📎Because V1 LLM Beta Update July 7, 2025

What We’ve Learned + What We’re Trying Next

Hi Because LLM Beta Users,

Quick update on what we’ve learned — and where we’re going next.

TL;DR:

1. Subdomains not working as expected (won’t get traffic fast enough)

2. Need to switch to primary domain content

3. Two experiments:

   1. “View as Markdown” button on PDP

   2. Prompt-optimized blog content

   3. More to come…

Why we chose subdomains & why it’s not gathering data fast enough

As part of the beta, we started by testing markdown product pages hosted on a subdomain using llms.txt files. Our original plan was to host the markdowns directly on your primary domain, but due to Shopify limitations, we had to adjust. Shopify only allows a llms.txt file to be uploaded as a "custom page" — but this functions more like a file upload, redirecting the main domain to a file URL. We don’t believe LLMs reliably follow these redirects when crawling, so we held off on that approach.

However, after 11 days of testing this setup on a subdomain with our first 11 beta users, we’ve seen virtually zero views on those pages outside of our own internal traffic.

While the content appears to be correctly structured and technically crawlable, our current hypothesis is that subdomains are not inheriting SEO authority from the main domain — which could be limiting how LLMs crawl and prioritize the content. While we can’t say that definitively yet, it’s a strong enough signal for us to try a different approach while continuing to let this one run in the background. (It’s possible the subdomain content may still gain traction over a longer timeframe like months but we’re searching for quick wins for you).

---

In the spirit of looking for quick high impact wins, we are proceeding with a few other experiments. You can feel free to participate in either or both:

Two new experiments we're kicking off

1. “View as Markdown” Button on Your Product Page

Instructions: You can find snippet and setup instructions here.

Please help us increase your LLM visibility by adding a small “View as Markdown” button to your 5 selected product pages that links directly from the PDP to its markdown version. This is quickly becoming a standard in industries like SaaS help docs. Since these markdown pages will now be accessible from your primary domain, we believe they’ll be easier for LLMs to crawl and index over time.

2. Prompt-Optimized Blog Content Featuring Your Products

We’re excited to be expanding beyond product page content faster than expected in this beta! Similar to the dynamic versions of your product markdowns that are optimized for different prompts, we’re also interested in experimenting with generating blog posts designed to directly answer high-intent prompts shoppers are asking LLMs (e.g., “best men’s skincare for dry skin” or “gifts for new dads”). Each post will feature one of your products as the recommended solution to that prompt.

To start, we’re going to be generating 2 blog posts per week for you. If this approach proves valuable, we’ll scale with automation to support more of your catalog and build features for you to auto-create and auto-optimize blog content over time.

Ping us when you're ready for your blog content and we'll send over for review. Once approved, you’ll just publish them to your blog as normal! Let us know if you need help with image generation as well.

---

Closing Thoughts

Bear with us - will take time to experiment

We believe LLM discoverability is one of the biggest untapped growth opportunities in ecommerce today. But the truth is: it’s still a black box. No one has nailed a repeatable solution yet — especially not for ecommerce catalogs — and your participation is helping shape what that solution could look like.

Based on what we’ve seen so far and how long we expect it to take to drive results, we’re shifting to a longer time horizon for our beta to run for 2 months instead of 1. There’s no additional cost to you. Our shared goal is to experiment, gather learnings, and lay the groundwork for a channel that could eventually become a major traffic driver, brand differentiator, and revenue opportunity.

You can reference all of the experiments we'll be running throughout the 2 months here. Please feel free to participate in as many of the experiments as you would like (we'll be running 1-2 per week). We promise to try to keep them as low lift as possible for you.

We’re so grateful to have you on board. Thanks again for leaning in and helping us build this together.

If you have any issues or questions, feel free to reach out to the Because team via slack, chat, or email [e-mail]. Feel free to book time with us here to walk through the set up of these next experiments.