How SEVEN scaled from ₹32,588 a month to ₹35L of total revenue in 66 Days

At-a-Glance

The store was open and almost nobody arrived

In the 30 days to 19 June 2026, SEVEN took six orders. Some days had one order. Most had none.

Surbhi Gupta was not selling a cheap product into a crowded feed. The store's average order across its whole history is ₹5,547.95, and every single order is prepaid, paid by card, UPI, net banking or wallet through Razorpay. That is unusual in Indian apparel, where cash on delivery (COD) is the default and returns of undelivered parcels (RTO) eat the margin.

So the problem was not trust and it was not payments. It was that a print-heavy set at ₹6,000 has to be explained to somebody before they will buy it, and nothing was doing the explaining.

A cold account has no idea who your buyer is

We launched on 20 June. That day cost ₹2,314 and produced no order, and so did 23, 25 and 27 June.

A cold start has nothing to learn from. The pixel, the piece of code that reports purchases back to Meta, is what trains delivery. With no purchase signal in the account, Meta cannot tell a browser from a buyer, so the learning phase, the first days after a change when delivery is still guessing, runs long and expensive. Four days with spend and no orders is what that looks like from the outside.

A print is a considered purchase, not an impulse. On 25 June the statics for a two-looks-in-one-skirt concept drew interested traffic that reached checkout and stopped. Attention was never the bottleneck. On 9 July an isolated Pink Guava reel held a 22.85% hook rate, the share of viewers who stay past the opening, and produced clicks at ₹62.92 with no add-to-carts at all. It was paused.

One account cannot optimise for two jobs. Cold prospecting builds intent; the Advantage+ catalog campaign, Meta's automated engine that serves products from the store feed, harvests it. Run together with no separation, the automated side takes credit for demand the manual side paid to create, and budget drifts to the wrong place.

The fix wasn't a new audience. It was a reason to buy.

Mythos, the creative advantage, was about the claim, not the picture. Angle testing means testing the idea an ad makes rather than a new photograph of the same idea. The reversible set became "two looks in one outfit". The dresses became a vacation edit. Then, on 13 July, a reel of Surbhi explaining the clothes herself returned 23.5 times its spend with a 7.12% click-through rate. High-trust storytelling closed evening traffic that statics only warmed. Customer-made videos (UGC) followed the same pattern.

Sentinel, the scientific media buying, ran one rule. An ad earned more budget in one way only: by bringing orders at a cost SEVEN could afford, a target of about ₹900 of ad cost per order. Winners moved up 10% to 20% at a time so delivery kept learning. Losers were named and stopped. On 7 August a best-sellers reel was cut after spending ₹2,834.72 without a sale.

Vault, the brand value engine, protected the price. The brand stayed prepaid throughout, so revenue booked was revenue collected. Discounts were used once as a dated offer rather than as a permanent floor under the price.

Your Shopify export already knows what to advertise

If you sell clothes, the export sitting in your Shopify admin has already run half the test you are about to pay for. Look at which products repeat across sizes in your top ten, because that is proven demand waiting for distribution rather than a hypothesis.

For SEVEN, the top ten products are 25.17% of product revenue, so the catalogue is broad but the money is not evenly spread. The ads followed the concentration.

Geography said the same thing. Of 667 lifetime orders, 289 came from Maharashtra and 255 from Mumbai alone.

The result

The buying period produced 630 orders and ₹35,00,535 on ₹5,13,245 of spend, 6.82 times what went out. Ad cost per order settled at ₹815 against an average order of ₹5,556.40.

The founder's own comparison is the month. In the 30 days before any spend the store took ₹32,588 from six orders. In the last 30 days of buying, 26 July to 24 August, it took ₹18,87,201 from 328 orders, and the average order rose to ₹5,754.

Meta's figures answer a narrower question. Over its own 90-day window the platform attributed 613 conversions and ₹33,12,691.55 to itself, a 4.59 return. Neither figure is profit.

The single best day was 3 August: 22 orders and ₹1,66,152 on ₹7,484 of spend.

Growth has a price, and the account showed it

More budget does not buy more of the same buyer. It buys the next one out, who is less interested. Split down the middle, the engagement reads like this.

Both halves grew the business. The second one just cost more per order to get there, because by August the cheap attention had already been bought and new creative was being launched into higher auction prices. The growth was real, and it came with a bigger bill.

"Will this actually look good on me?"

On 6 July, SEVEN installed Style Me, Arlox's virtual try-on app for Shopify. It lets a shopper see a look generated on herself before she buys. Traffic reached it on 11 July and never stopped: every one of the 47 days recorded activity, 71% of opens became a finished look, and 98% of attempts succeeded.

The useful finding is where the demand sat. Eight products absorbed 357 of the 513 looks, and the electric sangria set alone took 89. Every one of the eight is a print or a coordinated set, which is exactly the item a flat product image describes worst. Sunday carried 13 of the 36 add-to-carts, one seventh of the calendar holding more than a third of them.

A try-on isn't a sale, and an add-to-cart isn't an order. The tool's own revenue attribution reads zero, so nothing here connects it to the store's total. Wiring cart adds through to confirmed orders is what would settle it.

What this means for a brand like yours

"A creative with great retention must be working." The Pink Guava reel held 22.85% of viewers past its opening and produced clicks at ₹62.92 with no add-to-carts. Hook rate tells you whether an opening holds attention. It can't tell you whether an ad sells.

"Polished product content will do the job." The single strongest asset in the account was the founder talking to camera, at 23.5 times its spend on 13 July, and customer-made videos were next. People buying a ₹6,000 print want somebody to explain it.

"A discount is just a lever you pull." SEVEN ran a 15% offer into mid-August. When it was switched off on 17 August, conversion on full-price pages collapsed and the day produced one order against ₹9,090 of spend. The offer had stopped being a nudge and become the reason.

What it cost, and what is still open

The failures are named in the logbook and they were not small. A tropical-print reel took ₹3,376 with no transaction and was paused on 6 July. A best-sellers reel took ₹2,834.72 and was paused on 7 August. On 12 August, newly launched assets absorbed 43.7% of the day's spend in their first 24 hours and returned nothing, which is what the learning phase costs every time you test.

Two days broke badly. On 29 July a checkout failure reversed ₹6,998 of sales across every ad. On 17 August the day after the offer ended took ₹999.

The customer-made video route also stopped for a reason that had nothing to do with performance: on 20 August the brand raised a concern about the influencer involved, and those ads came down. A good ad also has to be one the brand is happy to run.

What is unresolved is the dependence on one asset. By 24 August the founder reel had fatigued to a 1.98x cumulative return and its budget was cut 30%. The next winner has to be found the same way this one was.

Questions

About Arlox

Arlox is a performance marketing agency for D2C fashion brands, working on profitable ad scaling, RTO reduction, COD-to-prepaid conversion, and getting brands CM2-positive at scale. This is one of 450+ brands we've worked with, and hundreds of those founders have recorded on-camera interviews about their results. Founded by Varinder Singh Gakhar (Vann Laniakea).