Why I rebuilt Chiblu after a Mumbai pop-up

v1 was a directory. Free pre-launch had 50+ makers signed up. Then I booked a pop-up in Mumbai - two days, in person. The product I thought I was building wasn't the one makers needed. Four problems they showed me reshaped v2. We went free, storefront-first, and SEO-indexed.
a directory v1's flat list-of-makers two days, in person a Mumbai pop-up, Dec 2025 four problems DMs, screenshots, trust, the crowd free, storefront-first the v2 we built next

I built v1 of Chiblu like most marketplaces start. A directory of makers. Their bios. Their photos. A way to browse, contact, follow. Free during pre-launch. Paid subscription planned for after.

By month two, 50+ makers had signed up. The dashboard looked right. Numbers were trending up. I told myself the product worked.

Chiblu v1 - the directory model, screen 1
Chiblu v1 - the directory model, screen 2
v1 - the directory. Makers had pages. Buyers had a feed. Everything looked right.

Then in December 2025, I booked a stall at a pop-up in Mumbai.

What I went looking for

I'd done the right things on paper. 600+ user interviews. Forms. Calls. A waitlist. I knew what makers said they wanted.

But I'd never sat next to one while she tried to sell something. So I paid for a small folding table, packed two days of patience, and showed up.

The plan was to demo Chiblu to passing makers and watch what happened.

What actually happened was different.

The pop-up stall in Mumbai, December 2025
The pop-up stall. Mumbai, December 2025. Two days. A folding table, a few prints, and a lot of listening.

The four things makers showed me

The maker in the next stall started talking to me. Then the maker across from me. Then a textile artist walked over because she'd heard I was building something for people like her.

Over two days, the same four problems came up. Not in the language they'd typed in my forms. In the language they used when they were tired.

1. DM overload. Every maker I talked to was running their whole business through Instagram DMs. "I get 30 messages a day. By the time I reply, half of them have bought from someone else."

2. Screenshot chaos. Customers ask for screenshots of every variation - this color, that angle, that fabric, that print. Makers were sending hundreds of photos a week to people who never bought.

3. No trust on DMs. Buyers wanted to pay. But sending UPI to a stranger based on an Instagram message felt sketchy. Makers were losing sales at the exact moment they should have closed.

4. Lost in the crowd. The ones who'd tried big marketplaces - Amazon, Etsy, Flipkart - all had the same complaint. "We're listed. Nobody finds us. We're three pages deep behind someone selling mass-produced versions of the thing we hand-make."

I'd missed all four. The directory model wasn't a directory problem. It was a trust problem, a workflow problem, and a discovery problem dressed up as a layout problem.

A second view of the Mumbai pop-up stall
Day two. Different maker, same four problems.
The form said "I need a portfolio." The pop-up said "I'm drowning in DMs and nobody pays what hand-weaving is worth."

What I changed in v2

I came back to Bangalore and rebuilt for the next four months. Four big calls came directly from the pop-up:

Before · v1
After · v2
Pricing
Paid subscription for sellers
Free. No monthly fee, no per-product fee.
Product surface
A directory listing per maker
A custom storefront per maker - real product pages, variants, share-able URLs.
Discovery
Internal feed only
SEO-indexed from day one. Google finds the maker, not the marketplace.
Trust
Same as Instagram DMs
Marketplace as the trust layer. Buyer protection, verified sellers, UPI through Chiblu.

The biggest call was the first one. Dropping the paid subscription. v1 was going to monetize through monthly fees. After the pop-up, that was obviously wrong. Makers can't pay a monthly fee while they're losing sales to broken trust. Free for sellers became non-negotiable.

The second call: every maker gets a custom storefront, not a listing. Real product pages. Real variants. Real share-able URLs. The opposite of a screenshot-by-DM workflow.

The third: SEO indexing from the first day. Every product page is its own URL. When a customer searches for "hand-block-printed cotton kurta Jaipur", they find the maker - not page three of Amazon.

The fourth: marketplace trust signals. Verified seller flags. Buyer protection. UPI through Chiblu, not stranger-to-stranger. The platform becomes the intermediary that Instagram DMs can't be.

Chiblu v2 - the current product, with custom storefronts per maker
Chiblu, today. Each maker gets a full storefront. SEO-indexed. Free. Trust built into the workflow.

Four months later

v2 went live in February. Steady month-on-month growth since:

47 Sellers today
(from 13 in Feb)
71 Live products
(from 20 in Feb)
~12 New sellers
each month, avg

Steady, not exponential. Which is also not the point of this post.

The point is: the four problems makers showed me at a Mumbai pop-up are still the four problems v2 is solving. Every new seller signs up because of one of those four things, not despite them.

The forms missed it. The pop-up didn't.

What I learned about research

You can't fix what you can't see. Forms abstract away the part that matters - the moment a maker realizes she's lost track of 12 DMs while replying to one. Surveys average out the urgency.

That's why v2 looks the way it does. Not because of the form responses. Because of the pop-up.

Your product is what you build after you watch someone try to use the first version.

Chiblu is live on the App Store. Building something? Tell me what you'd build.

Let's connect

Building something interesting? Want to collaborate? Always happy to chat about design, products, and the messy middle of building.