How This Indian Fashion Brand Stopped Fake Orders Draining Its Ad Budget — By Excluding One City

A growing Indian womenswear brand was receiving sequential fake orders from a single region — contaminating Meta pixel data and wasting fulfillment effort. Arlox identified the geographic fraud pattern and applied surgical targeting exclusions. The fake orders stopped within 48 hours. The brand went on to scale to ₹24L/month.

In late January, the founder messaged the Arlox team with something unusual. "I'm only receiving fake ones — all from Delhi." It wasn't a delivery complaint. It wasn't a ROAS issue. It was something nobody had warned her about: her brand was being systematically targeted with sequential fake orders, almost certainly from a competitor operation or bot network. Within 48 hours of identifying the pattern, the team had excluded the entire region from Meta targeting. The fake orders stopped.

BRAND SNAPSHOT

• Industry: D2C Fashion

• Category: Women's clothing (dresses, seasonal collections)

• Geography: India

• Stage: ₹1L/month → ₹24L/month over 7 months

• Services: Meta Ads (Scientific Media Buying), Creative Strategy, WhatsApp CRM

THE PROBLEM

In late January, this Indian womenswear brand began receiving a suspicious pattern of orders — consecutive, from the same geographic cluster (Delhi, and adjacent areas of UP and Haryana), and consistently non-genuine. The founder flagged it immediately: "Received one fake order." Then another. Then: "I'm only receiving fake ones." The orders looked real from the outside — they triggered ad attribution, consumed inventory processing time, and fed purchase events into the Meta pixel. But they never converted into actual revenue. In a month where the brand was scaling carefully, every fake purchase signal was teaching Meta exactly the wrong thing about who buys from this brand.

WHY IT WAS HAPPENING

The D2C fashion space in India has a documented dark side: competitor-driven fake order attacks. A competitor — or a contracted bot operation — places enough real-looking orders to contaminate a brand's conversion data, inflate apparent return rates, and waste fulfilment bandwidth. The specific mechanism here: Meta's algorithm was serving ads to Delhi-region users who were systematically placing and rejecting orders. Because the pixel saw "purchase events," it flagged these users as converters — and started serving more ads to audiences that matched their profile. The result: worse real conversions, worse ROAS, and a growing pile of fulfilment-side noise.

The brand had no operational filter for this. Orders came in, were processed, and eventually failed — but the damage to the pixel happened the moment the purchase event fired, before the order outcome was known.

THE SOLUTION

The Arlox team diagnosed the pattern within hours of the founder's first message. The fix was surgical.

Sentinel (Scientific Media Buying): The team cross-referenced the fake order cluster against ad delivery geography. The data was unambiguous — Delhi, Uttar Pradesh, and Haryana were the source. Geographic exclusions were applied to all active Meta campaigns immediately, blocking the entire region from ad delivery. Fake purchase signals stopped entering the pixel. The team then ran a structured audit of the prior two weeks to assess how much the contaminated data had skewed audience targeting, and rebuilt affected ad sets from a clean baseline. Performance benchmarks were reset to reflect what real buyers — not bots — looked like.

Vault (Brand Value Engine): Beyond the targeting exclusion, the team added a secondary layer: an order monitoring protocol. Any future concentration of orders from a single pin code cluster within a short time window would trigger manual review before processing. The founder received a simple diagnostic framework: if the same region produces 3+ orders within 24 hours with no corresponding organic brand activity, treat it as a fraud signal and escalate immediately rather than processing normally.

THE RESULTS

The fake order flow stopped immediately after geographic exclusions were applied. Clean purchase signals resumed — allowing the Meta pixel to optimize against real buyers for the first time in weeks. The downstream impact compounded: cost-per-purchase became more predictable, conversion rates stabilised, and the algorithm stopped serving ads to the contaminated regional audience profile it had been building. Over the following 7 months, this brand — starting from ₹1L/month and dealing with a corrupted funnel — reached ₹24L/month. Clean data was the foundation that made scaling possible.

LESSONS FOR SIMILAR BRANDS

• Fake orders are an ad strategy problem before they're an operations problem. Every fake purchase event trains your Meta pixel on the wrong audience profile. If you're receiving high-volume orders from a region with zero brand awareness, investigate before those events corrupt your algorithm.

• Geographic exclusion is a precision tool, not a blunt instrument. Excluding a city or state doesn't mean abandoning that market. It means protecting your pixel data while you identify and resolve the source of contamination — then re-entering the market cleanly.

• Track order geography against your organic footprint. If your brand has no Instagram engagement, no search traffic, and no word-of-mouth from a region — and you suddenly receive five orders in two days from there — something is wrong. Investigate, don't process.

CHALLENGES WE FACED

• Attribution lag delayed initial diagnosis. Because Meta's pixel attributes purchases with a 7-day click window, it took careful day-by-day order tracing to confirm the pattern was systematic fraud rather than a genuine regional demand spike. The diagnosis required correlating order geography with ad delivery data — not just looking at surface-level order reports.

• Geographic exclusion temporarily reduced top-of-funnel reach. Delhi is a major fashion market in India. Excluding it while the team investigated the fraud source reduced overall ad reach and required rebuilding performance benchmarks based on the cleaner remaining geography. The short-term reach loss was the necessary cost of protecting long-term pixel quality.

BELIEFS CHANGED

• "Fake orders are a delivery team problem." Before this engagement, the assumption was that suspicious or fake orders were handled at the logistics level — identify, cancel, blacklist. This case proved they're a media buying problem first. The pixel damage happens before the fulfilment team even opens the order.

• "More purchase events is always better signal." The brand initially treated order volume as a pure positive. After this, they understood that order quality matters as much as order quantity — and that fake purchase events actively degrade the intelligence driving your paid ads. Volume without quality is noise, not signal.