How we shipped Kailash Holiday in 48 hours.
A working website, multilingual, photo-led, deployed to a global CDN, in two days. No magic — just a tight scope, a refusal to over-engineer, and one all-nighter.
The brief came in on a Sunday morning. The pilgrimage season started Wednesday. The founder had photos, a rough sense of what should be on the site, and no website. We agreed to ship by Tuesday night.
This is what 48 hours of focused work looked like.
Hour 0–4: scope murder
The most important thing on day one was figuring out what to cut. We started with a list of 11 pages: home, packages, helicopter, overland, Mansarovar, dates, FAQ, gallery, blog, about, contact.
After a 30-minute call we ended up with 1 page.
Everything that mattered — the trip pitch, key dates, contact — could fit on one long scrolling page. Anything that didn't fit there probably didn't matter.
Hour 4–14: design + first build
Started in Figma for 90 minutes. Settled on:
- One hero with a Mansarovar wide shot, big serif headline, English/Nepali/Hindi gate
- Three-section content stack: pitch, packages, dates
- Sticky WhatsApp button as primary CTA
- Footer with phone, email, social
By hour 14 we had a working Next.js build with placeholder content, language gate, and a deployment pipeline on Vercel. No CMS — content lived in MDX files committed to git.
Hour 14–24: real content
Founder sent us:
- 18 photos
- 3 paragraphs of pitch copy
- A pricing sheet for 4 packages
- 6 trip dates
We dropped them in. Cropped photos to consistent aspect ratios. Wrote alt text. Translated key labels into Nepali and Hindi (Google Translate + a friend who speaks both for proofreading).
Hour 24–36: polish
The middle stretch is where most builds die. We forced ourselves to ship before polishing further.
Cut from the to-do list:
- Custom 404 page
- Schema.org structured data
- Newsletter signup
- Animations beyond fade-in
Added before launch:
- Open Graph image (one wide Mansarovar shot)
- Favicon
- Meta description in 3 languages
- WhatsApp deep-link with prefilled "Hi, I'd like to know more about your Kailash package."
Hour 36–46: testing on real devices
Borrowed three phones from people in the office. Tested:
- iPhone 12 (3G simulated)
- Samsung A52 (real Nepal Telecom 4G)
- An older Android with Chrome 80
The iPhone was fine. The A52 was fine. The old Android revealed three CSS bugs and one font that wouldn't load on slow connections. Fixed all four. Shipped.
Hour 46–48: live
Pointed the domain. Tested the WhatsApp link. Sent the founder the URL. He sent us a thumbs up. We slept.
What two-day builds teach you
You don't need 6 weeks for most websites. You need:
- A scope cut in half from what you think you need
- Photo and copy ready before you start
- The discipline to ship before polishing
- One person who can decide things without committee
Most "6-week" projects are 2 weeks of decisions stretched over 6 weeks of waiting for them. We just stopped waiting.