The way shoppers find and choose products has fundamentally shifted and so has the playbook for showing up on the digital shelf. In our latest webinar, Benny Blum (SVP of Product at Syndigo) sat down with Justin Maner (Co-Founder & CTO at DetailPage) to unpack what’s actually working in today’s omnichannel customer experience, from semantic similarity and AI-driven personalization to product images that get indexed by LLMs.
Here are the top takeaways every brand should walk away with.
1. SEO Isn’t Dead. It’s Stacked.
A lot of LinkedIn chatter will tell you keywords are obsolete. The reality is more nuanced. Today’s discovery engines work in layers:
- SEO is about content: learning to talk about your products and the way customers shop for them.
- GEO (Generative Engine Optimization) is about context: matching your product to the persona and situation behind a query.
- AEO (Answer Engine Optimization) is about access: making sure LLMs can actually pull the information they need.
The big mindset shift: these are additives, not alternatives. Semantic similarity is still the foundation, and contextual richness is what wins in the LLM era.
2. LLM Shopping Just Crossed the Trust Threshold
Adobe’s study of over a trillion visits revealed a stunning turnaround. In March 2025, LLM-referred traffic converted 38% worse than non-AI traffic. One year later, it converts 42% better.
Why? Because LLMs use rich context about the shopper to serve only the top one to ten recommendations, not 24 results across endless pages. Shoppers are starting to trust the recommendations, and that trust is reshaping the omnichannel experience across Amazon, Walmart, Shopify, ChatGPT, Gemini, and the newly relaunched Alexa Shopping.
3. Personalization Is the New Battleground (and Amazon Is Showing Its Hand)
Amazon recently published a science paper introducing “comprehensive synthetic personas,” and the company is now aggregating up to 43 dimensions of shopper data including children, hobbies, brand preferences, and product history into a single persona profile. ChatGPT just upgraded its entire memory system. Gemini and Shopify are doing the same.
Translation for brands: ecommerce personalization software and persona-driven content are no longer nice-to-haves. Your PDPs need to speak to who is shopping, not just what they typed.
4. Hit the 15 Contextual Dimensions or Get Left Behind
DetailPage’s team validated that Amazon (and other LLMs) use roughly 15 contextual dimensions to match products to personas: things like used-for events, lifestyle fit, value tier, and more. Most brands are missing obvious ones (think ring pops not mentioning “birthday parties”).
Action step: audit your top SKUs against these dimensions and prioritize the highest-volume gaps. You don’t need to optimize your entire catalog. Apply the 80/20 rule and grade your catalog to find the biggest ROI wins first.
5. The New 75-Character Title Rule Is a Hidden Opportunity
Amazon’s tightened title limit comes paired with a new 125-character “Item Highlights” field that is indexed (and likely weighted above bullet points). The winning structure:
- Title (75 chars): Brand + your single highest-ROI descriptive phrase + 2 to 3 defining features
- Item Highlights (125 chars): Parallel phrases customers use, longer descriptive “B terms” that contain “A terms,” plus salient attributes like size, materials, and dimensions
Brands that adapt fast will leapfrog competitors who treat this as a constraint instead of an opening.
6. Product Images Are Now Traffic Drivers, Not Just Conversion Tools
For years, product images were treated purely as conversion assets. That changed in the last few months. LLMs are now spending real money to index images, which means rich media is doing double duty: driving discovery and sealing the sale.
What a high-performing image looks like today:
- Compliance ready
- Lifestyle imagery in the carousel (still the gold standard)
- Clear feature coverage with text overlays
- LLM and mobile ready with a single-minded message (one idea per image, not 20)
Why single-minded matters: LLMs use cheaper, faster models to index images, so a “dumber AI” needs an obvious signal. A messy image gets lost. A focused image gets surfaced.
7. Reviews Are a Goldmine for Both Discovery and Conversion
This is where review management software and customer feedback software become strategic, not tactical. Reviews tell you exactly why customers buy, in their own words. DetailPage uses that data to:
- Identify the top features customers actually care about (the “laugh-out-loud funny” insight for Taco Cat Goat Cheese Pizza came straight from reviews)
- Map those features against competitor imagery to find gaps
- Build creative briefs and even production-ready images that match the way customers describe products
This is the closed loop in action: a user generated content platform like Syndigo’s PowerReviews feeds the insights that fuel Amazon A+ Content, omnichannel experience optimization, and the next round of image and copy updates. Pair that with social proof marketing tactics and a bit of FOMO marketing (think “1 in 5 customers say…”) and you’ve got messaging that converts.
8. Proof It Works: A Real-World ROI Story
For Breathe Right, DetailPage replaced a technical lead image with a simple “Peel, Place, Press. Done in seconds” visual based on review insights. Results:
- Conversion rate jumped from 12% to 18%
- Revenue lift of $43,000
- Total spend on the image stack: $394
- ROI: nearly 11,000%
That’s not theory. That’s what happens when you align product images, contextual signals, and customer-voiced features.
9. Diagnose Before You Optimize: Traffic Problem or Conversion Problem?
A simple framework for figuring out where to spend your next dollar:
- Share of voice > share of units: You have a conversion problem. Tighten your PDP content, images, and trust signals.
- Share of voice < share of units: You have a traffic problem. Invest in semantic similarity, GEO, and discoverability.
Example from the webinar: Dyson had 8% share of voice but only 3% share of units (conversion problem), while Eureka had 6% share of voice and 10% share of units (traffic problem).
10. DIY, Generalist, or Specialist? Choose Wisely.
You can change your own oil. You don’t perform your own LASIK. Justin’s argument: SEO, GEO, and image optimization at scale sit firmly in the specialist category because of the specialized datasets required, the speed at which the landscape is changing (two major launches happened in the two weeks before the webinar alone), and the fact that most ecommerce teams are under-resourced for catalog-wide execution.
The Bigger Picture: A Closed Loop for the Digital Shelf
This is why Syndigo and DetailPage fit together so well. DetailPage provides the intelligence layer, prioritization, and proof points. Syndigo solutions like Enhanced Content, PowerReviews, and Taggstar deliver the conversion rate optimization, rich media, and review collection that feed the loop. Whether you own SEO and search, PDP content, or the operational side of the digital shelf, there’s a clear next step:
- Own SEO and search? Run a GEO readiness check on signal health and inputs.
- Own PDP content? Close trust gaps by connecting reviews, insights, and how you present your product.
- Own digital shelf operations? Lean into the feedback loop and put your scraped, collected, and reviewed data to work.
Ready to See Where You Stand?
DetailPage is offering a free audit of your SEO, GEO, and image performance. Email syndigo@detailpage.com to find out where your biggest opportunities live and how fast you can act on them.
