We Had Everything — Except Sales
Riya and her mother started Kaavya Jewels in 2022 from their home in Jaipur. Handmade silver jewelry — traditional Rajasthani designs with a contemporary twist. The pieces were genuinely beautiful. They built their Instagram presence organically, posting process videos, styling reels, and close-up shots. By 2025, they had 10,200 followers, a consistent engagement rate above 5%, and a well-organized website on Shopify.
They were spending ₹12,000 a month on Meta ads. Getting 800–1,000 website visits per month from ads alone, plus another 400 from organic Instagram traffic.
And converting at 0.3%.
That means roughly 4 orders per month from paid traffic. Average order value around ₹2,400. Monthly revenue from the website: ₹9,600. Monthly ad spend: ₹12,000.
They were running a business that was losing money on every sale it generated.
Monthly website visitors
Conversion rate
Lost in ad spend vs revenue/mo
What the Website Was Actually Doing
Riya assumed the problem was the ads. Wrong audience. Wrong creative. She kept tweaking campaigns, trying new audiences, testing different videos. The conversion rate didn't move.
A heatmap analysis told the real story. Visitors were landing on the homepage, scrolling for an average of 8 seconds, and leaving. The bounce rate was 81%.
The problems were structural, not creative:
The hero section had a beautiful banner image, but no text. No offer. No value proposition. Just a photo.
Product pages had no size guide, no materials breakdown, no care instructions. Just price and photos.
There was no trust section — no reviews, no social proof, no 'as seen in' mentions despite having press coverage.
The checkout was 4 steps long with mandatory account creation. On mobile, this was brutal.
WhatsApp DMs with product questions were getting replies 1–2 days later. By then, the customer had moved on.
The ads weren't the problem. The ads were working perfectly — they were sending interested, qualified people to a website that immediately destroyed their interest.
Fixing the Foundation First
The conversion website rebuild came first. Not a redesign — a restructure. The goal was simple: every element on the page should either build trust or push the visitor toward a decision.
The new homepage opened with three words above the hero image: "Handmade in Jaipur." Below that, a two-line value proposition, a CTA button, and — critically — four customer review snippets with photos directly on the hero section.
Product pages got a complete overhaul: materials listed explicitly, a size guide with a photo, a care instructions tab, and a "Customers also loved" section pulled from real WhatsApp reviews with the customer's first name and city.
The WhatsApp Problem Was Separate — And Worse
While the website rebuild was underway, they looked at the WhatsApp situation. Riya was handling all DMs herself, mostly in the evenings after her day job. During peak times — sale announcements, reel virality moments — the inbox would fill with 50–80 messages overnight.
Common questions: "Is this available in gold?" "What's the weight?" "Can you customize it?" "Do you ship to Canada?" All of these had known answers. But each required Riya to personally type a reply.
The silent killer: people who DM'd with purchase intent and didn't get a reply within 2 hours would often place the order elsewhere. They weren't angry — they just moved on.
The WhatsApp automation setup handled the top 12 FAQ categories automatically. Custom orders — which needed Riya's personal input — were routed with a clear tag. Everything else got an instant reply with the relevant info and a link to the product page.
Where the Numbers Landed
Two months after both systems went live, Riya ran the numbers.
Website conversion rate
WhatsApp first reply time
Monthly website orders
Monthly net (ads vs revenue)
The ads budget stayed at ₹12,000. The traffic stayed roughly the same. What changed was what happened to visitors after they arrived — and what happened to DMs after they came in.
Riya didn't shut down the store. She hired a part-time packaging assistant instead.