0:00 Well, welcome, everyone. Glad you're here. Today, we're going deep on something that a lot of brands know that they need to pay attention to, but maybe haven't had a concrete framework for yet. 0:11 This is how to actually measure and improve your visibility in AI powered search. Over the next session, Marta and I are going to guide you. 0:20 We're gonna give you something you can take back to your team and actually use this week, not just strategic framing, but a real scoring methodology, a measurement playbook, and a direct line from your Bazaarvoice investment to your GEO performance. 0:34 So let's get into it. So quick introduction before we dive in. 0:42 I'm Megan Danielson, VP of global solutions at Bazaarvoice. My role is sitting at the intersection of our platform capabilities and how brands actually put them to work. 0:51 I've spent the last several months building out the measurement framework you're going to see today, and I'll hand it over to Marta to introduce herself. 1:00 Hello. Marta Kelsey here, senior director of global product management here at Hertsenberg Voice, and I oversee our ratings and reviews as well as contextual commerce product lines. 1:10 So excited to be here today. Amazing. Alright. So let me tell you exactly how today is structured so you know what you're getting. 1:20 We have four goals for this session. First, we want you to understand why GEO matters right now, not just as a trend to watch, but as something that's actively affecting your brand's discoverability today. 1:32 Second, we wanna show you how Bazaarvoice is a direct strategic lever for GEO performance. Not because we're Bazaarvoice and this is a Bazaarvoice webinar, although that helps, but because UGC and review data are literally the content layer that LLMs are drawing on to build brand answers. 1:49 Third, we're gonna walk you through a five pillar measurement framework that you can apply to your own brand today. No special tools required, spreadsheet, a browser, and about two to three hours of your time or what it will take. 2:01 And fourth, Marta's gonna connect each of these pillars to specific Bazaarvoice capabilities and road map investments so that you know not just where your gaps are, but what you can do about them. 2:15 Alright. So let's start with what's actually happening. Seventy nine percent of consumers are expected to be using AI enhanced search within the next year. 2:24 That's not a research projection about twenty thirty. That's twelve months from now. And, honestly, this stat is actually a bit old, so that's actually the time we're in right now. 2:33 And here's the thing that matters most for brands. When those shoppers ask an AI assistant or an LLM a question, such what's the best sustainable sneaker or which dog food prevents health issues, the answer they get shapes their consideration set before they ever visit a website, before they see an ad, before they search on Google. 2:52 If your brand is not in that answer, you effectively don't exist for that shopper in that moment. The shift that's worth naming explicitly, this isn't just about being found. 3:03 It's about being described accurately and positively. Traditional SEO was about ranking on a page. GEO is about what the AI says about you when someone asks. 3:13 And, critically, the signals that determine that, your content quality, your schema markup, your review volume, are almost entirely within your control. 3:22 What's important to note as well is that Gen Z and millennials are leading the shift most aggressively, which means that brands most at risk are in categories where those demographics over index, things like beauty, fashion, consumer electronics, pets, and food and beverage. 3:41 The questions consumers ask haven't changed. What's changed is when where they're asking them. Look at the three stages on this slide. Education and discovery. 3:50 Shoppers are asking, what's the best everyday fashionable sneaker? They're no longer typing that into Google, although they may be in getting an AI overview. 3:59 They're also asking this in ChatGPT, Perplexity, Claude. And these tools don't return ten blue links. They return a synthesized answer with a brand recommendation. 4:09 The middle stage, which is personalized product selection, is where it gets really interesting. Which are the best sustainable options asked to an LLM doesn't just return a list. 4:18 It returns a ranked recommendation that reflects what the LLM has ingested. Reviews, q and a, product descriptions, editorial coverage, the brand with the richest, most structured content wins. 4:30 And the validation stage, where they're asking which has the highest rating, is really where review data becomes the digital currency. LLMs are pulling aggregate ratings and review sentiment to answer that question. 4:43 If your reviews are structured and accessible, you show up with specifics. If they're not, you show up vague or not at all. The through line across all three stages is the AI is making decisions your brand used to make for itself in the consideration layer. 4:57 You need to be feeding these AI systems the right content. And UGC is the training data for how LLMs represent your brand. 5:08 When generative engines build an answer about your product, what it's made of, how it fits, whether it's worth the price, how it compares to a competitor, they're drawing from content that exists on the web. 5:18 There are multiple types of content that are important, but they tend to fall into three buckets. The first is the content you are generating about your own brand and products. 5:27 In this content, it's important to have distinct proprietary claims that help the LLMs understand how your brand is differentiated. The second is third party content. 5:37 Who knew that traditional publishing would be so important again? How large and small third party publishers talk about your products helps LLMs to understand your products more in-depth and in relation to competitors. 5:49 And the third bucket is user generated content. Ratings and reviews, questions and answers, and influencer or creator content all serve to validate claims that you are making about your products. 6:00 These are the voice of the consumer that give LLMs the confidence to recommend your brand for specific prompts. To wrap the context section before we get into measurement, in the AI first world, brands need what we're calling triple a UGC. 6:19 The first pillar here is accessible, so content that AI systems can actually find and read. This is what the schema and render accessibility work that we'll cover in one of the pillars for the measurement framework. 6:30 Your review stars, your review text, your q and a, if those only load via JavaScript, LLMs can't see them. Accessible means your content is in the initial HTML response structured with the right markup. 6:42 The second pillar is authentic. So genuine UGC that meets the authenticity bar LLMs are trained to wait. Moderate moderated verified reviews from real customers, high specificity language, the kind of content that LLM cite for credibility questions rather than deflecting to vague generics. 7:02 And the third pillar is abundant. You need a rich volume of detailed specific UGC that gives AI systems deep product insights. 7:10 This is the volume floor we'll talk about in the measurement framework. Brands with fewer than a thousand reviews are capped on their trust and sentiment score because the signal is too thin. 7:20 Review collection programs close that gap. The triple a framework is also the lens through which to look at your Bazaarvoice implementation if you already work with us. 7:29 If your review content is abundant and authentic but not accessible because it's only loading via JavaScript, you're leaving visibility points on the table. 7:38 We'll show you exactly how to measure each of these and where you stand. Alright. 7:46 So now that we've established why GEO measures and where Bazaarvoice and UGC fit, let's get into the actual measurement framework. Five pillars, each one scored, each one pointing to a specific intervention. 7:58 Let me give you a quick lay of the land before we get into the pillars themselves. What you're looking at is a five pillars scoring framework for measuring how well a brand shows up when shop shoppers ask questions in an LLM like Chat, GPT, Gemini, Perplexity, or Claude. 8:13 The reason this exists is that LLM visibility isn't one thing. You can be mentioned frequently but from weak sources. You can have great content, but it's invisible to the crawler. 8:23 You can dominate, share a voice, and still carry negative sentiment. Each of those is a different problem owned by a different team with a different fix. This framework breaks that down. 8:33 We'll walk through all five pillars and then look at how the scores combine and how to benchmark them against competitors. By end, you'll have a clear picture of where the gaps are and what to do about them and where we at Bazaarvoice fit into the picture. 8:46 To be clear, Bazaarvoice will not solve all of your GEO issues, but UGC is a powerful tool in your toolbox that you should be optimizing. 8:54 This framework will expose opportunities beyond just optimizing UGC or your Bazaarvoice program, but we will touch down after each pillar on how exactly Bazaarvoice can impact these things for you and help support your GEO visibility efforts. 9:08 This framework does not require any tooling and can be done just with a browser, a spreadsheet, and some time. If you are using a tool today, that's amazing. 9:16 You're ahead of the curve. Just look for the same metrics that we're highlighting within the tool that you're using. And just for context, we're gonna be using Allbirds as a sample brand to really bring this framework to life. 9:27 Full transparency, Allbirds does not use Bazaarvoice today, but they are a highly visible brand that still has some opportunities for optimization from a GEO perspective. 9:42 Alright. So pillar one is the most visible number, literally. It measures whether your brand appears in LLM responses to commercially relevant prompts, where in the answer you appear, and how much of the conversation your brand owns relative to competitors. 9:57 The measurement is straightforward. You build a set of twenty five prompts across three types, category prompts like best sustainable sneakers, comparison prompts like allbirds versus veija, and intent prompts like most comfortable shoes for standing all day. 10:12 You run those prompts signed out of each LLM, chat, GPT, Claude, Gemini, and Perplexity three times each to account for nondeterminism, and you log where you showed, in what position, and alongside which competitors. 10:25 The scoring looks at three things. Did you appear at all? This is your presence rate. Where in the answer were you? This is your position score. And what share of the total brand mentions across all your responses did you own? 10:37 That's your share of voice. The key thing to understand is that this pillar is the most market dependent. You're competing against everyone else in your category for the same limited real estate in an LLM response. 10:48 And Bazaarvoice's contribution here is indirect but real. Every product page enriched with syndicated reviews, every q and a answered by the brand, every piece of UGC that LLMs crawl on retail partner sites adds to the body of content that mentions your brand in context. 11:04 Syndication to retail partners puts your brand into more LLM crawlable pages, and volume and surface area of your brand mentions across the web really drives pillar one over time. 11:17 Here's what that an here's what this actually looks like in practice. So Allbirds was tested across twenty five prompts. They appeared in sixteen of them, a sixty four percent presence rate. 11:28 When they appeared, they were showing up at an average position of three point two in the answer, and they owned eighteen percent share of voice behind Nike at twenty four percent and ADDOS at twenty two percent. 11:39 Their composite score is a sixty seven, which puts them solidly in the competitive band. They're definitely in the consideration set, but they may be not winning every time. 11:49 The heat map on the slide tells you the story more granularly. They're dominant on sustainability framed prompts, but they're falling off the map for general comfort and everyday sneaker queries. 11:59 And that's really a content gap, not necessarily a technical problem. The fix for them could be editorial coverage in those adjacent prompt categories. 12:07 The practical takeaway for you, this measurement is something you can run yourself with a spreadsheet and a couple of hours. The prompt templates are something that we can share after this webinar, and the value isn't in the number. 12:19 It's in the heat map that shows exactly which queries you're missing. So now, this is where I get to jump in and show you how you can leverage Bazaarvoice to optimize for GEO visibility. 12:36 After each section that Megan's gonna go through each pillar, I'll come in with the slides that have the icon at the top called optimize with Bazaarvoice to walk you through how to use Bazaarvoice to achieve and to optimize in that pillar. 12:52 But I will also have a slide at the very end that aggregates this all, called the screenshot slide for you to screenshot and take home and help you set your priorities. 13:04 Okay. So this first one, issue isn't that you need more reviews. It's that your reviews might be thin and might be causing challenges with getting that visibility up within LLMs. 13:16 So for example, if you're trying to answer and LM's trying to answer what's the best sneaker for someone who works on their feet all day, It needs more than just a five star rating or review descriptions that say things like great shoe. 13:29 It needs that context. And so this is where you can really lean into an old but true technique around improving your secondary review questions and ratings of your review submission forms. 13:42 So asking consumers things like, how long have you owned this? How often do you use it? Where are you located? 13:50 The reviewers answer those questions and suddenly, you're generating content that's rich with the kind of language, the actual phrases that match the real consumer queries in AI search. 14:02 Same thing with the secondary ratings, fit, value, performance, quality. These aren't just visual nice to haves. They're really signals to the LLMs to understand what a product what the product is good at. 14:16 So when your review content reflects that, you really close that gap between what shoppers are asking and what AI can find on your pages. 14:24 So just this is super straightforward. Make sure your review submission forms are configured to collect this data. It's pretty low effort, but the payoff really compounds over time as your review dataset gets richer. 14:40 And the second one here is is that it's important to ensure your content shows up, not just on your own site, but across the surfaces where AI is making those recommendations. 14:52 So here's three surfaces that are be that are owned themselves by Google. In the first shot on the left here, someone's asking what's the best face sunscreen for oily skin? 15:05 The AI response is pulling in content from product reviews as well as dermatology sources. And the results are basically showing AI based product discovery. 15:16 Your brand either appears in that answer or it doesn't. So the key to unlock your product reviews into Google's ecosystem of AI product discovery is Bazaarvoice syndication. 15:29 So when your views flow through our Google syndication integration, they will become eligible to appear in Gemini, AI overviews, AI mode, Google Shopping, all of the things. 15:43 So these are not separate features that you need to build into. It's really one activation through Bazaarvoice. So our clients today can opt activate this today via simple opt in form. 15:56 I will say most clients do this today, but if you aren't, this is an easy way to start showing up immediately. 16:05 And this really matters because Google AI surfaces are where a massive chunk of consumer discovery is happening right now. If all of your product review content isn't accessible to those Google systems, you're essentially going to be invisible in those moments. 16:25 Alright. So pillar two is where it starts to get interesting for us. Two brands can have the exact same pillar one score showing up in the same number of prompts, but one is being described by The New York Times and Bloomberg, while the other is being described by a Reddit thread from twenty twenty one. 16:42 We all know those ones. Those are not equivalent positions, and LLMs do weight source quality even though we all know that they do love Reddit, at least for now. 16:52 Citation strength measures four things. It's the tier quality of your sources, how recently those sources were published, how many distinct domains they come from, and this is the one that's easy to miss, how many times your brand is actually named within each cited piece. 17:08 The last one is what we call mention density, and it matters because a piece that mentions all birds eight times is doing much more reputational work than a piece that mentions them once in passing. 17:19 The measurement uses the same twenty five prompts from pillar one. You capture every URL the LLM site, tier each one using the six tier taxonomy on the left that you'll also see more in the next slide, flag anything older than two years, and count unique domains. 17:34 And for each piece, you can log how many times your brand name appears in the body text. One important nuance, when you look at step four, we've updated that this to explicitly include brand mention counting because it's easy to collect at the same time, and it feeds the scoring directly. 17:51 Bazaarvoice actually shapes three of the four citation inputs. The reviews and q and a content on your brand dot com become brand owned citations. 17:59 Syndicated reviews on retailer pages become retailer citations. Aggregate or review content on platforms like Bazaarvoice owned Influenster, where we can pull or push review volume from, becomes another citation source. 18:13 And because reviews naturally mention the brand name multiple times, Bazaarvoice powered content tends to score well on mention density. Bazaarvoice also drives fresh review content continuously, which helps the recency score it keeps your recency score healthy. 18:32 For all birds, forty seven unique citations were captured across those twenty five prompts. You can see the tier breakdown, seventeen percent from major editorial, twenty three percent from trade and category press, and then nineteen percent that's brand owned, and twenty one percent long tail. 18:49 The brand owned piece is a positive signal. It means Allbirds is being ingested by LLMs, which is a structured readiness win, which bleeding into the citation strength. 18:59 The drag is that twenty percent long tail bucket. Notice the four stat boxes at the right, forty seven total citations, seventy percent of them published in the last twelve months, thirty one unique domains, and importantly, two point four average brand mentions per piece. 19:15 That two point four puts them in the two to three range, which scores which scores a seventy on mention density. If they drive more tier one placements that are specifically about allbirds rather than just mentioning them in passing, that number moves up. 19:29 Their overall score here is sixty five. The fix really is targeted PR, at least at at the surface, not a spray of press releases, but six to ten deliberate tier one and tier two placements that specifically cover all birds, not just mention them. 19:43 One important thing to note is that these citations will shift frequently. We've seen YouTube making more and more of a presence in citations for a lot of our brands and retailers recently. 19:54 This is largely on Gemini and likely due to the close relationship between Gemini and YouTube. However, it speaks to the increasing importance of video and visual content as a citation source. 20:05 Just because tech sources dominate today doesn't mean you shouldn't be looking at video as an important citation source for the future. So to reiterate, citation strength is about diversity and consensus as much as it is about volume of the opinion and data. 20:25 So for example, ten citations from your own site is weaker than ten citations from ten different authoritative sites. 20:34 So this is where the Bazaarvoice network becomes a real benefit to you. So we're talking about the network that makes up of more than thirteen thousand brand and retailer sites for reviews to be displayed at. 20:46 So when product reviews syndicate across the Bazaarvoice network, not just getting more reviews to consumers, you're building that cross domain consensus about your product quality. 20:57 And as Megan mentioned, that's exactly what LLMs are looking for. So when retail sites as well as your own PDP have all consistent high quality review content saying similar things about your product, that's the trust signal AI systems weight heavily. 21:14 The specific action here is to make sure you have Bazaarvoice syndication turned on if you do not. But also to go in constantly or regularly, I should say, review your syndication coverage insights within the portal. 21:31 You may find that some product matching gaps exist between your catalogs as well as retailers. And we released a couple months ago, the ability and tools for you to go suggest matches to optimize that review, display and volume across the retail product or surfaces. 21:50 Excuse me. So definitely be leveraging the reporting as well as the self-service matching tools to increase this citation. 22:02 And another angle on the citation string is to ensure you're proactively generating content on products without reviews. 22:11 So this is where product sampling comes in and we have a few ways to help you do that. First is Influenster. Megan mentioned this review aggregation site, but it's more than that. 22:22 It's a ten million plus community that allows you to segment by audience to go sample products and drive reviews from as well. 22:31 So this is great for getting reviews from the right type of consumer, not just any consumer. Think about trying to drive reviews from a certain segment. 22:39 We use the example often of hair care products and trying to get more reviews from people with curly hair. That's where this can come in. 22:48 And then those reviews are displayed on the Influenster aggregation site, which helps with the citation string. We also support creative creator partnerships, nano mid tier creators with engaged followings. 23:03 So think about this as your long tail kind of editorial presence. When a creator writes about your product and that content lives on their site or social, you can get picked up as a citation. 23:15 And this is where to reference what Megan was saying earlier, the YouTube videos can really come into play here. We can sampling and create a partnership support those video call to actions. 23:28 And finally, loyalty communities. Seating your own customer base is really underrated and you make sure that you're leveraging them to drive your review content. 23:40 So the goal here is to make sure that you're running a steady ongoing cadence across these sampling and creator and ambassador programs, that you're really running this as a program and not a one off campaign. 24:02 Alright. We're on to pillar three. And this really shifts from discovery to conversion. You might be showing up. Pillar one is fine. Pillar two is decent. 24:11 But when a shopper asks a specific question about your brand, is the LLM getting it right? This pillar asks, when an LLM is asked something specific, does it answer at all? 24:21 Does it cite a credible source? And is the answer specific and accurate rather than vague and generic? The question set for Allbirds covers five categories, product specs, return and shipping policies, quality and comfort, sustainability claims, and brand comparison. 24:37 The categories you choose for your own brand may be the same, or you may some variants depending on your unique claims in the market. Each one is designed to test a different layer of the brand's content ecosystem. 24:49 One important note that I wanna flag before we look at the example, hallucinations are treated as a zero, worse than no answer. A confidently wrong answer about your return window or your carbon footprint is more damaging to a purchase decision than the LLM saying, I don't have that information. 25:05 So hallucination rate is tracked separately and should be watched closely. So Allbird scores well here, eighty two overall. 25:15 They have a ninety percent answer rate, strong source quality, and good specificity. But the pattern in the breakdown is worth pausing on. Product specs and policies are almost perfect. 25:25 Sizing and fit are solid. The gap is really sustainability claims. That category score is a sixty six, dragged down by both dragged down by both source quality and specificity. 25:37 What that tells you is that when a shopper asks an LLM about Allbirds' carbon footprint or supply chain, the LLM is either going to a weaker source or giving a vague generic answer. 25:48 Depending on the specific question, the fix isn't necessarily PR. It could be brand owned content. Allbirds may need more detailed structured sustainability FAQs or an answer page on their own site that LLMs can pull from directly. 26:03 And that content probably needs to be structured properly, which brings us right into pillar four. This is the clearest example of how the pillars connect. 26:11 A pillar three gap in one category is almost always pointing to either a pillar four problem, the content exists but isn't crawlable, or a content gap, the answer page doesn't exist yet, for example. 26:25 And this is a bonus layer on top of pillar three that I just wanna spend a minute on because it's one of the more practically useful outputs. 26:33 The premise is simple. Tier alone doesn't tell the whole story. A brand owned citation for what's your return window is actually ideal. 26:42 A brand owned citation for is this the most sustainable shoe on the market could be a red flag. You're essentially citing yourself for a credibility question. 26:51 The matrix the matrix maps question type against source tier and flags whether the LLMs are pulling from the right kind of source. 26:59 When you see a column of risky or bad answers for a question type, like Allbird sustainability claims, that tells you there's a content gap somewhere. 27:07 Either the right source doesn't exist, it's not structured for LLM ingestion, or it's just not been authoritatively published yet. This is the most direct bridge between measurement and your content and PR road map. 27:19 It tells you not just that something is wrong, but what kind of intervention will fix it. For example, look at the quality and comfort row. 27:29 Under t three, which is retailer, it's actually ideal for subjective experience questions. That's really Bazaarvoice territory. When a brand has high review volume with specific credible review bodies, LLMs use that review content for exactly the questions it's best suited for. 27:48 Now let's talk about how you can optimize with Bazaarvoice here. One of the best kept secrets of GE right now is this product q and a. 27:57 Structured question and answer content on your product pages does something really specific. It teaches the LLM how to answer questions about your product in your voice from your site. 28:09 So think about it from the LLM's perspective. If someone asks if the if a moisturizer is fragrance free, the only source and the only sources from a blog post from about two years ago that vaguely mentions it, the AI might get it wrong or hedge or just not answer. 28:28 So if your PDP has a q section that directly answers that exact same question with a verified response, that's the source that LLM wants to use. 28:40 So this matters for really three things in scoring what Megan talked about. Your answer specifically goes up, your source fit gets better because brand known as the ideal source for product specs and policies, and the hallucination risk drops significantly. 28:58 If the right answer exists on your page in a readable format, the AI doesn't have to guess. The common objection we hear from our clients when they're hesitant in leveraging q and a as a strategy is that it can be hard to scale, hard to answer all those questions with humans. 29:19 And so we've developed the answer assistant to really help humans respond more quickly. So it lets you respond to your consumer questions effectively at scale with every response not being a one off manual effort. 29:33 So the brand really gets to shape that canonical answer. So this has been in pilot for a while and will be available to all q and a clients here any day. 29:44 It's really right on the brink of releasing. So we're excited to get this out into the Bazaarvoice portal for everyone to help them scale their q and a. 29:56 On the review side, the quality of each individual review matters as much as volume. So now I'm trying to answer a really specific question, how does the serum wear underneath makeup or is this good for sensitive skin? 30:11 It would need reviews that actually talk about those things. As I mentioned before, reviews that are generic that say, quote, I love this product that are thin, they don't really help. 30:22 And so we developed content coach that helps solve this by prompting reviewers to write more detailed content. This has been out for a while and has pretty widespread adoption. 30:34 But if you've not turned this on, this is definitely a quick win. It helps drive things like texture, fragrance, cost value, ingredients, ask the the review writer how long they've been using it. 30:47 And that specifically feeds directly into LM answer quality. Because now when someone asks those precise questions, there's a body of review content to draw from. 31:01 So your content needs to match that. One other thing I'll say on content coach is it's been out for a while is we're continuing to invest in this by expanding it to more languages to be available on more country websites. 31:20 All right. Pillar four is the one that surprises people the most and the one with the highest leverage. It measures two things. 31:29 The first is whether your pages are properly tagged with schema dot org markup so LLMs understand what they're looking at and whether your critical content is available in the initial HTML response without requiring JavaScript to render. 31:42 The second part is the the one most brands haven't thought about. Most LLMs can't execute JavaScript at crawl time. Gemini is the one consistent exception to this. 31:52 So if your review stars, your review text, or your product prices only appear after a JS bundle loads, they're actually invisible to the LLM. 32:01 You can have perfect schema and still be invisible if the content is JS dependent. The scoring for this pillar has three components, forty points for foundational schema, so product offer, organization, breadcrumb list, etcetera, thirty points for review and q and a schema, which is aggregate rating, review body, author attribution, date published, and question and answer markup, and thirty points for render accessibility across your three main page types. 32:28 That thirty point review and q and a component is significant, and I wanna be direct. That's exactly what Bazaarvoice provides when properly implemented. 32:36 The how to measure section walks you through running this audit yourself. It's using a simple schema validator and dev tools, roughly about an hour per page type template. 32:51 Allbird scores sixty here, which sounds okay. But when you see the breakdown, it's clear that there are three distinct fixable gaps. The foundational schema on PDP and home page is in place, but breadcrumb list is missing on the category template. 33:05 That's a pretty quick engineering ticket, maybe a day of work. In the review and q and a block, aggregate rating and review body are present and valid, but author attribution is missing and q and a markup is missing entirely. 33:17 And I think they just don't have q and a on their site right now. Those are fifteen points gone right there. The most significant gap is render accessibility. 33:26 The category page is essentially invisible to LLMs. Products require JavaScript to load. So from an LLM's perspective, the page is empty. 33:34 The PDP is partial. Core product info is server rendered, but reviews load by a JS, so the LLM sees the product but not the social proof. 33:43 The home page looks good. The render fix is the longest cycle item. It's likely a server side rendering project, but it's also the highest upside because fixing fixing it amplifies every schema improvement already in place. 33:55 A brand that fixes the JS rendering gap and adds the missing q and a markup could move this score from sixty to close to ninety. As I mentioned earlier, Allbirds does not use Bazaarvoice today for their reviews program. 34:07 But if you're listening, we'd love to chat. But if they were, we would recommend that they sync with their account team to upgrade to some of our latest and greatest display options in order to overcome JavaScript rendering issues that I just mentioned. 34:23 So structuring review content has been something we've been spending a lot of time on the product development team with clients on lately. 34:31 And as Megan mentioned, it's been surprising clients a little bit because it's not just about marketing and getting your review content out there. 34:40 This one's about making sure your infrastructure is set up to support the proper crawling. Structuring user generated content, ratings and reviews, questions and answers, etcetera on your site is entirely in your control. 34:55 So here's the problem with solving. As Megan mentioned, most review content on product page loads via JavaScript, which is fine for human visitors. 35:04 The page renders, you see the reviews. But many AI crawlers don't execute JavaScript at crawl time. We believe this will change over time, but that is how it is today. 35:16 So a bot hits your page and the reviews, they just aren't there. The authentic discovery API that we've launched recently really solves this by serving your review contents in that server side HTML layer. 35:31 So the first thing a crawler sees when it hits your page is the actual review content, also including q and a and review summaries. 35:39 It's structured, it's readable, and it ingests into those crawlers. So this API, it's patent trademark pending and we're excited to have released this recently. 35:52 It is available today to our enterprise tiered clients. We are working to release this out to our essentials and advanced tiered clients soon as well. 36:03 In addition, we're solving a native solution for clients that leverage Bazaarvoice Shopify's plugin. And our Shopify solution will work natively within the platform with no additional development effort. 36:17 So clients who are looking to implement the authentic discovery API today, they can expect to take about two to four weeks to implement it. 36:26 There's some clear instructions to follow, but overall, it's not too much of a heavy implementation left to get them crawled right. 36:38 And then structured readiness, this one just builds on it. It's pretty simple. LLMs love summaries. Review summaries do the synthesis work for the LLMs. 36:51 So we've had review summaries out for for years now. It's been available. We've got pretty wide adoption on it. But think about this, instead of an LLM having to read all those five hundred reviews, which it will do to make sure answers right, We can also provide it a little earlier cues to provide really the the synthesis of what are the product strengths, weaknesses, and key themes. 37:15 So this reduces some of that work for the AI. So when you implement the authentic discovery API, review summaries are included as well as, as I mentioned, the the ratings, the reviews, as well as the q and a content. 37:32 So we continue to invest in review summaries. We've added a number of new languages to this product over the last several months, and we are continuing to add languages going forward. 37:46 Alright. Onto the final pillar. So pillars one through four measure exposure and accessibility. Pillar five measures whether that exposure is helping you or hurting you. 37:57 The core insight is that LLMs don't just surface your brand. They amplify the signal that's already out there. High visibility with negative sentiment is actually worse than low visibility with neutral sentiment. 38:09 This pillar has two streams. The first is editorial sentiment. You could take all the tier one and tier two sources from your pillar two audit, classify each brand mentioned as positive, neutral, or negative, and calculate net sentiment. 38:23 The second pillar is review sentiment, your aggregate rating from first party review data weighted by volume. If you have fewer than a thousand reviews, the score is capped at seventy five because you don't have enough signal yet. 38:36 The way these two combine is important. Editorial sentiment is a leading indicator for how LLMs will frame your brand and opinion based answers, and review sentiment is what shows up when an LLM is answering a question about customer experience or product quality. 38:50 Both matter, and they point to different interventions. So Allbird scores a seventy one here. 38:59 The owned channel data is very healthy. Four point two stars across fifty one thousand verified reviews. That's a really strong review base. 39:08 The editorial sentiment is what is what pull is what is pulling the composite down. So look at the bar chart. Sustainability and product coverage is strongly positive, which you'd expect for Allbirds, but there are twelve negative editorial mentions in the finance and business category and three in store closures and operations. 39:27 That's really layoff coverage, declining sales coverage, and stores store closure closures coverage. It's not product criticism, and it's not experience criticism, but LLMs don't make that distinction when they're summarizing what's being said about Allbirds. 39:42 The strategic implication is clear. The PR team at Allbirds should be actively counterweighting financial coverage with product sustainability and customer storytelling. 39:52 You can't control the business news cycle, but you can make the positive signal louder. Those fifty one thousand reviews with a four point two average star rating are the single most powerful asset in Allbirds LLM trust profile. 40:06 When an LLM is asked, are Allbirds comfortable, or what do people think of Allbirds? That set of reviews is the ground truth it's drawing from. 40:15 Control what you can control and ensure that reviews are part of your sentiment narrative for LLMs. I'm gonna dig into the trust piece a little bit here, specifically around user generated content. 40:30 Bazaarvoice has always put trust at the forefront of our products, forefront of our company, and as the product continues to evolve to support LLM product discovery, the same thing holds true. 40:43 So here's a nuance that matters. LLMs can look for signals to understand whether review content is authentic or trustworthy. 40:52 They may look for signals within the text to determine if there's fake reviews, maybe the contents thin as I mentioned before, that's not really relevant. 41:03 These are patterns that the AI systems are learning to discount the content over time. So trust signals aren't just for human shoppers anymore. 41:12 So there's two kind of things you can do within Bazaarvoice to optimize for this. One is enable verified purchaser badging and this does something specific. 41:22 It confirms that the reviewer actually bought the product directly from your site. Pretty straightforward, pretty clear direct signal to the LLM to trust the review. 41:33 The second thing is the Bazaarvoice intelligent trust mark. And this goes a layer deeper. It provides a machine readable tag that tells LLMs your reviews have been authenticated. 41:46 Think of it as giving AI systems a direct trust signal for your reviews overall. Instead of it having to infer trust from those indirect signals I mentioned within the text itself, it's explicitly telling the crawler this content's real, it's verified, it's safe to cite. 42:02 And it's reducing that friction in the citation decision. And here's the last recommendation I have. 42:11 And I wanna close on this because it kinda ties the whole strategy together. LLMs have a freshness preference. So as we mentioned, if you have five thousand reviews but most are from twenty twenty one, you're at a disadvantage compared to maybe a competitor with a thousand reviews in the last six months. 42:30 So recency is a ranking signal inside AI systems just like it was in traditional SEO. So sentiment's important, recency is just as important. 42:41 So this is where that sampling strategy shouldn't be random or just volume focused. It can really be data driven with two tools within Bazaarvoice. 42:50 So we are we have our existing sentiment reporting metrics within the portal. We've also recently released these new recency metrics and opportunities and that's what's displayed on this the screen right here. 43:03 This is to help you identify and focus your review collection. So take advantage of these new of the new recency metrics as well as our existing sentiment metrics to run those sampling campaigns to target specific products where freshness or maybe positivity have eroded. 43:25 Alright. Now that we've been through all five pillars, let me pull the lens back. Scores on a zero to a hundred scale map to five bands, critical gap, below par, competitive, leading, and best in class. 43:40 Allbirds sits at sixty nine, solidly in the competitive band. They're in the consideration set, but not necessarily winning every prompt. So two important callouts here. 43:50 First, pillar four is high control. Both schema and render accessibility are completely within the brand's own hands. No PR cycles required. 43:58 Any brand should expect to reach seventy five plus on pillar four within a quarter of focused work. A score below sixty on pillar four is a structural failure, not a competitive position. 44:09 Second, absolute scores only tell you so much. A sixty nine is competitive in the abstract, but the category leader is a seventy six, then you're losing. 44:18 The brand tells you or the band tells you the floor. Benchmarking tells you the ceiling. So now this is where measurement really becomes strategy. 44:31 This table shows all birds alongside four major competitors across all five pillars. Looking at all birds in Vesia, they have similar composites, sixty nine versus sixty six, but really completely different profiles. 44:44 Allbirds wins answer quality. Veja wins visibility and sentiment. Same headline number, totally different strategic situation. 44:53 Now look at ADDOS. Seventy six composite, the category leader. How? A perfect one hundred on structured readiness. 45:01 They've solved the schema problem completely. That one pillar at twenty five percent weight is doing enormous work in their composite. And here's the direct implication. 45:10 If Allbirds closes their pillar four gap from sixty towards ninety, fixing the JS rendering and adding q and a markup and probably q and a to their site in general, their composite moves to roughly seventy five or seventy six, so pass beja up to ADIDAS' level, that single work stream without touching any other pillar. 45:30 Vesey is really the cautionary tale. Fifty three composite, critical gaps in citation streak and visibility. They're effectively invisible to LLMs for most category prompts. 45:41 This framework isn't a onetime audit. It's a quarterly operating rhythm. Pillars four and three move fast, weeks to a quarter. Pillars one and two move slower. 45:49 Editorial cycles take time. And pillar five is continuous. Know which lever you're pulling and what the expected timeline is. I think your last slide oh, there it goes. 46:07 Here it goes. Yes. Okay. So let's wrap up on just the view overall. This slide really summarize and aggregates all the ways you can optimize GEO with Bazaarvoice today. 46:20 So this is the screenshot slide. Please take advantage of the opportunities and prioritize them on your roadmaps. I'll say our knowledge base and developer portal all have these documented for ease for you to configure and implement. 46:37 And as always, reach out to us with any questions you have. So thanks, Megan, for joining today and appreciate it. 46:45 Yes. Thanks, everyone. Thank you.
