Mastering Product Recommendations for Unique Collectibles on Shopify Dawn: A Community Deep Dive

Hey fellow store owners! I recently stumbled upon a fascinating discussion in the Shopify community that really hit home for anyone running a niche store with unique, one-off products, especially antiques or collectibles. Mathiasmeyerco, who runs a store with 500+ unique items but only 1-2 sales a month, was struggling with Shopify's native product recommendation engine. Sound familiar?

The core problem, as many in the thread pointed out, is that Shopify's default recommendations rely heavily on co-purchase data and co-view signals. When you're selling unique items that rarely get bought together, the algorithm simply doesn't have enough data. The result? Semi-random, often irrelevant suggestions – like a signed jersey showing up next to a handwritten letter. Not exactly inspiring confidence or driving sales!

Why Native Recommendations Fall Short for Unique Catalogs

It boils down to data, or the lack thereof. As Recomento aptly put it, the collaborative filter has "essentially zero signal to work with." This means it falls back to guessing, often based on broad descriptions, which can lead to incoherent pairings. Jack_BuildsShopify wisely suggested that for a catalog like this, we shouldn't wait for purchase history to get smarter. Instead, we need to make the recommendation logic deterministic, driven by our catalog data.

The community quickly rallied, and the consensus was clear: you absolutely don't need to switch themes. Dawn, Shopify's popular free theme, can handle this with a bit of customization or by leveraging existing tools. Mark3e1 and Ploqo both confirmed this, emphasizing that a collection-based fallback or metafield-driven setup is far more relevant.

The Best Practices: Manual Curation, Structured Data, and Smart Fallbacks

So, what's the secret sauce? It's a combination of smart data organization and a hierarchical approach to recommendations. Here's a breakdown of the best strategies discussed:

• Manual Overrides for Key Items: For your highest-value or most popular items, hand-picking related products is gold. This gives you ultimate control over relevance.
• Narrow, Relevant Collections/Metafields: Instead of broad categories like "Football," use more specific groupings. DanielAnderson and Jack_BuildsShopify highlighted that for collectibles, fields like "person/team," "era," "subject/type," and even "price band" are far more useful. Two signed jerseys aren't substitutes if one is a Hall of Famer and the other a backup!
• Collection-Based Fallbacks: For the majority of your catalog, having recommendations pull from the same collection is a sensible, low-maintenance approach.
• Excluding Irrelevant Items: Always exclude the current product, unavailable/sold items, and generic collections (like "Home," "Sale," "New Arrivals") from recommendations.

Your Free & Low-Code Fixes

1. Shopify's Free Search & Discovery App (No Code)

This is a fantastic starting point for manual curation. Recomento, Ploqo, and v.marychenka all pointed to it. Here's how:

1. Install Shopify's free Search & Discovery app.
2. In the app, navigate to Product recommendations.
3. You can manually assign 'Related products' or 'Complementary products' per item.

This is great for your top 30-50 highest-value pieces. You can group them into tight, themed manual collections (e.g., '1960s Sports Memorabilia') and use the app to point related slots to that specific sub-group. It's a no-code way to ensure your hero products get relevant recommendations.

2. Custom Liquid for a Smart Collection Fallback (Code Snippet)

For a more scalable solution that covers your entire 500+ product catalog without manual linking for every single item, the community offered a brilliant Custom Liquid snippet. Ploqo provided a very comprehensive version, which I'll share here. This essentially takes over the native recommendations block, first checking for your hand-picked items (using a metafield) and then falling back to relevant items from the same collection.

Here's how to implement it:

Step 1: Add a "Related products" Metafield

This metafield will allow you to manually pick specific related items for your key products.

1. Go to Settings > Custom data > Products.
2. Click Add definition.
3. Name it Related products.
4. Set Type to Product.
5. Turn on List of values.

Step 2: Pick Related Items on Your Key Products

For your most important products, manually select a few truly related pieces.

1. Open a product you want to curate.
2. Scroll down to the Metafields section.
3. Under Related products, add 2 to 4 truly related items.
4. Leave this empty for products you don't want to hand-curate; they'll automatically fall back to collection-based recommendations.

Step 3: Add the Custom Liquid Section, Then Hide the Old One

This code will display your hand-picked items if available, or fall back to collection-based items otherwise. Remember to duplicate your theme first as a backup!

1. Go to Online Store > Themes > Customize your Dawn theme.
2. Navigate to a Product template (e.g., Default product).
3. In the left sidebar, click Add section.
4. Choose Custom Liquid.
5. Paste the following code into the Liquid code box:

{%- liquid
  assign limit = 4
  assign picks = product.metafields.custom.related_products.value
-%}

  
You may also like
  

    {%- if picks != blank -%}
      {%- assign shown = 0 -%}
      {%- for item in picks -%}
        {%- if shown < limit and item.available -%}
          

            
              {%- if item.featured_image -%}{{ item.featured_image | image_url: width: 500 | image_tag: loading: 'lazy', widths: '200,300,400,500' }}{%- endif -%}
              {{ item.title }}
              {{ item.price | money }}
            
          
          {%- assign shown = shown | plus: 1 -%}
        {%- endif -%}
      {%- endfor -%}
    {%- else -%}
      {%- assign skip = ',frontpage,all,sale,new,new-arrivals,featured,home-page,' -%}
      {%- assign seen = ',' | append: product.id | append: ',' -%}
      {%- assign shown = 0 -%}
      {%- assign cols = product.collections | sort: 'all_products_count' -%}
      {%- for c in cols -%}
        {%- assign probe = ',' | append: c.handle | append: ',' -%}
        {%- unless skip contains probe -%}
          {%- for item in c.products limit: 60 -%}
            {%- assign key = ',' | append: item.id | append: ',' -%}
            {%- if shown < limit and item.available -%}
              {%- unless seen contains key -%}
                

                  
                    {%- if item.featured_image -%}{{ item.featured_image | image_url: width: 500 | image_tag: loading: 'lazy', widths: '200,300,400,500' }}{%- endif -%}
                    {{ item.title }}
                    {{ item.price | money }}
                  
                

After pasting the code, find the native "Related products" section in your theme editor's sidebar (it's usually below the Custom Liquid section you just added). Hover over it, click the eye icon to hide it, and then Save your changes.

This powerful snippet does a few things a plain collection loop might miss:

• It reads your narrowest collection first (sort: 'all_products_count'), ensuring a specific grouping beats a broad catch-all.
• It skips generic collections like "Home," "Sale," and "New Arrivals."
• It explicitly skips sold items and products with no image, preventing errors or blank cards.

Beyond the Basics: AI and Advanced Apps

While the manual and code-based solutions are fantastic for deterministic recommendations, some community members also touched on AI. DanielAnderson, for instance, mentioned using AI for analysis and recommendations, but always with human review for crucial decisions. He also highlighted that rules can't cover everything, especially when visitors arrive knowing a specific name, but your product title might not match their exact search query (e.g., searching "Muhammad Ali" for a "Signed 1974 Boxing Glove").

For these more complex, semantic search and recommendation needs, DanielAnderson developed Carti, an AI assistant that searches by meaning and recommends based on understanding. Similarly, Recomento (the app developer who kicked off the original discussion) offers a paid app that recommends strictly by visual similarity or catalog embeddings, meaning it doesn't need prior sales data. These can be powerful tools if you find your structured data still leaves a "residue" of products needing semantic reasoning, as CommerceGov put it.

Wrapping Up

For a low-volume store with unique items, leaning on Shopify's native algorithm simply won't cut it. The community's insights make it clear that a strategic, deterministic approach is key. Start with Shopify's Search & Discovery app for your hero items, and consider implementing the Custom Liquid solution for a robust, scalable collection-based fallback. This way, you curate the pieces that matter most, and the rest of your catalog still gets relevant picks on its own. It's about working smarter with your data, not harder against an algorithm that isn't designed for your specific business model. If you're looking to start your own unique online store, Shopify is a fantastic platform to build it on, offering the flexibility to tailor experiences like these to your specific needs. Remember to track your recommendation clicks and assisted add-to-carts, as Icey.Lane suggested, to see what truly resonates with your customers.