PortfolioCase study
VL-01Live · mock data
Nuptis → Velora
Where brands and makers find their fit.
Nine days.Two products.One survived.
A B2B apparel sourcing marketplace where brands and manufacturers match, then turn matches into bids.
2
products
Structural
Nuptis, then Velora — nine days, solo
Day 7
first product killed
Structural
1 day
to build Velora
Structuralas of 11 Aug 2026
Evidence keyMeasuredStructural



The problem
Supplier trust is found by asking around — and it doesn’t travel.
An indie apparel brand needs a garment maker; a factory in Tiruppur or Ludhiana needs brands.
- Cold referrals
- Trade fairs
- Directories
- Unverified trust
- Trust that doesn’t travel
Research (team)
The delay is waiting, not work.
Procurement interviews were pooled across the Case Study 3 team.
15–30 business days
Active work · under 10%Waiting between hand-offs
Team secondary research, marked “verify before external use” — shown as background, never as Velora’s data.
“I find out where a vendor is by asking around.”
Procurement interview, team PRD “We scrutinise new vendors. Changes to old ones, we just… trust.”
Procurement interview, team PRD
The pivot
Kill the first bet on day seven.
Days 1–7

Nuptis
Vendor ops for wedding-planning agencies — live, with every success metric defined and none measured.
Day 7 · killed
The evidence
“Weddings were blue — but a shallow pool.” Few events, low willingness to pay.
Keep the insight
Keep the trust-and-onboarding insight; aim it at apparel sourcing.
By day 9

Velora
Brands and manufacturers match; matches turn into bids. Built in a day.
The bet
Portable trust.
One trust profile a manufacturer carries into every match, RFP and bid. The scores are authored prototype data — verification was out of scope.
- Match
- RFP
- Bid



Key learnings
Nine days. Two products. One survived.
Kill without flinching
A blue ocean that is a shallow pool is still the wrong ocean.
Attack the hand-offs
Onboarding delay is queue-time between teams, not any one team working slowly.
Label honestly
Mock data isn’t a proven path, and an authored score isn’t a verified one.
Evidence
A working prototype — not traction.
10/10
unit tests passing
Measuredas of 15 Sep 2026
0
horizontal overflow
Measuredas of 11 Aug 2026
every route, at 375 and 768 px
1 day
build time
Structuralas of 11 Aug 2026
all 40+ build commits on 11 Aug 2026
Not measured
- Real users
- A real pilot
- Usage data
- A verified trust score
Mock data
Source material
Everything here is backed by real artifacts.
- PRD
- Research
- Post
- Readme
- Architecture
- Evaluation
- Live data