How a D2C home décor brand went from "I don't know why it's not working" to a 67% revenue jump — after switching agencies

He had asked the question more than once. Why aren't these ads working? The agency had no clear answer. Not because the answer was hard to find — but because they'd never looked. The cause was diagnosable. The data was there. And the plateau that felt like a ceiling was actually just a problem nobody had bothered to investigate.

BRAND SNAPSHOT

• Industry: Home Décor (D2C)

• Category: Pop-culture themed rugs — machine tufted, custom/handmade

• Geography: India (pan-India, southern states as highest-AOV market)

• Stage: ₹6,00,000/month → ₹10,00,000/month in 2 months

• Services: Meta Ads, Creative Strategy, Conversion API Integration, Audience Architecture

THE PROBLEM

The brand had built ₹6L/month and a 3.25x ROAS without agency help. Decent for a self-built operation. But growth had plateaued — and after bringing on a marketing partner, the founder was watching the same numbers month after month with no explanation. "There is lack of consistency," they wrote in their onboarding form. "At times they are not answerable on why ads are not working." No benchmarks. No defined BAU. No accountability when results fell short. Just silence.

WHY IT WAS HAPPENING

Three separate failures were compounding — and none of them were hidden. The Meta algorithm was flying blind: the payment gateway the brand used had a Conversion API that wasn't firing purchase events back to Meta. Event coverage score: 0%. The algorithm was optimizing on add-to-carts and checkouts, not actual buyers. Every rupee spent was training the machine to find browsers, not purchasers. On top of this, there was no campaign architecture separating awareness from conversion, and ad budgets were added sporadically — constantly resetting algorithmic learning phases. The problem wasn't inexplicable. Nobody had diagnosed it.

THE SOLUTION

The first step was an audit of data infrastructure — not creative performance.

Vault (Brand Value Engine): Within the first week, the team identified that the payment gateway wasn't correctly routing purchase events to Meta's Conversion API. Event coverage: 0%. This was treated as a critical infrastructure failure, not a background issue. Working directly with the gateway's technical team, the fix was implemented via Stape.io — a server-side tagging solution that correctly captured and deduplicated purchase events. Event coverage moved from 0% to 100%. For the first time, the algorithm had real purchase signal to optimize against.

Sentinel (Scientific Media Buying): Campaign architecture was rebuilt from zero. Two parallel campaign types were launched: a Catalog campaign for dynamic retargeting and a bottom-of-funnel conversion campaign built on historically top-performing creatives. Geographic targeting was surgically adjusted — specific manufacturing hub cities that had been generating high click volumes from trade buyers with zero purchase intent were excluded entirely. Southern Indian states, identified as the highest AOV market, were prioritized. Daily budget was set with a clear protocol: ROAS had to hit defined thresholds before spend was scaled. Daily monitoring replaced the sporadic oversight of the previous arrangement.

Mythos (Creative Advantage): Creative strategy leaned into what made the product genuinely extraordinary — its cultural specificity. Product-specific creatives were built for the top-converting SKUs. Advantage+ Video emerged early as the standout format — showing the texture and "reveal" of the product outperformed static and catalog ads across every tracked period.

THE RESULTS

₹6L → ₹10L/month. Two months. 67% revenue increase.

First tracked week after CAPI was fixed: ₹39,943 in ad spend generated ₹1,81,516 in sales — a 4.5x ROAS. Best four-day stretch: 5.06x. Advantage+ Video produced a 5.7x ROAS in its best weekly period. Single-day peak: ₹70,000+ in gross sales at 10x+ ROAS.

The founder who came in describing themselves as "Unsatisfied" — managing the entire brand alone — now had a system running with professional oversight, daily tracking, and no unanswered questions.

LESSONS FOR SIMILAR BRANDS

• When an agency can't explain why your ads aren't working, that's not a communication problem — it's a competency problem. A plateau has a cause. If your partner can't find it, they're either not looking or don't know what to look for.

• Before blaming creative, check the data layer. The most common invisible ceiling for Indian D2C brands on non-Shopify gateways is a broken Conversion API. Zero purchase event coverage means the algorithm is training on the wrong behavior — and no amount of creative testing fixes that.

• Switching agencies without a diagnosis is a gamble. Switching with one is a strategy. The previous engagement failed not because ads can't work for this category — but because three specific, fixable problems were never identified.

CHALLENGES WE FACED

• The CAPI fix took three weeks. Resolving the payment gateway configuration required coordination with the gateway's technical team, test orders to verify the fix, and monitoring to confirm deduplication was working. This compressed the early optimization window — the algorithm couldn't fully train on purchase data until coverage hit 100%.

• Budget continuity was a recurring disruption. As a solo founder managing operations and marketing simultaneously, the brand periodically ran low on ad account funds mid-flight — causing campaign pauses that reset algorithmic learning. A 10-day replenishment protocol was established to prevent this from compounding.

BELIEFS CHANGED

• "My previous partner couldn't explain it, so maybe it just can't be explained." The explanation was precise. Conversion API at 0% event coverage. Manufacturing-hub traffic polluting the audience pool. No funnel architecture separating awareness from conversion. These aren't mysteries. They're infrastructure failures with known fixes — and diagnosing them is the job.

• "A niche product has a small ceiling." A single-day ROAS of 10x+ happened within the first two months. The ceiling wasn't the product category. It was a broken data layer that had been invisible to everyone responsible for fixing it.