Not Using ImportYeti
Is the New Not Using AI.
Here Is the Full Playbook.

A few years back I was paying $5.50 a unit landed from a U.S. manufacturer for a paper product. I moved it to a factory in China at $2.80 landed. Forty-nine percent reduction in cost. Returns dropped from 3 percent to 0.3 percent. The product got better and got cheaper at the same time.

That move was not AI-assisted. I researched it manually. I typed competitor names into ImportYeti, read every supplier page, cross-referenced on Alibaba, checked addresses on Google Maps satellite, sent emails. Old-school diligence with a free modern tool. Two factories replied. One quoted. One is still our manufacturer.

I signed up for ImportYeti in 2021, close to when it launched. Early paid subscriber. In 2024 and 2025 I started layering ChatGPT on top of the data: pasting supplier lists, asking it to classify factories versus trading companies, building outreach drafts. Now I do the same thing in Claude, which handles larger data sets and longer histories more cleanly. The platform did not change. What changed is that AI turned a research tool into an intelligence system.

I have used ImportYeti to research paper, ceramics, wood, puzzles, textiles. I have pulled competitor sourcing across the gift and puzzle space. I have bookmarked suppliers across all those categories and tracked how they move between customers over years. The intelligence I have built using public data and a free tool would have cost fifty thousand dollars a year on Panjiva five years ago.

That is what this article is about. If you run a gift or lifestyle brand and you are not in ImportYeti every time you evaluate a new product, a new supplier, or a competitive threat, you are working without data your competitors may already have. Not using it right now is exactly like not using AI two years ago. You will feel it eventually. This is the playbook.

Every time a container ship docks at a U.S. port, the cargo manifest is filed with U.S. Customs and Border Protection. Under the Tariff Act of 1930 (19 U.S.C. Section 1431) and the CBP regulation at 19 CFR Section 103.31, most of that manifest is a matter of public record. The disclosure regime was codified in 1981. It is not a data leak. It is the law.

ImportYeti founder David Applegate built a system during COVID downtime that indexed roughly 70 million of those records, cleaned and structured them, and made them searchable for free. Bootstrapped, unfunded, around three employees. Launched in 2020, went viral on Reddit. It is now the standard free tool for any operator who sources internationally.

What is in a bill of lading. The shipper name and address (your competitor's supplier), the consignee name and address (the brand or its 3PL), the notify party, vessel and voyage, port of loading, port of discharge, container number, piece count, weight, and a product description. ImportYeti layers on an HS code classification from the description text and builds a chart from the history.

What you see on a company page. Every supplier used since 2015, volume by supplier, country breakdown, port mix, shipment frequency, shipment descriptions, and a time-series chart of activity. You can see whether a brand is growing or declining from their import volume. You can see when they switched factories. You can see which factories serve multiple competitors in your category. You can map the entire supply chain of a product you are about to launch, in about forty-five minutes.

Free vs paid. The free tier covers unlimited company and supplier searches, the full dataset back to 2015, and all charts and visualizations. You can see everything. The paid tier (around fifty dollars a month for the sourcing plan) unlocks CSV export and Power Query. CSV export is the door that lets you take the data somewhere interesting, specifically into an AI.

Search the LLC, not the brand. Brands almost never import under their consumer-facing name. Step zero is finding the legal entity. Run the brand name through the USPTO trademark database and read the owner field. Or check the brand's terms of service or privacy policy, which usually discloses the legal company name. ImportYeti will then find every shipment filed under that entity, including name variants and alternate spellings that show up at the bottom of the company page.

The neighbor brand technique. Once you find a supplier, click into their ImportYeti page and read their full U.S. customer list. A factory that ships to multiple premium brands in your category has been independently vetted by each of them. That is a tier-one signal. A factory you find in ImportYeti that serves only one brand you have never heard of is a different conversation. The neighbors are the quality signal, not the factory's own claims.

Read the volume over time. Shipment frequency tells you how a brand manages inventory and whether their business is healthy. A regular monthly rhythm means a stable factory relationship and a working reorder cycle. A long gap followed by a volume burst often means a supplier swap, a quality dispute, or a product pivot. I have spotted competitor pivots from import data months before any public announcement.

FCL vs LCL. Full container load versus less-than-container-load shipments are now visible on ImportYeti. LCL shipments are small orders. A factory that regularly ships LCL to brands your size will almost certainly accept small initial orders from you. FCL-only factories have real MOQ floors. This single data point can save you weeks of back-and-forth with factories that were never going to take your volume.

Country mix over time. The year-over-year country breakdown shows whether a brand is China-only, actively diversifying into Vietnam or India, or already running a multi-country supply chain. One important note, and Applegate has said this publicly: many Vietnam and Cambodia shipments are Chinese goods finished or relabeled offshore to avoid Section 301 tariffs. Do not read a Vietnam supplier as confirmed China decoupling without verifying the finished-goods production location.

Watch for the swap. A supplier that represents more than 15 percent of a brand's volume and then disappears within 60 days, with a new supplier of similar product appearing in roughly the same timeframe, almost always means a quality dispute, an IP issue, or a deliberate sourcing upgrade. That pattern shows up clearly in the timeline view.

The most valuable thing ImportYeti shows you is not the brand. It is the factory behind the brand. But there is a persistent problem. A significant portion of the shipper names in import records are not factories. They are traders, agents, or sourcing middlemen presenting themselves as manufacturers. The commission they collect is invisible in the data, but it is real and it is coming out of your eventual cost structure.

Name alone does not always tell you the difference. This is where AI becomes the actual tool.

I use Claude, ChatGPT and Manus in combination. I do not pick one and stay there. I run the same supplier data through all three. Claude is better for large-context full-history analysis. ChatGPT Data Analyst is faster for charting and quick tables. Manus is useful when the research needs to reach across the open web and combine with public business registry data or Alibaba listings. I make them check each other. The disagreements between models are often where the real insight sits.

The workflow: export the supplier CSV from ImportYeti (paid plan) or copy-paste the supplier table from the free-tier company page. Feed it into the prompt in the next section. Ask for factory versus trader classification, a confidence score, and the top three signals behind each verdict. Take the likely-factory list, cross-reference each on Alibaba, pull the address on Google Maps satellite to confirm it is a real industrial building, and find a named sales contact on LinkedIn before you send anything.

That process used to take two weeks and a sourcing trip. With AI it takes an afternoon.

Yiwu address is almost always a trader. The city is China's largest small-commodity wholesale hub and very few actual manufacturers are based there. Shenzhen CBD addresses are high-probability agents or sourcing reps. Dongguan, Chaozhou, Ningbo, Foshan, Cixi: real production clusters. Karur, Tamil Nadu: real home textile production. Jaipur, Rajasthan: real block-print production. Sivakasi, Tamil Nadu: real safety matches. If the address is in the right city for the product and the building is in an industrial park, you are probably looking at a factory.

The way I work: paste or export the supplier list from ImportYeti, run the same prompt through Claude and ChatGPT, compare the outputs, and use the disagreements as starting points for deeper research. Manus comes in when I need the AI to independently verify something on the open web, like checking whether a Ningbo address actually shows as a factory on Baidu Maps or whether an Alibaba profile matches what the BOL says.

Six prompts that do the most work.

Each product category in the gift and lifestyle space has its own manufacturing geography. Knowing which city to expect when you look up a supplier is one of the fastest ways to validate whether a name in ImportYeti is a real factory or a trading intermediary.

Air freight is invisible. The public-disclosure statute covers ocean vessel manifests only. Air freight, rail and truck are not included. A brand that ships by air leaves no trace in ImportYeti. If a competitor has no ocean import history but clearly has product in stores, they are either manufacturing domestically, using air freight, or have filed a confidentiality request.

DDP shipments erase the supplier. When a brand buys on Delivered Duty Paid terms, the supplier becomes the importer of record and their name goes on the BOL as the consignee, not the brand. A significant portion of small DTC brands use DDP, especially on early production runs. This can make a brand look invisible in ImportYeti even when they are actively sourcing.

Domestic manufacturing leaves no trace. U.S.-made product does not generate an ocean BOL. Brands that manufacture in the U.S., Canada, Mexico by truck, or EU domestically are outside the dataset unless they also import components.

Confidentiality requests remove brands entirely. Free to file, twenty-four hours, two years of coverage. If your competitors figure this out and you do not, your sourcing is visible to them while theirs is hidden from you. Worth knowing.

What the data cannot tell you even when it finds the right factory. Quality is not in the BOL. MOQ for new accounts is not in the BOL. Whether they will respond to your email is not in the BOL. Lead times, defect rates, communication speed, certifications, ethics and compliance are all invisible. ImportYeti finds the factory. What happens next is still your job.

Transshipping distorts country signals. Many shipments that appear as Vietnam or Cambodia origin are Chinese goods finished or relabeled offshore to avoid Section 301 tariffs. A Vietnam supplier in your search results is not automatic proof that a brand has meaningfully diversified its China exposure. Verify the finished-goods production location separately before drawing supply-chain conclusions.

Naming inconsistency is real. The same supplier may appear under four different name spellings across different BOL records, especially when different freight forwarders file the document. ImportYeti does some de-duplication but it is not perfect. When building a complete supplier picture, check the "other names" section on every supplier page and run partial-name searches.