What Fashion Brands Should Be Doing Before A Drop
Something I've personally noticed across several drop-model brands I've worked with: the ones that struggle to capture AI-driven traffic during a drop almost always have the same problem. Not their paid strategy. Not their creative. Their entity data hasn't been touched in months. Sometimes even longer and by the time the drop window opens, they're presenting a stale, low-confidence identity to every AI agent fielding queries about their brand.
The thing about AI-native discovery that most ecommerce teams don't fully appreciate is that it's not a real-time system in the way paid search is. When someone asks Perplexity or a shopping AI "where can I buy the new X,Y,Z Super Collection Hoodie today?", the agent isn't doing a fresh crawl of the internet. It's working from a combination of cached entity data and recent signals. Recency matters. A brand that has been actively feeding the entity graph in the weeks prior is presenting a richer, more recently validated identity than one that hasn't.
This matters more in a drop model than in continuous retail, precisely because of the compression. You have 24 to 72 hours where your brand is getting the highest concentration of AI-driven queries in a hours. If your entity isn't ready going into that window, you can't fix it in real time. The preparation happens in the four weeks before.
Here's what that preparation actually looks like.
Get your new SKUs into the entity graph before launch day
This is the most urgent thing and the most commonly left until it's too late.
New drop products are unknown entities on launch day unless you've done the work beforehand. Your Product schema needs to be live, your GTINs need to be mapped and verified against GS1, and your brand and manufacturer fields need to be correctly populated. AI shopping agents need this before the product page goes public. Not on launch day. Not the week before. Four weeks out is when I want this on someone's checklist. AI shopping agents need to be certain they are recommending and sourcing exactly what the user is looking for.
The reason it gets missed is that product setup for a drop is usually handled close to launch, when creative assets are finalised and copy is approved. Often these are always last minute. Schema and GTIN work tends to sit in a different team's lane, or in nobody's lane, and gets treated as something to tidy up after launch. By which point the first 48 hours of your highest-traffic window have already passed with your new SKUs sitting as unresolved entities.
One thing I now specifically check before any drop I'm involved in: are the GTINs for new products assigned and mapped before the staging environment is built, not after. It sounds obvious. It almost never happens by default.
Feed the entity graph with authored content
This is the layer most ecommerce teams don't connect to drop preparation at all, and I understand why. It doesn't feel like it belongs in the same conversation as schema and product data. But it's doing real work.
AI shopping agents consider recency. An entity that has published fresh, authored content in the weeks before a drop is a more recently validated entity than one that hasn't published anything since the last collection. When I say authored content I mean anything. It could be a collection story, a design note, a piece on the trend the drop is responding to. Something that is properly marked up with Author schema linking back to a real person in your Organisation entity. This builds trust and confidence in the AI shopping agent.
This isn't content marketing for its own sake. It's entity maintenance. Publishing one solid piece of authored content three to four weeks before a drop, marked up correctly, refreshes your authorship signals and gives AI agents recent, verifiable content to associate with your brand. It's twenty minutes of schema work on top of content you're probably creating anyway.
If your creative or brand team is writing anything in the pre-drop period, which they most certainly should be, make sure it's going on your domain with proper Author markup. Not just on Instagram. Not just in the press release. On your site and attributed correctly.
The review timing problem for drop brands
Most brands I work with solicit reviews about a week after a drop, which is roughly right on timing. The problem is almost none of them are thinking about where those reviews sit in their entity structure.
For a standard ecommerce brand with a stable catalogue, product-level reviews accumulate over time and build a rich signal layer for each SKU. For a drop model where a lot of your products are one-time only, a product-level review on a SKU that will never be available again has limited forward value in terms of entity trust. That product isn't going to be recommended to a future buyer.
What matters more in this instance is brand-level review signals. Are your post-drop reviews being captured on your own domain with Organisation-level Review schema? Or are they sitting on Trustpilot or Google associated with your entity but not reinforced by your own structured data? The distinction matters. A review that exists on your domain, marked up correctly, is a direct input to your entity's trust signals. A review on a third-party platform is an indirect one. Many brands are tied into TrustPilot so that’s a strategic decision. I would personally opt for your own domain.
There's a second opportunity here that I don't see many brands using. If you have early access buyers before launch, their reviews during the drop window are the only window you'll have for that SKU. Two or three verified reviews on a PDP before the drop goes live, properly marked up, means your new products are entering the entity graph as reviewed, verified items rather than blank unknowns. That's a meaningfully different starting position for a limited window.
What prepared looks like versus unprepared
Four weeks out, a prepared brand looks like this: new SKUs have GTINs assigned and verified in GS1, Product schema is live in the staging environment, at least one piece of authored content is published or scheduled, and the post-drop review flow is set up to capture brand-level reviews on the domain with proper markup.
An unprepared brand four weeks out looks like most of the brands I walk into. Schema hasn't been touched since the last site update. New product data won't be finalised for another two weeks. Nobody has thought about authored content. Reviews are going to Trustpilot by default because that's how it was set up three years ago and nobody changed it.
The gap between those two situations isn't enormous in terms of effort. It's mostly a coordination problem. Getting the right things on the right checklist early enough that they actually happen before the window opens.
Why this compounds across drops
Each drop where you do this work makes the next one better. Your entity gets richer. Your GTIN coverage expands with every new collection. Your authorship trail lengthens. Your brand-level review depth builds. By the third or fourth drop where you've been deliberate about this, you have a fundamentally different entity profile than a brand doing it for the first time.
This is your competitive advantage.
The counterfeit operations trying to intercept your drop traffic have none of this history. Their entity is shallow by definition. The gap only widens the longer you keep going.
It's not the most visible part of drop preparation. But in a model where everything rides on a compressed window, it's often the part that determines how much of that window you actually own.