Office Paglu
D2C Brand · E-commerce
Office Paglu is a print-on-demand merch brand for Indian corporate life,. I built this brand with my girlfriend. Office Paglu runs on a custom Next.js storefront, an Express and Supabase backend that routes orders to print vendors, WhatsApp order tracking and cart recovery, and an AI pipeline for product and ad imagery.
For working unprofessionals
Office Paglu is a clothing brand for Indian corporate folks. It's made for people in their 20s and 30s who work in engineering, design and product. Every tee is designed in-house to be funny and relatable, the kind of thing you wear to standup and get a laugh.
My girlfriend and I started it in January 2026 as a side project, with one rule: we'd build everything ourselves and see where it went. The brand, the storefront, the backend, the admin panel and the content pipeline all came together between January and March. The store runs on print-on-demand, so there's no inventory and no warehouse. That kept it cheap to start, and it meant most of our hard problems turned out to be software problems.
How we built it
We skipped the off-the-shelf store builders and wrote our own backend. It took longer up front, but it meant we could change anything the moment the business needed it, which turned out to be every other week.
Supabase holds everything: products, carts, orders, coupons and customers. When someone checks out, the order, its items and any coupon are saved together in one go, so we never end up with half an order. Payments go through Razorpay, and every paid order is sent automatically to our print partner, who prints, packs and ships it.
In India, people live on WhatsApp, so we made it our front desk. Using the WhatsApp Business API, customers get a message when their order is placed, confirmed, shipped, out for delivery and delivered. There's no app to install and no email to dig through.
For our side of the conversation we use Periskope. If a delivery fails because the customer wasn't home or the address was incomplete, our team group gets an alert, and we can call the customer before the parcel heads back. Customer replies land in an inbox inside our admin panel, so support happens in the same place as everything else.
Building a POD store taught me that commerce is mostly logistics. The storefront took weeks. Getting a correct answer to "where is this order, and is it costing me money" took months.
Problems solved
Once real orders started coming in, the data told us what to fix next. Each of these was a small change, shipped as soon as we spotted the problem.
People visited, but didn't buy
Plenty of people landed on the site, looked at a few tees and left without buying. Checkout was simply too easy to miss.
The fix: we made the cart and checkout much more visible and quicker to reach from anywhere on the site. Once someone is in checkout, it's now harder to wander off: fewer exits, fewer distractions, one clear next step.
Traffic grew, Instagram didn't
More people were reaching the store, but our Instagram still had almost no followers. That's where our next customers come from, so it mattered.
The fix: a secret coupon. Shoppers get ₹50 off by DMing us on Instagram and following the page. It turned one-time visitors into followers we could reach again, and our follower count jumped.
Small orders
Most people bought a single tee, so our average order value stayed low.
The fix: bulk discounts. Buy two tees and you get a percentage off the order. The cart lets you know when you're one tee away from the deal, which makes adding a second one an easy call.
Cash on delivery was hurting us
COD sat front and centre at checkout, and COD buyers returned their orders far more often. Every return costs shipping both ways, so a lot of those orders lost money.
The fix: prepaid became the default option, with free delivery. COD moved into a drawer you have to open, and it now carries a ₹50 delivery charge. Anyone who really needs COD can still use it, but prepaid is now the easy choice.
Related tees were hard to find
Shoppers usually saw only the tee they came for. The rest of the collection was a few taps away, and most people never got there.
The fix: a frequently-bought-together section on every product page. It shows tees that pair well with the one you're looking at, so adding a second is a single tap. It also nudges people toward the bulk discount.
Design
How it sounds
Before we designed a single screen, we wrote down how the brand talks. The reference is Jim Halpert looking at the camera: someone who knows exactly how absurd office life is and plays it straight. The Paglu isn't miserable about work, just amused by it.
The language is light Hinglish. It has an English backbone with Hindi words that need no translation, like kaam, naukri, chai-sutta and yaar. Product descriptions follow three beats: a short deadpan punchline, a relatable scene, then an honest product detail, usually signed off with "For working unprofessionals."
The guide also lists what we never say. Nothing LinkedIn-motivational, nothing that punches down at interns or support staff, no profanity (the tees have to be wearable to office), and none of the usual D2C filler like "curated", "quirky" or "premium". Corporate jargon like "synergy" and "circle back" only ever shows up as a joke.
We built a set of Claude skills that write our captions and product copy, and each one reads the guide before it writes anything.
The brand language guide is the most reused file in the project. The storefront copy, product pages, ads and AI-generated captions all go through it, which is why they sound like one person wrote them.
How it looks
The storefront uses a neobrutalist style: thick borders, heavy type, flat blocks of colour and grainy gradient backgrounds. It's loud on purpose, so it looks nothing like the grey corporate tools the jokes are about.
Most of our traffic comes from Instagram on phones, so we designed for mobile first. Picking a size opens a bottom drawer instead of a tiny dropdown, buttons are big and easy to tap, and small animations make the store feel alive.
Every tee gets the same set of product photos: a ghost mannequin shot, a flatlay, a close-up of the print, and full-body shots front and back. We generate them with AI from one base mockup, so a new drop gets a full set of photos without a photoshoot.
Future plans
These are our immediate milestones:
- Grow to 1,000 customers a month.
- Automate most of the day-to-day running of the business.
- Grow our Instagram to 10k followers.
