Fixing Product Listing Issues in Ecommerce

Most product listings fail before they go live. Here’s how Shopify fixed it—for millions of merchants—using one small AI model: Everyone wants their products to rank high in search. But messy listings kill visibility. Wrong tags, unclear descriptions, missing categories. Shoppers scroll right past. Shopify faced this at scale. Millions of merchants. Billions of product images. Over 10,000 potential categories. Manual tagging wasn’t cutting it. Closed AI models were too expensive to run daily at that volume. So they built their own solution. A fine-tuned, open-source LLaVA model powered by Llama 2 7B. It analyzes product images, suggests categories, creates metadata, and even flags violations—all in real time. It now processes tens of billions of tokens daily across 100 GPUs. Merchants get smarter listings. Shoppers get better search results. No expensive fees. No bloated infrastructure. The entire product catalog updates faster, more accurately, and with zero added effort. Want listings that actually work? Start by fixing your metadata. Or let the right AI do it for you.

Sellers blame their product when returns spike. Sometimes that's right. Most of the time it isn't. I've sat with founders who tore apart their entire supply chain trying to fix a return rate problem that had nothing to do with the product. Better materials. New supplier. Tighter QC. Months of work and tens of thousands of dollars chasing a defect that didn't exist. The actual problem was usually one of three things, and none of them required a product change. The listing promised something the product wasn't built to deliver. Bullets oversold the feature set. Images made the product look bigger than it actually is. The lifestyle shots set an expectation the unboxing couldn't meet. Buyers received exactly what was shipped and returned it because it didn't match what they thought they were buying. Or the instructions failed. The product worked fine, but the customer couldn't figure out how to use it in the first 10 minutes. They didn't email support. They didn't watch a YouTube video. They just clicked return and left a one-star review calling the product cheap. Or the fulfillment experience created the impression of a quality problem. Damaged box. Slow delivery. Tape that took ten minutes to cut through. The customer opened a package that felt like a bad first date and decided the product inside must be bad too. All three of these look identical on the return dashboard. "Product not as described." "Quality issue." "Defective." None of them are actually about the product. Before you redesign anything, read 50 of your returns and 50 of your one-star reviews in one sitting. Sort them by what actually went wrong. You'll usually find that the listing copy, the instructions, or the unboxing is doing more damage than the product ever did. The cheapest product upgrade you'll ever ship is a better explanation.

Promotions stuck? It’s not your offer, it’s what Amazon thinks your product is. Pretty annoying when you plan a whole promotion strategy, but the system gives you errors. That’s what happened to a seller who came to us, stuck with error 5885. Here’s what that means: Amazon flagged their branded ASIN as generic, which means they "could not change a product detail page, or add offers on another seller’s generic product, or copy the product detail page to another store". Which made no sense because it wasn’t just branded. The actual brand owner was the one making the changes. And the brand name was clearly displayed on the detail page, and the backend data supported it. But still, no listing updates. No ability to create promotions. Just that dreaded error message. First, we diagnosed the issue, and what we found was something really interesting. What Amazon calls “branded” isn’t just about having your logo on the product or your name in the backend. There are three deeper internal attributes (that I will dedicate a post to) that need to align perfectly for the system to recognize your ASIN as branded. In this case, the Amazon Retail team had contributed to one of the attributes, the brand name (which made it look legit), but some of the other internal attributes were off. To the system, that meant the product was generic, end of the conversation. So we built a workaround. We downloaded a new category listing report. Filled in only the most critical fields. Then we submitted a full update using the product’s UPC instead of the ASIN, All this to recreate the listing from scratch, essentially. That reset the branding attributes. And it worked. Listing: unlocked. Promotions: enabled. Error 5885: gone. It sounds simple in hindsight, but I'm not going to lie—it took us some solid tries to make it happen. That's why understanding how Amazon reads your backend is so important to problem-solving on Amazon After this case, I think we are sure of some things. • Just because your listing looks branded doesn’t mean Amazon agrees • Misaligned attributes can quietly block your growth • Flat files (done right) can be your best friend • Solving the issue isn’t about fighting Amazon, it’s about speaking its language If your listings are misclassified or your promotions are stuck, there’s probably more happening under the hood than you realize. But what's crazier is that this is just part one of the issue, because even though the error was gone, Amazon still held one attribute. Want me to break down how brand attribution really works and get more control over this in a future post? Let me know. #AmazonSellers #AmazonFBA #ListingErrors #Error5885 #BackendOptimization #FlatFiles #EcommerceStrategy #AmazonBranding #AmazonSupport #CaseStudy

In eComm, PDPs aren’t just product pages—they’re potential drivers of SEO success. So, what’s the best approach when a product is out of stock or sold out? Every decision you make about an unavailable product page can influence your site’s visibility, potential search performance, and, most importantly, user experience. And with a thoughtful strategy, you can retain the SEO value you’ve worked so hard to build while keeping the customer journey seamless. Here’s how to approach it: 𝗜𝗳 𝘁𝗵𝗲 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝗶𝘀 𝘁𝗲𝗺𝗽𝗼𝗿𝗮𝗿𝗶𝗹𝘆 𝗼𝘂𝘁 𝗼𝗳 𝘀𝘁𝗼𝗰𝗸... Keep the PDP live. This helps preserve its SEO equity and ensures it’s ready to pick up where it left off once the product returns. ↳ Add a “Notify me when available” option to capture customer interest. ↳ Ensure your Product Schema is updated to show accurate availability in search results. 𝗜𝗳 𝘁𝗵𝗲 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝗶𝘀 𝗽𝗲𝗿𝗺𝗮𝗻𝗲𝗻𝘁𝗹𝘆 𝗱𝗶𝘀𝗰𝗼𝗻𝘁𝗶𝗻𝘂𝗲𝗱... If the page still holds value (organic traffic or backlinks): ↳Redirect it to the most relevant category page. This maintains SEO equity and helps guide users to similar options. If the page doesn’t hold value: ↳ Use a 410 (Gone) status to remove it from search engines. Clean up internal links and remove the page from your sitemap for a cleaner structure. 𝗜𝗳 𝗶𝘁’𝘀 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁 𝘁𝗼 𝗰𝗼𝗺𝗺𝘂𝗻𝗶𝗰𝗮𝘁𝗲 𝘁𝗵𝗮𝘁 𝘁𝗵𝗲 𝗽𝗿𝗼𝗱𝘂𝗰𝘁 𝘄𝗼𝗻’𝘁 𝗿𝗲𝘁𝘂𝗿𝗻... Leave the PDP live, marked as “Discontinued,” while prominently featuring alternative products. This approach supports user experience and ensures visitors still find value on your site. With a proactive plan, out-of-stock products don’t have to be a roadblock—they can become opportunities to maintain traffic, enhance user trust, and avoid annoyances your SEO strategy.

7 out of 10 of my projects start with fixing what most people ignore. This includes: - making copy easier to read - making images informational - making product name impactful Simple, but yet forgotten. In this post, using URturms example, I'll be sharing 11 underestimated changes that can increase your website sales. 1. Adding breadcrumbs. Important if you drive ad traffic to the PDP directly. They take shopper to the parent category page. Reducing bounce rate. 2. Adding a badge. Like "Bestseller", "Most Loved", "Few Left". This reassures the shopper that they're making the right decision. 3. Making images easier to swipe. Add a sneak peek of the next image along with navigation dots that show the count. Cap them at 8. 4. Making the product name impactful. Add key USPs. Show your current product name to 10 people. Do they understand what it is? 5. Add a short description below product name. Keep it in 1 line. Highlight it's most important feature here. 6. Consider adding an offer close to price. This motivates the shopper as they see some potential savings or benefit. 7. Highlight key product strengths in bullets or with icons. Avoid sentences. Keep this before the add to cart CTA. 8. Keep your add to cart CTA full width. Don't combine it with quantity or another CTA next to it. Make sure it's readable and prominent. 9. Highlighting shipping time or return policy below the CTA. This solves for common questions - when will I get it? can I return it? 10. Cross-selling complementary products. Like bottoms with tops. Earrings with necklace. Do this close to the add to cart CTA. 11. Adding 'Benefits' to your accordion. This gets a higher click through rate, while helping shoppers understand why they should buy this. Other UX/UI changes I did: - Removed quantity button - Made the information bar non-moving - Removed log-in, moving search next to cart - Changed the font for product name and CTA - Increased font size in places for better readability Found this useful? Let me know in the comments! P.S. If you want to maximize your PDP’s potential, start by understanding your visitor's behavior and the gaps. Get heat maps for your site (Microsoft Clarity is free). Observe what they like to (and don't like to) interact with.

Most ecommerce product pages fail at SEO because they optimise for just keywords. Not entities. Not search intent. Not AI visibility. After working on hundreds of e-commerce projects, I've built a complete framework for Product Page Optimization that covers everything from semantic foundations to AI search readiness. Here's what it includes: → Entity-Attribute-Value framework → Query intent mapping for transactional, commercial, informational, and navigational searches → Title tags, meta descriptions, and URL structures that actually work → Schema markup implementation → Image and visual content optimization → Reviews and UGC strategy for SEO → Technical SEO checklist for product pages → Internal linking and site architecture → Handling product variants without creating duplicate content → Discontinued product strategy (stop losing that link equity) → Conversion elements that support both rankings and sales → AI search and conversational commerce optimization The mind map below breaks down each section visually. Use it as a checklist next time you audit or build a product page. If you prefer a detailed PDF version for a clearer view, I'll DM it to you: 1- Make sure we're connected 2- Comment "FRAMEWORK" P.S. - Repost for priority access

Still treating your product detail page like a digital flyer? ❌ That “good enough” mindset is quietly killing your revenue. 👇 𝗟𝗲𝘁’𝘀 𝘁𝗮𝗹𝗸 𝗽𝗿𝗼𝗼𝗳: A client came to Blikket frustrated—tons of traffic, but conversion rates stuck at 3.5%. We overhauled their product detail page using 3 core UX shifts: → 𝗜𝗻𝗳𝗼𝗿𝗺𝗮𝘁𝗶𝗼𝗻 𝗵𝗶𝗲𝗿𝗮𝗿𝗰𝗵𝘆: Focused on ONE headline, one primary CTA.  → 𝗩𝗶𝘀𝘂𝗮𝗹 𝗰𝗹𝗮𝗿𝗶𝘁𝘆: Stripped distracting banners, highlighted product benefits above the fold. → 𝗙𝗿𝗶𝗰𝘁𝗶𝗼𝗻𝗹𝗲𝘀𝘀 𝗽𝗮𝘁𝗵: Reduced variants, surfaced key reviews with icons (not endless text). 𝗥𝗲𝘀𝘂𝗹𝘁: Conversion rate jumped to 𝟳.𝟭% in 30 days. That’s a 𝟭𝟬𝟬% 𝗹𝗶𝗳𝘁—not from guesswork, but from focused UX changes. ✅ Stop stuffing. Start prioritizing what the buyer actually needs. Where’s your biggest UX blocker right now? Drop your pain point below—let’s troubleshoot. https://lnkd.in/gedvrCnM #eCommerce #UXDesign #CRO #ProductPages

Is Your Amazon Rank Dropping Without You Knowing? 🚨 What if I told you Amazon could secretly be changing your product’s classification—like switching your item type keyword—without any notice? This silent killer can tank your organic rank, force you to overspend on ads, and drain your sales without you realizing why. The worst part? Everything looks fine on the front end, but in the backend, Amazon’s bots might fill in missing fields with irrelevant data, completely wrecking your listing’s chances of success. Here’s what sellers often miss: • Amazon Bots Love to Change Things: One day, your product is categorized as “Pajama Sets,” and the next, it’s “Pajamas Sets.” A small change? Sure. But this tweak can derail your rank, PPC performance, and visibility. • Seller Central Suggestions Aren’t Always Helpful: Just because Amazon suggests it doesn’t mean it’s right. Accepting these changes blindly could lead to costly mistakes. Here’s how you stay ahead: 1. Audit Monthly: Regularly check for backend errors if your rank or ad performance suddenly drops. 2. Use the Category Listing Report: Download it from Seller Central to uncover and fix hidden issues. 3. Verify with the Browse Tree Guide (BTG): Ensure your item type keywords and browse nodes match your category. Missteps here can crush your performance. 4. Fill Out Every Field: Skipping optional attributes? Amazon bots will fill them in—and usually get it wrong. 5. Stay in Control: Never let Amazon automatically update your listing. Reclaim it before it’s too late. I dive deeper into all these issues (and real case studies) in my new 20-minute video. 📹 If you’re serious about protecting your rank and sales, check it out—link is in the comments or search my name on YouTube. 👉 Don’t let backend mistakes destroy your listings. Take control before it’s too late. #AmazonSellers #AmazonAdvertising #AmazonPPC #AmazonFBA

Your listing is no longer just being evaluated as a static page built around keywords. It is now being interpreted by Rufus based on shopper context, and you are being judged on whether your PDP provides the right proof for the question being asked. With the recent update to Rufus’ shopping memory, answers are shaped by intent, not just search terms. That means your product detail page is scanned for evidence that supports specific use cases in real time. Take a blender. A shopper searching for “smoothies with easy cleanup” will trigger different surfaced details than someone searching for “high-protein shakes fast.” One context prioritizes dishwasher-safe parts and durability. The other prioritizes motor power, texture, and speed. Same product, different proof. The same pattern applies across categories: - Running shoes for marathon vs trail - Air purifiers for allergies vs quiet bedrooms - Backpacks for carry-on travel vs hiking - Baby monitors for newborn sleep vs travel Rufus looks for signals that support the query's intent. If that proof is not clearly shown in your images, specs, comparison charts, and FAQs, you are less likely to be surfaced as the best answer. Most sellers still build listings around generic feature lists. That worked in a keyword-driven environment. In an AI-assisted environment, evidence needs to be structured around shopper scenarios. This is why we built the COSMO / Rufus Readiness Report. We score your listing based on how well it supports real shopper contexts and how clearly your proof is structured for Amazon’s AI systems. If you want to see where your PDP stands and what gaps may be costing you visibility and conversions, get your free report here: https://t2m.io/3dnRQxS

20 critical errors that silently destroy your organic reach on Amazon After auditing 400+ brands, I see the same mistakes killing conversions daily Your competition isn't beating you with better products, they're winning because they fixed these foundational errors while you're still guessing. 👉 Suppressed listings lose 80% of search visibility overnight 👉 Missing backend keywords = zero indexing for high-value terms 👉 Wrong categories kill your browse path completely 👉 Flat file corruption makes detail page changes impossible Here's what separates winning Amazon Sellers from the rest: structure beats creativity every time. You can have the most beautiful listing in your category, but if your ASIN is suppressed or your keywords aren't indexing, you're invisible. So treat listings like technical systems, not just marketing brochures Every backend field matters. Every image specification counts. Every keyword placement is strategic. Which error is your bane to existence?