When Sourcing From India for Ecommerce, Stay Away From These Blunders

Your first container from India will teach you more than any course, and most of the lessons cost money. I've watched sellers on Amazon and Shopify make the same eight mistakes for years, and I've made a few of them myself. This post lists them so you don't have to pay the tuition.

One thing straight up - every number here is Indian and dated, because that's where I work and where our research is deep. Most of the patterns travel to the rest of Asia too, but I'm not going to invent a Chinese or Vietnamese figure to prove it.

1. Trusting The Directory Price And The Directory Badge

The listed price looks great, the badge says "verified", and you build your whole margin on it. Then you enquire and the real quote is nowhere near the listing.

In India that's the norm, not the exception. Buyers on forums (2023) describe IndiaMART and TradeIndia prices that are months old or set low to draw an enquiry. IndiaMART's TrustSEAL is a paid seller package (check the current price with IndiaMART), and it checks that a business exists, not what it makes. Its buyer payment protection doesn't cover foreign buyers, and "Pay with IndiaMART" was discontinued (IndiaMART FAQ). The sellers who reply fastest are on the paid tiers that see your enquiry first, so speed shows subscription level, not seriousness.

The pattern travels. Whichever country and whichever platform, ask two questions before you trust a listing: is this price current for my quantity and Incoterm, and does the platform's protection cover a buyer in my country? On Alibaba, India-based sellers reportedly can't use Trade Assurance and ask for PayPal or a wire instead (buyers on forums, 2020 to 2023, not re-checked since, so verify it before you rely on it).

2. Assuming The Person You're Talking To Runs The Machines

Most first contacts you make in India are traders or agents, and that's fine. Traders are a normal business model, and large retailers buy through them every day. The blunder is not knowing, because then you don't know who answers for quality when a batch goes wrong.

India makes this easy to check for free, which is one reason I like sourcing here. India's tax number and its export code share one identity key (the PAN), so a single number lets you check that every record belongs to one legal person. A composition-scheme GST number can't export at all under Indian law, so if one quotes you FOB then someone else's registration is doing the shipping (CGST Act s.10(2)(c) and IGST Act s.7(5)(a)). A manufacturer-exporter RCMC is the strongest public "we make it" signal India offers, because the export council asks for evidence before issuing it (Handbook of Procedures 2023, para 2.79(a)). And the cheapest check of all is that the legal name on the GST record, the proforma invoice and the bank account match (buyers on forums, four voices, 2023 to 2025).

None of that proves quality. It proves who you're dealing with, and that's the question. The full walkthrough is in how to check an Indian supplier. If you'd rather someone ran the checks, including the deeper tax view that needs an Indian login, SourcingSync (opens in a new tab) does that.

3. Skipping The Pre-Production Sample To Hit A Launch Date

Your launch is in eight weeks, the sample looked great, so you skip the pre-production sample and go straight to bulk. The costliest disputes I've read about follow exactly that decision. In one case on a forum in 2025, a Canadian buyer skipped the T-shirt sample to save time, then withheld the whole balance when the bulk came in wrong, including payment for the styles that were fine.

The sample you approved was probably a proto, made in whatever fabric was on the shelf. Only a pre-production sample in the actual bulk fabric and trims proves the order (buyers on forums, 2013 to 2026). Two India-specific tricks to know - on heavy knits, some units reach the quoted GSM with a starch finish that washes out, so ask for a weight test on unwashed fabric and again after one wash (production-side voices on forums, 2026). And small heavy-knit orders that get rushed are cut straight off the roll without a relaxation rest, and they shrink 5 to 8% after the first wash. Write a maximum shrinkage into the spec and tie a payment milestone to it.

Budget the calendar too. Many small and mid-size Indian makers have no formal sampling procedure, so plan for two or three sample rounds (an India-based founder, 2023, who met 12 to 15 makers and had three sampling failures). One maker's first five-style, 300-piece order took about three months against a two-month demand (one case, not a norm, but the right shape).

4. Ordering Below The Inspection Threshold, Then Not Inspecting

Here's a pattern nobody tells first-timers. Third-party inspection usually gets bought only when "the order size is good", so small first orders go uninspected (an Indian maker on a forum, 2025). And the small uninspected order is exactly where things go wrong. It's the order too small for the main line, so it's the one that gets quietly sent to a subcontractor across town.

The people who know India best still inspect. A UK importer who owns a company in India said in 2026 that they still pay for checks before every shipment leaves. A former sourcing manager's rhythm: two inspections for a new order, pre-shipment only for repeats. Inspect regardless of quantity. On a first order of a few hundred pieces the inspection fee is small next to a container of returns, and it's the one thing that turns "they said" into "I saw". Quality control in India sets out the four inspection stages. SourcingSync (opens in a new tab) books and runs inspections at any quantity.

A factory visit is not an inspection either. One India-based founder visited factories in person and still couldn't get sampling right (2023). A visit tests intent. An inspection tests output.

5. Testing To One Border When You Sell Into Another

This is the blunder our own data surprised me with. We joined about 22,000 official border rejection records for Indian goods from January 2021 to September 2026 across the EU, US, Australia and UK. The same product fails for different reasons at different borders.

Spices are the clearest case. In the EU, farm pesticide residues drive the notifications (202 of 453 spice notifications) and Salmonella is just 19. In the US, Salmonella is the charge on 690 of 1,320 refused spice lines, with filth on 355. In Australia, Salmonella accounts for 77 of 88 failing spice records. One "export quality" certificate can't cover both, and a supplier's lab report is not a shield: 149 refused US spice lines and 113 rice lines had a private lab package on file and were refused anyway (RASFF, FDA import refusals and DAFF failing-food reports, our counts, to September 2026).

If you sell non-food, the same logic applies to the categories ecommerce lives on. In children's clothing, 22 of 27 EU safety alerts on Indian goods were cords, drawstrings and detachable poppers, and 8 of 45 US recalls of India-made products were sleepwear that failed flammability. Those are design faults a pre-production review catches. In fashion jewellery, cadmium ran 38 to 95% by weight in six EU alerts: an alloy choice, not contamination. In leather, chromium VI ran 5 to 20.5 mg/kg in eleven items against a 3 mg/kg EU limit (limit reported, not re-read) (EU Safety Gate and US CPSC, our counts, 2021 to 2026).

The rule - set your lab tests to your own destination's limits and to your own product's failure mode, and commission the test yourself on a sample drawn by an independent inspector. Whatever country you buy from, ask the same question. What does my border actually reject this product for?

6. Letting The Supplier's Courier Account Run Your Customs Entry

Your supplier's courier rate is two or three times cheaper than anything you can get on your own account (buyers on forums, 2024 to 2026), so you let them ship. Then the courier shows up as the importer of record, and any duty refund goes to the courier, not to you. That's exactly what happened to US buyers with the 2026 IEEPA refunds (buyers on forums, 2026).

The ground moved under everyone in 2025 and 2026, so date-check anything you read. India Post stopped taking US parcels in August 2025, and the US de minimis exemption is suspended, so every India-to-US parcel is now dutiable, and shipments above US$250 need a formal entry because the extra 10% Section 301 duty sits in chapter 99 of the tariff. In the UK, couriers bill import VAT even when you're set up for postponed VAT accounting, plus their own admin fee on top, even on a small parcel (buyers on forums, 2023 to 2025). In New Zealand, since 1 April 2026 every consignment under NZ$1,000 pays its own levy, and on a consignment over NZ$1,000 the Customs and MPI levies for sea come to roughly twice those for air (NZ Customs guide, April 2026), which flips the usual sea-is-cheaper logic for small orders. The thresholds for all six markets, and the parcel-versus-pallet sum, are in low-value import rules by country.

And if a supplier offers to split the invoice, mark the parcel "personal" or under-declare to "save you duty", say no. Buyers on US and UK forums call it what it is, customs fraud, and it lands on the importer. That's you.

7. Counting On "FTA = Free" Without The Origin Paper

Australia, the UK and soon New Zealand all have 0% deals with India, and sellers read the headline and stop. The duty saving is won or lost on paperwork after the goods have moved.

In Australia, ECTA preference needs a certificate of origin on each shipment. No certificate, no preference, and the general rate is charged by default (ABF). You can claim it back later with refund reason code 23A10A once you get the certificate, but the Indian side can only issue one retrospectively within 12 months of export. In the UK, CETA has been live since 15 July 2026 with three ways to prove origin. In New Zealand, the FTA takes effect on 20 October 2026, and brokers are already warning that goods won't qualify "just because they were shipped from India". Detail for each market: Australia, the UK, New Zealand.

Two traps inside the paperwork - Indian finished goods can contain Chinese cotton or yarn, and an input you never see can defeat the origin rule (a UK importer, 2026). Ask for the fabric source in the RFQ. And since April 2026, India's DGFT reportedly requires the invoice number on a preferential certificate of origin to match the shipping bill exactly, with a mismatch denying the benefit at destination (trade advisers on LinkedIn, 2026, so verify it on DGFT before you rely on it). Whatever the origin, ask the supplier whether they've issued the document for your market before, and for which product.

8. Paying Everything Before You've Seen Anything

Small Indian makers commonly want 70 to 100% up front, and they'll tell you why: foreign buyers, including long-standing US buyers, have paid late or not at all (seven exporter voices on forums, 2025 to 2026). Merchant exporters want an advance because their own suppliers want cash, so your money buys your goods. Trade-finance voices put typical advances at 10 to 30%. The blunder is paying the whole thing before you've seen the output.

Three things I'd do every time - pay only a corporate account in the exporter's legal name, and confirm any change of bank details by phone on a number you already had. Structure the payment around milestones you can see: PO, pre-production sample approved, pre-shipment inspection passed, bill of lading. What each method protects, letters of credit included, is in how to pay an overseas supplier safely. Platform protection is thinner than you think. PayPal disputes have made buyers ship the goods back to India at their own cost before any refund, and card chargebacks worked better (buyers on forums, 2023 to 2026).

One scam script to memorise - you pay the deposit, then a "minimum order fee" appears, then a "customs fee", then an "export fee" (buyers on forums, 2024 to 2026). Stop at the first unplanned fee. A real exporter quoted the whole price before you paid. More on terms in MOQ, pricing and payment.

The Short Version

Most of these blunders are one blunder wearing different clothes: trusting a promise where you could have checked an output. Check the record, not the badge. Approve the pre-production sample, not the proto. Inspect the order, not the factory tour. Test to your border, not to a certificate. Hold the customs entry and the origin paper yourself. Pay against milestones.

None of it is hard, just unglamorous, and it's what separates the sellers still importing in year three from the ones who quit after container one. Hope this helps. If you need help on the ground, you know where to find me.