Google Image Display Within Web Results Ep. 573

Hi, and welcome to the SEO podcast, Unknown Secrets of Internet Marketing. My name is Chris Burris, one of the owners of EWR Digital. And my name is Matt Berks from the Lead Strategist at EWR Digital. Oh, that's an improper pacing. EWR Digital. Welcome back to another fun-filled edition of our podcast. This is podcast 575. That's a lot of podcasts. We're excited. We're going to be covering an article about the Google Image Pack. Like that's pretty, we haven't done one on this before. Well, I can just tell you there's been a lot of activity in that area. So I'm sure that there's a lot of interested SEOs out there that want to know what's going on with Google and images and videos. Yeah, well, and it makes a lot of sense, right? So we know that the trend is going to video. We know that the trend, like people just are, we're visual people. We did do a podcast. I don't know if it was the last one or the one before where we were just talking about how images are so very valuable and why images are valuable. Because our brain processes them faster. Yeah. Yeah. The phrase, an image is worth a thousand words. Oh, I like that, Chris. I think that's... Quote Chris Burris, scheme of markup. Hashtag, hashtag said before. This article that we're going to be covering is by Jean-Christophe Chouinard. I don't know. Hopefully I did that pretty well. JC. And it's how Google displays images within web search results. Yeah. JC Chouinard. That's actually how he has his name at the top of the article. Before we get into the article, we have a review that we need to read. This is from Bakar and he says, it is, of course, five stars. Oh, my voice can't go that hard. He says, fantastic podcast. That is with the appropriate exclamation mark. Very informational. Did a consult and it was very educational as well. Thank you so much. Patif, to you, Bakar. That's a nice... Yeah, no, he's... So we're

offering audits, we're offering workshops, and we're offering paid consulting. Did some paid consulting for a few agencies the last two weeks. Been really busy, so I know we've missed a podcast or two, but lots been going on out there and we've had a lot of new stuff come in. And so sorry about that, but we're going to try to stay consistent and we're going to maybe record more of these to make sure that we can give you your daily dose. Literally, we have a plan to have like a recording day and knock out like four of these. Yeah. And then we haven't had a time when we can do that. So that's one piece. And as soon as we record them, we'll stay on track, right? Like we'll keep doing new ones and just have those in the can. And anyway... Chris is starting to fund, so... I did go to an event, Fund Launch Live. It was a spectacular event, I got to say. Really well done. You know, when they think... When they talk about getting into a business, go into a higher-end business. And there's... I don't... There may be, but is there a higher-end business than starting funds? I mean, people, they were giving out awards for people who, you know, raised the capital for their $10 million fund, who raised the capital for their $100 million fund. Sell money. That's the best thing. Yeah. Yeah. Yeah. For sure. I'm going to Mexico tomorrow. That's cool. Yeah. Where are you going? I'm going to a YEC event. Oh, YEC event. Okay, excellent. Doing some network down in Mexico. Yeah. But not starting a fund. Hashtag not starting a fund. All right, let's get into this article with JC. Have you ever wondered how Google decides whether or not they should or even where to show images and web search results? Maybe you have. I don't know. And they're combining videos and images and text. Yeah. Yeah. Let's cover... Are you seeing the screen blink? I see it blink. Oh, it's doing

some shifting. I don't know. That's interesting. Let's cover how Google ranks. What is often called the image pack. According to the specifications of the Google patent title, this is the title of that Google patent, image display within web search results by Grant Dasher and David R.H. Miller. Let's jump into this. In order to know if it's going to show an image pack or what kind of size it is, Google needs to identify if, and there's three ifs. One, if the query has image-seeking intent, are there enough high-quality images to satisfy the user? And do users often search for the query in Google image search engine? So let me sum this up. Google wants you to click stuff. So is the image clickable? Well, yeah, I mean, it's a couple of things, and I'm going to be facetious here. We could end with, like, we could end this article right now because those are the key points. And we could get cut out here. Like, we're going retro. It seems to be getting worse and worse. Well, you know, I've been really running this computer hard, and so that's probably my graphics card. You think so? Probably. Would it be worth... Oh, no, I think it's style. We'll just keep going. If you're on YouTube, please don't have a seizure, and we apologize. After it understands these things, it will define the size of the image display module and then rank the module against regular search results based on three more things. Ooh. Query intent score, normalization score, and image quality score, right? Yeah, so like our video right here, if this was a video quality score, bad. It would be a low... So y'all that are listening, and I know most of you are, you're not missing anything. Yeah. This time on the video, well, I mean, you're missing a light show. Our old videos, like go back to like 200, okay? Podcast number 200. Those were awesome, and there was props, and there was real intent on videos, and we've started to release some

of them on shorts, and they've gotten some good responses. People were like, you look really young. Like, I was like, you don't have gray hair. All right, so when and where images are shown in chirps. This is a little technical. Remember, they're pulling this out of the actual Google patent. The image display environment and position relative to general search results varies in proportion to the inferred user intent in images and the quality of the images that are responsive to the query. All right, that's a lot. We'll be breaking that down through the rest of this podcast. If many high-quality images are available, but there is little inferred interest in images, then no images would be shown. Conversely, if there are a few high-quality images available, not too many, right? But there is a significant interest in images, then some image results will be shown. Well, I think Google's rolling out. Basically, if you're not on the first page of Google, if you're not at the top of the first page of Google, you're not getting any clicks. You don't matter. You don't matter. I mean, there's the top 100 search results. I mean, there's billions of searches or terms or callbacks for different things. I think that they're going to roll out images to hopefully get people another chance to actually get some traffic, right? So, and they're able to, on mobile, show these images side-by-side. So, instead of like one text callback, it might be two images, and you can process images faster so you can see more. So, they're actually trying to do it to help, but I don't know if a lot of people are positioned for this rollout, so. Yeah. Well, this is to help you position for this rollout, so you know what's driving. That is why we're doing this protocol for you, yes. How the image-seeking intent score is determined, right? So, intent score is determined by the image display trigger module, right? This is, again, based on the patent, based on the property, search property ratio, and image

CTR for a query. All right. Here are the factors. This is going to get a little technical. We'll kind of summarize it. Here are the factors that influence the image-seeking intent score. So, are they actually seeking an image? Number one, overall quality. Again, our podcast, if you're watching the video. Well, no, I mean. It's got a, yeah. It's got a. Yeah. We'll just do a WebP images. They're good. Yep. Yeah. Ratio of images to web pages available for a particular search query. You should do ratios on images so you don't have the layout shift for Google. Oh, no. Okay. So, this is that you're talking about. You need to define the image so you don't have image shift. Yeah. What is this? Ratios is saying, if I do a search, let's say, I don't know, we'll say tequilas in Mexico, right? Because you're going to Mexico. If there are a ton of images, right? Of tequilas that are identified as tequilas in Mexico and there are no articles, then that's going to shift the image-seeking intent score towards the images, even without the user doing anything. I like that. Because there's just a lot more images to show. So, obviously, that's probably- I would rather see images of tequila than text. Than read text about how it's made from the blue agave. Thank you, Google. All right. So, next. Click-through rates of images available for a query, right? It's past data. Basically, there's an audition period. Ratio of number of searches for a query in the image corpus, so that's in Google image, compared to the number of searches of the same query in the regular web search, right? So, if people search for tequila images a lot in Google search, Google image search- It's going to know that you're looking for images. Yeah. Yeah. And I've seen that on a lot of searches. They try to understand the intent and try to serve you up what you need, right? And in that example, I said tequila images. Let's get rid of the word images.

If you just look tequila bottles, and a lot of people search for tequila bottles in Google image search, and they don't search for tequila bottles in Google regular search, then it's going to understand even when you search in regular search, you're probably looking for images. Well, and I think they probably have enough data, and that's why they're starting to combine the searches. Yep. Next, explicit image intent terms, right? Not explicit as a nude, but specific image intent terms. So, if you use the phrase pictures, it says explicit image. So, I got to say this. So, Google, when it first started doing images, like back in the day-day, okay? Like, one of the first things that it worked on was the percentage of skin that you saw in the photo because they were trying to figure out- Eliminate nudity. Yeah, yeah, yeah. Yeah. Right. So, again, this goes back to my example of tequila pictures or tequila bottles. If you use the phrase pictures, photos, diagrams, images in the search, Google understands that you're probably looking for pictures, photos, images, maybe diagrams. So, you're going to be looking for images. And then, explicit non-image intent terms, right? So, have you added something into your search phrase that is specifically saying, I don't want images? So, if you say articles, dissertations, studies, you're probably not, I mean, Adam, you may want some images in those articles, dissertations, or studies, but you're actually looking for the text, the body of those. Yeah. Or you can do the Boolean search and say, not. Not, yeah. Images. Yeah. Skip, negate images. That could work, too. There's a lot of search capability. That is what I would tell you is learn how to use search engines. Like, did you know that 75% of all searches are new searches still? Like, because everybody's like, what's the new phrase? Like, we'll type in a different way, different way to do it. But there's a lot of really powerful Boolean searches. And, you know, I think the younger generation, like, grew up in this,

so they, like, have it better. But, man, like, searches are so powerful. I've had conversation with young kids, and I'm like, that kid can't out-search me. All right. Typical case where image display environment would not show. The search contest. It's going to be like eSports gaming. Yeah. A user searches for SEO on Google, and there are a few images. There are few images labeled SEO compared to a lot of web pages talking about SEO. That's an example where it's not going to show you images. Hat, hat, hat. Yeah. Typical case where image display is inserted first and takes up a lot of space, a user searches for travel images. Every travel web page have at least a few high-quality image, right? Every web page. You better if you're a travel site. And then he jumps into this, so I thought this was really good. So note for that, that the query travel by itself. So you've got travel images, and he's showing the SERP, right? And there's lots of images. In fact, it's, you know, a whole section of kind of the Google image area, and then the next set is like unsplash.com, which has a bunch of images. Um, but if you just search, that's, that's for the phrase travel images. Yeah. If you just search for travel, there's a section of images. Now he's scrolled down. So it does meet the threshold for Google to show images. It just doesn't place them as high. And it's only one carousel instead of two carousels. So I would just say too, uh, when we're talking about images like unsplash, if you're using those stock images, uh, unsplash is already getting the credit for that. You're basically, I mean, you're not getting penalized for dupe content, but it's not helping you. So just think about that. So the opposite is also true where the intent means that the query is specifically or explicitly not interested in showing images. And he gives the example, travel dissertations. So like, if you're going to type in travel dissertations, there

is not going to be any sort of, um, Google show or image pack. What is a travel dissertation? Um, well, it's a paper about, you know, what it's like, like, look, there wasn't digital marketing in school, right? So there's like now go on a trip, right. And write about it. Yeah. I like that. I'll take that class. I mean, the first example is from, uh, georgetown, uh, dot edu, uh, funding grad students that is pleased to invite nominations for dissertation research travel awards up to 5,000 each. Oh my gosh. Yeah. My wife would win one of those. All right. So what determines the size of the image displayed in the SERP, right? Cause you, it's not just, Oh, does it meet the, does it, you know, trigger it so that it shows images? Like how big is it? Um, so the SERP is determined, the size in the SERP is determined by the image display and positioning system by calculating the image intent score and the flatness score. All right. Let's get into this higher intent score, higher quality images, the larger the image dimensions of the image display area. So again, if you're searching for things like image picture photos, and, and there are lots of high quality images available, then you're going to get a larger area note. If the intent score is high, but image quality is low overall, the number of rows may be lower. So you may only get one, uh, carousel of images versus two, where if you had better, more high quality images, it might show two carousels. So it's, it's like saying, what do we have and what are they looking for? Yeah. They're clearly looking for images, but all the images we have are garbage. We're not going to show them that many images. If they're possibly looking for images and there's a ton of high quality images, we're more likely to show those high quality images. All right. What determines the position of the images on the SERP? Again, we're going to go back. You

know, I told you kind of facetiously, we could have stopped at those first three. In fact, we'll go back and read them again and it'll make sense. Um, the insertion location. So that is, is it in the first position? Is it further down the page? Is it towards the bottom of the page? Yeah. Google decide where, what image it wants to grab. Yep. Yeah. Uh, query intent score. So the intent based on the image quality and search of the click data. Query normalization score. Normalization of image score, ensuring to process the placement of the image display environment in the same manner that it processes the placement of general search results. Did that make sense? Read it again. All right. Um, normalization of image score, ensuring to process the placement of the image display environment in the same manner that it processes the placement of general search results. So it's, it's, it's got a, a piece of the algorithm is saying, Hey, based on how does this answer the question? Right. Yeah. Again, when you say travel images, images, answer the question better than non images. Right. So, you know, even ADA, right. Like is really important for websites, but man, like you need an alt text, you need a description because Google's trying to serve the best search. And that best search is not just a tech search anymore. Could be images. It could be images. So, you know, you really, you really need to think about that. I think that's a blind spot for a lot of people. So, and then finally the central tendency score of quality scores, mean of the relevant score of the image. This is just straight up like the computations that are presented in the patent. Oh, no. I love these images. A very high score may bring the image pack to position one in the SERP. Like, and that makes sense. Yeah. The insertion score has to be greater than the relevant score of the general search results. And then he actually shows the full process. This is

from the patent, right? Breaks that down the text. And really, this is what determines, and we've covered the information really from the layman's perspective. We can get more technical. I don't think that's worth doing. No. We can go back and let's. Let's go through the definitions. No. You want to go through it? If you want to go through the definitions, we can. I want to go through the definitions. All right. Where are the definitions? Page eight. Okay, so we've got query confidence score. A measure of accuracy of the query search property ratio. Okay. Query intent score. The intent based on image quality and search and click data. Click data. Search property. And remember, that click data is like, where are people clicking on a traditional search? And then how many searches are happening actually in the Google image search? Search property ratio. This is the image search versus the web search. That's that same ratio I just mentioned. Click through rate. So that's the impressions clicks on the image corpus versus any of the search corpuses, any of the search properties. Insertion score. So that's insertion score. The position to insert images in the search result. That's one of the variables. Flatness score. Distribution of image quality scores. Okay. So yeah, I didn't know what that was. Yeah, I actually, yeah. I'm glad. Distribution of image quality scores. So how good are these images? Yeah. Central tendency score. mean of the relevance score of the images. Okay. Query normalization score. Put image and web search queries on the same scale. So you got to normalize it. So, I mean, I think what this is, is think today we may not be thinking in terms of, there's a lot of thinking, in terms of, hey, we want images on our traditional search. And so, when we don't find them there, because Google's not going to display them, because that's not necessarily what people are looking for, we're going to go over to the Google image search and search for that same phrase because we want images.

And then today, that might happen a lot. And as that happens a lot, us shifting over to Google image search, Google's aware of this and going, hey, maybe more people in the traditional search would like to see images and will start to show that. Let me finish. So, as we shift this, less people will go over to the image search. So, there's this normalization to kind of make a traditional search comparable to an image search. Yeah. No, different searches, based on the data they have, they're testing it out and they're ramping it up and then they're ramping it back down because I don't think, well, I mean, people that use image search, which I use image search sometimes. Yep. So, I don't think of that as part of natural search, but it does make sense that, hey, I just want the best answer. It could be all kinds of formats. Also, more people are listening and using Google, right, or watching Google than anything else. People would rather hear their answer. So, again, voice search is on the rise. Yep. And you're just trying to ruffle that feather. Oh, yeah. Second resource score, measure of relatedness of query to a subset of the image collection similar to the query intent score. And we got our final two. First, resources, search for web pages. Second, resources, search for images. Now, if we just roll back to where we started, where I said we could have quit. All right. Google shows the image pack in web search. In order to do that, Google needs to identify, one, if the query is an image-seeking intent. Yeah. So, we covered that. Does it have photo, image, pictures? Two, are there enough high-quality images to satisfy the user? Yeah. And then three, users often search for the query in the Google image search engine. This is really kind of tying all these together. And there you have it. Like, this is what Google is actively, well, it's an active patent. And it's what it's actively doing to figure out what,

if, and how much of the SERP should be dedicated to images. Yeah. So, shout-out to JC over at TripAdvisor. Yep. He's over there in Australia. I have some friends out that way. So, shout-out to all y'all on the down-under. I don't know what... Good day, mate. Is that? Yeah. Good day. Yeah. And my Bitcoin friends down there. But this is a good article. I think that this is really important stuff. And it's going to be continuously more important. And we're trying to get y'all prepared for it so you are not shocked when the search results are littered with images all over the place for all your searches and your site is not optimized. And you're like, why haven't I been focusing on photos this time? This is the podcast, number 575, where you're like, you know what, I've got to get focused on images and make sure I'm doing images right. Hey, if you get any value out of this podcast, and we hope you do, we would kind of... We probably wouldn't stop doing it because it serves two purposes for us internally. One, it keeps us up to date, right? And training our team. You actually really know a subject when you teach it, right? So, that's our opportunity to really know a subject here. It is a way that we train our team. And we know that you or you know that our team is up to date. And it's also a way, because we have a lot of fun on this podcast, that we kind of let culture trickle down into the team, right? So, they understand by listening to the podcast that... More beer. More beer, more fun, more education. I love... Yeah, no. So, we're in a very nice office that nobody else is in. And everybody's still working remote. And we've hired people all across everywhere. The globe. Yeah. Well, yeah. North America, I would say, mainly. Yeah. And we have a lot of miles. And so, we're bringing people in. And I don't know. I think it's

time to start getting back in the office, Chris. I do. I think it's a good idea. I think it's a good idea. So, if we've delivered value to you, and we hope we have, we would ask you to do something for us. So, maybe you found out about this podcast on a tweet, or somebody mentioned it in social media, or however you found us. Just go do that again. Yeah, just share that way how you found us. And, you know, really, this is a community. And it grows, and it builds based on you. We're going to do a lot more stuff. It's slated for next quarter. There are stuff that we're already be working on. People have been asking for a lot of trainings. You know, community. We're going to start a Facebook group. You know, so, we've heard you. We're listening, and we're working on it. Really, next quarter is what all that's slated for. We've got some exciting stuff to unveil. And until the next time, Chris. My name is Chris Berruth. My name is Matt Bertram. Bye-bye for now. Bye-bye. Bye-bye. Bye-bye. Bye. Bye-bye. Bye. I'm sorry. Bye. Bye.