From First Copy to First Takedown: a 30-Day Plan

It was 23:15 on a Friday when Ife found it, which is when everyone finds it, because that's when you're on your phone looking at nothing in particular.

Her bag. Her exact bag, the one with the asymmetric strap that took 14 prototypes, photographed on her own white sweep, with her own model's hand in frame, for 34 dollars.

She spent the next 3 hours doing the thing everybody does. Reading the seller's other listings. Reading the reviews. Reading the 4-star review from someone who said the stitching came apart but the colour was lovely. Drafting a message to the seller at 02:00 and not sending it, which was the one good decision of the night.

By Saturday morning she had 2 problems. The copy, and no idea what to do first.

Here's the order. 30 days, roughly 6 hours of work in total, and a cadence at the end of it instead of a project.

Week 1: count, don't act

The instinct is to file immediately on the listing you found. Don't. You'll spend 40 minutes on 1 listing and discover in week 3 that there were 9, from 4 sellers, and that your first notice taught the biggest one to be careful.

Day 1 and 2. Find them all. Pick your 3 best-selling products. For each, reverse-image search the hero shot on Google Lens, on TinEye, and on Bing Visual Search. Different crawls, different results, 10 minutes each. Then search your exact product title in quotes on every marketplace you can think of, including the ones you don't sell on.

Then the highest-yield search in the whole plan, which takes 4 seconds: search your invented words. Colourway names, model numbers, made-up materials. A copycat lifting your listing keeps those without realising they're a signature. Your product photos are your fingerprint has the full method.

Day 3. Build one spreadsheet, organised by seller. Columns: shop name, seller ID, listing URL, listing title, price, date seen, notes. One row per listing, sorted so all listings from one shop sit together.

Capture the seller ID specifically. After a takedown the listing URL dies and the seller ID is how you find your way back. This is the single most commonly skipped field and the one you'll want most in month 3.

Day 4. Screenshot everything. Full page, with the URL and the date visible. Save into folders named after the shop, not the listing. Listings vanish; your proof shouldn't vanish with them.

Day 5. Write the count down somewhere. Ife's was 9 listings from 4 sellers, 2 of which turned out to be the same operation shipping from the same city with the same 12 to 18 day handling window.

Do nothing else this week. The count is the deliverable.

Week 2: work out what you can actually claim

This is the week that decides whether your notices work, and it's the one people skip because it feels like admin rather than action.

Day 8. Audit your image ownership. For each hero image, answer one question: can I prove I own this? A shoot you did yourself, yes. A commissioned shoot with copyright assigned in the contract, yes. An agency shoot where the licence stayed with the agency, probably not. A supplier catalogue image, no, and neither does the copycat, which means that listing isn't attackable on copyright at all.

Sort your images into those buckets once. It takes an afternoon and it tells you which half of your range is defensible.

Day 9. Check your trade mark position. If copycats are using your brand name in titles or shop names, copyright won't help and you need a registration. A UK application filed online through the Intellectual Property Office costs 205 pounds for one class since the fee change on 1 April 2026, plus 60 pounds per additional class, and you can do it yourself in under an hour. It won't help this month, and it's what gets you into the platform programmes for the next 10 years.

Day 10. Decide which notice for each row. Photograph copied, copyright. Brand name copied, trademark. Product shape copied and registered as a design, design right. Genuine goods sold without permission, not an intellectual property matter at all, and filing one anyway is how brands lose their reporting access. DMCA or trademark? is the decision in detail, including what the notice has to contain.

Day 11 and 12. Write the stockists page. One page on your own domain listing every official store and authorised stockist, by name, with links, plus a plain sentence about what an authentic one costs. 300 words. Stable URL. Linked from your footer.

This is the free asset that does the most work, and it does 3 jobs at once: customers check it, platform reviewers read it, and AI assistants quote it when a buyer asks whether a shop is legitimate. It also becomes your allowlist, so your own stockists stop appearing in your own match reports as findings.

Week 3: file, once, properly

Day 15. Pick one seller, not one listing. The worst offender, by volume and by price. If that seller has 6 listings, you're filing on 6 listings in one sitting, because the seller is the unit that matters.

Day 16. File. Find the right route first; the platforms move these pages and bury some of them behind a sign-in. We keep a free, hand-checked index at brandprotectiontools.com/data/takedown-routes, currently covering 12 platforms and channels at version 2026-11-06, with 10 routes confirmed by opening them and 1 still marked unverified.

One rule from it that saves real trouble. For a copy sold from a standalone website rather than a marketplace, the notice goes to the company hosting the site, never the domain registrar. A registrar can only switch off a whole domain, which is disproportionate and the quickest way to have your complaint thrown out.

Read what you sign. The DMCA notice includes a statement, under penalty of perjury, that you're the owner or authorised to act for the owner, set out in 17 U.S.C. 512(c)(3). Section 512(f) makes knowing misrepresentation actionable in damages. That's why this stays a human act and why no tool should file on your behalf.

Day 17. Log it. Date filed, platform, route used, seller, listings covered. That log is the thing that becomes an escalation case later.

Day 18 to 19. Nothing. Review cycles run from 2 days to 3 weeks. Resist the urge to file 8 more notices while you wait, because you want to see what one notice does before you repeat it 8 times.

Week 4: track, then set the cadence

Day 22. Check the status. Down, still up, or contested. If a counter-notice has arrived, the clock is statutory: under 17 U.S.C. 512(g) the platform restores the material in not less than 10 and not more than 14 business days unless you tell it you've filed a court action. For a 34-dollar bag the honest answer is usually that you won't, which is fine. Log it and let it feed the case you're building on that account.

Day 24. Check for relisting. Same seller ID, your photographs with a border or a mirror applied, your invented words still in the text. Relisting is the normal outcome, not a failure. The relisting problem is why the right response is to build a repeat infringer case on the account rather than to keep filing listing by listing.

Day 26. File on the second seller, now that you know how long the first one took and what the platform asked for.

Day 29. Write the cadence down. 30 minutes a month: reverse-image your top 3, search your invented words, check last month's filings, update the spreadsheet by seller. 1 hour a quarter: the full range, the repeat sellers, and a re-read of the stockists page.

Day 30. Decide whether to automate. If the manual month was comfortable, don't. If the sorting was the bottleneck rather than the searching, that's the signal. Brand Protection Tools, which Verity AI runs, starts at 99 dollars a month for 1 brand and 10 products with a weekly scan, no setup fee and no contract, and it does a free single-product scan that returns a count rather than an estimate. It searches Etsy and Alibaba.com today, with AliExpress, Temu, eBay and Amazon listed as next rather than live. Confirming and filing stay with you, which is the point. The small-brand playbook has the fuller comparison, including the 3 things not to buy.

What the 30 days actually bought

Ife got 5 of the 9 listings down. 2 were rejected because the sellers had reshot the product and dropped her name, so she had no claim on either. 1 came back after a fortnight. 1 seller stopped entirely.

More usefully, she has a count, a seller-level log, a stockists page and a 30-minute monthly habit. The Friday-night panic has been replaced by a spreadsheet, which is a worse story and a much better business.

The thing she wishes she'd understood on day 1 is that taking the copies down doesn't fix the recommendation. Buyers increasingly ask an assistant where to buy, and a machine can only name what it has been told about. When the AI shopping assistant recommends the fake is that argument in full, and it's the reason the stockists page in week 2 matters as much as the notice in week 3.