, over 90 days.

Before that, the site had foundations to fix: indexing and crawling on a catalogue with combined filters, canonicals for pagination, a menu that did not follow the way people search, and no content answering the technical questions asked before a purchase.

What we did

Phase 1, November 2024 to January 2025: the SEO foundation

Before content, the site had to be readable by Google. On the technical side: checking GA, GTM and Search Console, then indexability on every page type; canonical rules for pagination and for categories with 1, 2 or 3 filters; crawl budget optimised in robots.txt by removing useless filter combinations; 95 URLs prioritised for on-page audit and descriptive content.

On structure: keyword research across 14 product families, a new menu architecture on L1 and L2 levels, with filters derived from searches, new category opportunities, motion sensors for instance, and 12 blog articles: LED consumption, LED types, strip installation, sensor adjustment, lighting ideas.

Search (Romanian)	Monthly volume
bandă LED (LED strip)	14,800
proiectoare LED (LED floodlights)	9,900
bec LED (LED bulb)	6,600
bandă LED 220V (220V LED strip)	2,900
profil bandă LED (LED strip profile)	2,900
senzori de mișcare (motion sensors)	2,400
proiector LED exterior (outdoor LED floodlight)	1,900
iluminat exterior (outdoor lighting)	1,300

Phase 2, September 2026: the Google Ads audit

We calculated the break-even from real margin, tiered all 1,166 products by performance and implemented a single change, reversible in a minute. The details are in the sections below.

Phase 3: content

24LED customers ask technical questions before they buy. Content answers them directly and leads them to the product. The plan holds 100 blog articles: pillar guides, comparisons for the buying decision, use cases, B2B articles. Every article has an interactive element, a full tool, a micro tool or a value form, and is written to be cited by AI engines, not just found in Google.

In parallel, 100 “quotable” reels. A reel answers a single question, in the first 2 seconds: “Why does a new LED strip flicker?”. Each is produced once and published on Instagram, TikTok, YouTube Shorts and Facebook. Plus an LED strip calculator built for Gomag, which works out the right length and power supply.

The break-even

Of every leu reported in Ads, 0.787 lei remain after VAT and shipping. At a 33% margin, the threshold is 1 ÷ (0.787 × 0.33), roughly 3.85. Below 3.85, every leu spent on ads loses money.

ROAS in Ads	Profit per 100 lei spent
3.21 (starting point)	loss
3.85	0 lei
4	+4 lei
4.5	+17 lei
5	+30 lei

The conclusion: a ROAS of 4 barely covers costs. The real target for profit is 4.5 to 5.

Diagnosis: every product on a performance tier

We split the 1,166 products into tiers by cost, conversions and ROAS over the last 90 days.

Tier	Products	ROAS	Note
Hero	12	11.28	clearly profitable
Sidekick	69	9.17	clearly profitable
Borderline	10	3.53	around the threshold
Below threshold	99	3.46	under 3.85
Villain	37	1.35	15,377 lei in cost
No conversions	101	0	5,652 lei in cost

Only 81 products, hero and sidekick, are clearly profitable, and they were getting roughly 22% of the budget. Around 21,000 lei went into “villain” products and products with no conversion at all, at a ROAS close to 1.

Other findings: products over 300 lei have the best ROAS, 4.75; the 7 Performance Max campaigns overlap and the general campaign wins almost every time; two top products, an aluminium profile with 99 units sold and a staircase controller, were out of stock.

A single change, reversible in a minute. We did not touch the campaigns. We added a supplemental data source in Merchant Center that gives every product its tier label, in custom_label_2. All 1,166 products updated without errors, including the 8 atypical IDs in the Gomag feed. The working rule was set from the start: no change in Ads or Merchant Center without the client’s explicit approval, and every step has to be reversible. Excluding the 37 “villain” products was approved and is done through the same source, without editing the 7 campaigns.

What the accounts show

The scenarios for the last 90 days, calculated from the account data:

Scenario (90 days)	Cost	Value	ROAS
Starting point	42,570 lei	136,491 lei	3.21
Without the villain products	27,193 lei	115,732 lei	≈ 4.26
Without villains and without zero conversion products	21,541 lei	115,732 lei	≈ 5.37

These are estimates, not results. The scenarios are the theoretical maximum, because Performance Max will move budget toward other products. The realistic estimate is a ROAS of 4 to 4.5. Measured results are added here after the 2 to 3 week observation period.

The accounts, as the platforms report them, from the screenshots above:

• Google Ads, 1 January to 31 July 2026: 66.5K clicks, 142K lei in cost, 530K lei in conversion value, 3.72 lei of value per leu spent over seven months. The last 90 days, at 3.21, were below the period average.
• Meta Ads, 1 January to 31 August 2026: the catalogue campaign at 4.68 ROAS, with 58,027 clicks, 405 purchases and 185,824 lei in conversion value. Retargeting at 4.19 ROAS, with 111 purchases and 50,317 lei.
• Meta Ads, 2025: 468 purchases from the catalogue campaign, at 91.97 lei per purchase, from 43,043 lei. Dynamic retargeting brought another 120 purchases, at 92.32 lei each.

On Meta, the account is above the 3.85 threshold on the campaigns that matter. On Google, below. That is why the audit started with Google.

Next steps

• Excluding the villain products from Shopping ads. The products stay in Google’s free listings.
• 2 to 3 weeks of observation, with ROAS tracked weekly and labels recalculated periodically.
• Top products back in stock, because the algorithm loses exactly the products that sell.
• Restructuring by tier, only if ROAS does not rise enough: Heroes, Core and Test campaigns, with no overlaps.
• Local inventory and an enriched feed for the showroom: cost of goods, Google categories, better titles and images.

Services involved

• Paid Search, the break-even from real margin, per product performance tiers and automatic labelling in Merchant Center.
• Technical SEO, indexing, canonicals for filters and pagination, crawl budget.
• SEO & organic, keyword research across 14 product families, the menu architecture and the 100 article plan.
• Paid Social, the catalogue and retargeting campaigns on Meta, with ROAS above the break-even.