How N-gram analysis found a keyword with 200 clicks and 1 order

How N-gram analysis found a keyword with 200 clicks and 1 order Michael Erickson Facchin: What's going on Badger Nation? Welcome to The PPC Den podcast, the world's first and longest running show all about how to make your Amazon advertising life a little bit easier and a little bit more profitable. We have a fantastic resource in the description of this video where you can go and get our episode guide that's organized by topic and theme so that you can really dig into a lot of the topics when you need them, as you need them. Just that you go and get that. Today on the show, we're going to be touching something that is incredibly important to talk about. Ngram analysis, I've talked about it on the show for quite some time. But recently, I received feedback that, hey, after I get my Ngram analysis, how do I do Ngram optimization? What do I actually do with the data after I take my one, two-click search terms and I begin to extrapolate information from them and I find my one grams and I find my two grams? What do I actually do with this? So we're going to talk about four very common, very effective ways to optimize your PPC accounts using Ngram analysis. And I promise you, if you've never done this kind of analysis, you will find some gold and you'll have a good opportunity to trim some fat. So we're actually going to be talking about how to optimize Ngrams. Let's jump in. Unknown Speaker: Paint and pick keywords. I've got my bit. Set placements too. No bad mistakes. I've made a few. I've had my share of rankings. We are the people shooting, my friends. And we'll keep on the blazing. To all the PPC Den, we've talked about Amazon. No time for letting go, because we've fixed the game. Michael Erickson Facchin: Michael, you have to forgive me. I'm wearing a jacket today because it's blistering, like, 45 degrees here, which is unacceptable for me. Michael Tejeda: Very cold. Michael Erickson Facchin: I moved to Austin for sun and warmth, and if it gets below $50, I'm writing to the mayor and asking for some money back. Michael Tejeda: Yeah, I hear you. I hear you totally. Michael Erickson Facchin: And so what's new in the world of Amazon advertising for you? How are you feeling about Amazon these days coming back from a winter break? Michael Tejeda: Yeah, so it was good for me. I'd stepped away for a couple weeks, so I was able to get a fresh perspective on everything. And sometimes those little breaks do help you come back with refreshed eyes. So what's new for Amazon for me, just step away for that data and just seeing it kind of in a new light. Michael Erickson Facchin: I think it's a myth to like, You know, grind all the time, work so hard, burn the bridge, you know, burn the candle at both ends, all these different things. And it's like, I think actually you perform more better over time if you do take time to rest and rejuvenate, for sure. Michael Tejeda: Yeah, agreed. Michael Erickson Facchin: And today on the show, This is actually a very frequently requested topic, which is not so much Ngram analysis, but Ngram optimization. You know, the whole concept of analyzing grams, I must have first heard about like 10 years ago doing Google ads, and it was just a way to analyze your search terms that was so Useful, because most search terms have very few clicks. So it's a way to group data. But the thing that gets requested, and dear listener, if you don't know what Ngram analysis is, we'll break it down a little bit here for you. But we've talked about this topic a lot in the past on the show. But today we're going to be actually going through some scenarios of actually doing Ngram analysis, which is going to be awesome. I'm excited about it. And truthfully, this took me some time. When I first started doing Ngram analysis, I remember just sort of looking at it, thinking it was awesome, finding some cool trends, and then having a moment where I was like, what do I actually do with all this information? Michael Tejeda: There's a lot of data, that's for sure. But that's what it's all about, right? Breaking it down and finding those actionable steps. Michael Erickson Facchin: So let's start with just defining The problem that Ngram solves and what it is. So, you know, a very, very common approach to search term management is to download a search term report, look at your search terms and go to orders equals zero. So you can find your non-converging spend. And when you do that, you'll have thousands and thousands of search terms, many of which Have very few clicks. Which is very true and very frustrating to deal with. And most search terms will have very, very few clicks. And it's the hardest question to answer, which is like, you know, well, what do you do with all this low click information? I think I was looking at an account. It had like 70% non-converting spend. From its search terms. And most of them are like one click, two click, three click, four click. And historically, there wasn't a good way to do anything about this. And it's really, really tough to solve. Does this ever get brought up to you when you're talking to your customers? Michael Tejeda: Oh, yeah, all the time. It's one of those things of, you know, how do I turn off that tap on a wasted spin? And this tool has helped tremendously. And I like that it's built in door software. So that actually helps quite a bit. Michael Erickson Facchin: Totally. Yeah. There's definitely ways to do this outside of any tool. On the show, I always like to give people options by not having to pay for anything. We do have in our Amazon PPC checklist, which you should grab in the description of this video, we do have a spreadsheet that you can follow along with and run everything. We've got videos of The spreadsheet in use on this channel, which we'll also link to. But yeah, we did build some solutions of things that we were commonly doing just to speed up the process. Because the more frequently you look at your Ngrams, usually the better and more in control your account is. So anyway, you have a search report where many things have very few clicks. And it's an issue because it's hard to make decisions because, you know, what do you do with things with one, two, three clicks that are relevant? Hard to find trends. And basically, how do you define what Ngram analysis does to a normal search and report? How do you go from normal search and report to Ngram analysis? How do you define what an Ngram is? Michael Tejeda: Sure. So what I would define it as is not only finding the root problem, but also a lot of the Longer tail keywords that might be issues for you. So that's the way I would define it as being able to find, you know, and filter through everything to find the exact pinpointed spots that are major hotspots. Michael Erickson Facchin: Yeah, like imagine Imagine you're selling disposable cutlery and let's say you have a whole bunch of SKUs selling disposable cutlery and the word spork comes up. And you look at your first search term and the search term is like disposable cutlery with spork, one click. And you sell a bundle. You sell some forks, spoons and some sporks. So you're like, it is relevant. I don't necessarily want to get rid of it just yet. So I'm going to let it ride. It only has one click. But you look at your entire search term report and you spent like 50% on things without order. So you're trying to bring it down. And here's one random search term. Disposable cutlery, spork. Okay, I'm going to leave it. And then it's like, you go to the next search term. And it's like, pack of cutlery, miscellaneous with sporks. Let's say, okay, well, that's what I sell. And you keep going down the line and you find all these individual search terms with like one, two, three clicks. And if you summed all of the ones up with the word spork in it, you might realize that, well, when that word spork is in there, My A cost is like 900%. I get lots of clicks. I have 55 different search terms with the word spork in it and 200 clicks for all these different search terms and only one order. So that sort of tells me that, number one, that gives me product insight. That tells me people are buying my product, probably not for the spork, and I should probably put more emphasis on the You know, spoons and forks. And then from there, I can also see, okay, everywhere where the place spork shows up, I'm just like, you're having a really hard time converting. So then I can make some decisions on spork because I had 55 different search terms with the word spork in it. You know, I had two Clicks with the word spork in it, and I have data for all the search terms with the word spork in it, and then I can now go and do some optimization. So it's all about finding the root word that is common in a whole bunch of search terms instead of looking at them individually. So that's a perfect example. Michael Tejeda: Spork. I like that. Michael Erickson Facchin: Now, the hard thing is, of course, is what do you do once you have this list and you're looking at your list of Ngrams? We're actually going to look at some demo dummy data here, and we're going to make some decisions and some workflows on it. So this episode, if you've never You've never looked at your search terms in this particular way. I recommend you pause this video and go to, you know, we have some other content on Ngrams and run through the spreadsheet just on your own account just so you can see all of this information, see what it is that we're talking about, see how You take your search term report and you find those common root words and you're able to extract analysis from like search terms that have one, two, three clicks in bulk. It's a really powerful way to analyze. Anything else to say about the analysis? Of search terms into Ngrams? Michael Tejeda: Just the fact that we did use that spreadsheet before we had the tool and it worked beautifully. Definitely recommend everybody does check that out and run through it just to kind of understand what the process is and to dig deeper into that data for yourself just for your own sake of understanding your own data. You should really do it. Michael Erickson Facchin: And you know, I'll even put it up on screen here so everyone can sort of look at it together. So yeah, so this is the spreadsheet. Feel free to play around with this, but it's like running water bottle. It takes the word running water bottle and it breaks it up into three separate words, running, water and bottle separately. And then it will go and find every other search term with one of these words in it and create a one gram would be All the one words together. A 2-gram would be like, okay, running water. Find every other search term with the word running water in it. And it sort of builds on this specificity, you know, all the way up to 3-gram, 4-gram, 5-gram that allows you to get really good insight into things that have like one, two, three clicks, which is amazing. So I'm looking at this example. I have a 1-gram badger. And I can see that everywhere where the badger showed up, I had this particular ACOS and this particular amount of revenue for all the search terms with the word badger in it. And you know, how many search terms that I have with the word badger in it, you know, a ton. And they're all like one, two clicks. So it allows me to see this data rolled up in aggregate, which is really powerful. Now the question becomes, what do I do with this information? Like, what do I, how do I go from Looking at this, is actually doing something meaningful with it? Let's answer that question. So up on screen, we have four activities that we're going to do. Three of them help you find your worst terms, and then one of them helps you find your new great terms. So you can do some gold mining, And some coal mining to find the bad things and find the good things. So we've got three activities to find some bad things, add new negatives, and we've got one activity to find some new positives. So let's jump in. And we are going to be using our tool for this. So let the games begin. So the first thing that we'll be doing is talking about time frame. If you're downloading a search and report from Amazon, you very likely are downloading it for a 30 day time frame or a 60 day time frame, you know, like a one or two month period. And that's great for sure. And doing a short, I would call that short time. Would you also call like a one month period, a short time frame? Michael Tejeda: Yeah, for sure. Michael Erickson Facchin: And then, of course, a longer time frame, if it's available to you, you can combine multiple search term reports, or you can use a tool that does this kind of bulk analysis for you, where you can look at like a year's worth of search term analysis. So you had this comment on here. We had at this point, they saw something you mentioned, and it's something I intuitively do too, but I've never like written it down, which is actually run the analysis on a short time, short term and a long term. Tell me about that. Michael Tejeda: Yeah, so just taking different time frames into account, as you were mentioning, 30 days is very short. Sometimes we'll look at like 7 or 14. Also, just to see if there's immediate changes, especially if there's seasonality shifts like, say, Christmas or Black Friday, Several Monday, that sort of situation. But there's a lot of times where this Ngram analysis tool really gives us a lot more data if we're looking at a longer time frame. So combining those reports into getting longer than 90 days or using a tool like this that'll help. You know, put it in bulk aggregate. Awesome. Michael Erickson Facchin: Yeah. So yeah, I thought that was really interesting to like see if anything's popping out in the short term. Yeah, new that might get muddied in with like a really long. So all these activities I would go through once at a shorter time frame and then go through again. It was a longer time frame. I thought that was a really nice optimization point. So activity one to find some irrelevant things is just a scan. It's just a manual scan. That's definitely the first place to begin. If you think about what Ngram analysis is even doing, it's taking a search term report and it usually makes it smaller, meaning it's reducing the amount of stuff to look at. So what I mean by that is, you know, in this demo document, it had 194 sample search terms. But when I look at the one grams, it dropped to 125. I think we saw in another example that we had, you know, like 5,000 search terms, but only like 2,001 grams. So it's easier to scan for irrelevant words, just like literally, we're just looking for words that are irrelevant. It's much easier to do in an Ngram view than it is your search and report, because your search and report likely has thousands and thousands, where your Ngram analysis probably just has, you know, maybe half or a quarter of that. I generally have found that to be true. There are fewer Ngrams than there are actual search terms, because we're finding the trends. How often would you say you've actually find something irrelevant? Michael Tejeda: irrelevant, irrelevant, more often than I want, just because like if we're using a phrase term or even a broad term, we will get kind of those irrelevant things that will pop into it, like say, We're talking about disposable cutlery and sporks specifically. I might get something that's like a disposable straw. I mean, yeah, it's relevant as disposable, but it's not exactly what we're targeting. Michael Erickson Facchin: Exactly. So that's a really easy place to begin. Just visually scan. I'm just scrolling through. I'm looking for words that purely are irrelevant. And it's much easier and faster to do it in an Ngram environment than it is your search and report. And when you find one, there's a couple different actions to take here. Number one, you can just add new negative execs. You can reduce spending on the keywords. You can potentially stop bidding. Or you can add it as a negative phrase. So let's actually break this down. Let's maybe a real example in step two. The first one is just a visual scan, look through things, find something relevant. Let's actually get to something a little bit grittier here, which is going to be orders equals zero, sort by most spend. This activity would look like this. I usually like to start with one grams. And for the purposes of this example, let's go long time frame analysis. Let's scan through our search term reports from July all the way to January. So I'm looking for orders equals zero. So we're going to run a filter. Orders equal Zero. And now I'm going to find stuff where the orders equal zero and I'm going to sort this from spend high to low. Let's find a word like lid. So I found a word lid in my demo dummy data. I have the word lid and I spent $97 and to get no sales. So mission accomplished. I've identified something that incurred a lot of clicks. No sales. And I have some options with what I want to do here. So when you see this sort of demo data, this is kind of representative of a real account. It is very likely the first time you're running Ngram analysis, you go to one grams and you see your list that you find a route that is pretty misbehaving, something like 97 and spend. No sales. What comes up for you when, you know, you see stuff like this? It's like, dang, you know, this isn't an individual search term that spent 97. This is a group of, let me give you the count of how many terms the word lid was in. It was in 59 different terms, you know, compostable cup with lid, solo cup with lid, all these different terms with lid in it. You know, what goes through your head when you see something like this? Michael Tejeda: Definitely that there's some actions that need to be taken. So we need to start sorting things out and start making sure that we're negativing and adding those as negatives to make sure that we're mitigating some of that damage because sometimes it's just a little bit too much spent in the wrong places. Michael Erickson Facchin: The first activity is to really just scan for things that are irrelevant. So this is really simple. All I'm doing is I'm Looking at my one gram data, and I'm just looking for anything irrelevant, any irrelevant word, seeing if I spot any trends there. So a visual scan. And in order to illustrate what to actually do if you find something irrelevant, let's actually jump to the second activity, which is find things without orders and sort it by spend. So I love this sort of second activity. Basically what I'm going to do is I'm going to go to one gram, And for the purposes of this demo and this dummy data, I'm just going to go with a long time frame. So I'm going to go back all the way from July 1st to January 7th. And I'm going to run a filter for orders equals zero and sort it by spend greatest to least. And right off the bat, I've got plenty to chew on. I'm going to grab the one gram lid as the example here. So the word lid here is running away with spend and Not converting. And it's sort of just the nature of the beast being on Amazon. You would never let an individual search term get all the way up to 97 and spend without an order, but that's just not how search-based PPC works. And if I click on the word lid, I can see, you know, compostable cup with lid, solo cup with lid, disposable cup with lid. So like I'm seeing the word lid show up over and over and over again. And if I look at the amount of Clicks for each one, it's like one click, one click, two click, one click. I would have never found this if I was looking at individual search terms. But again, I'm using Ngram analysis, I find it, and now we get to take action. So what is the action to take when you sort by orders equal zero, sort by spend greatest to least, now what do I do here? Talk to me about what you would do. Michael Tejeda: Right, right. And that's what I love about this whole process is that you just are getting this data and aggregate and kind of breaking it out. But the first step that we do take is just adding negative exacts. So if there's something that's popping out to you that you can see that is just irrelevant or spinning a lot, just negative exact it right away. Michael Erickson Facchin: Pause there. So like I found the word lid and now I go into lid. I look at all the individual search terms. Michael Tejeda: Exactly. Michael Erickson Facchin: And then I can go through here and I can add some new negative exacts. Tell me about like what you're picking to add as a negative exact, like combustible cup with lid, you know, I've got one click for it. Should I add all of these as, you know, they all sort of have one or two clicks. Should I add all of them as a negative exact? Michael Tejeda: Definitely a great question, but the couple things that I kind of look at are, okay, well, What's our CPC, right? And do we have, obviously, these are things that don't have conversions on them right now. So if I have a very high CPC, that kind of brings us to the next step. Okay, well, am I going to want to negative exact it if I only have one click on it and it's a high CPC? Maybe not. It might convert at something lower. So let's give it a shot at something lower. So that's where we kind of start going into our other steps. So then it's right there, number two, right? Reduce spending on these keywords. So lower that CPC, lower that bid. And then the next part there is, okay, well, if we do stop bidding on these, is it something that we want to make sure that we just negative phrase match out completely? Like is everything with LID not converting? And I know right now we have it sorted by Zero sales, but are there some that have sales? Maybe they do and maybe they're running away at like a thousand percent ACoS. And obviously if it's, you know, we have a couple conversions and they're a thousand percent ACoS, not totally worth it for us. So maybe that's a negative phrase opportunity there. Michael Erickson Facchin: Yeah, exactly. So like the mental decisions that you're making are very similar to mine. You know, I'm looking at it and I'm number one, like where is the business? Is the business really struggling with ACoS right now? If so, I'm going to be A lot harder on this, right? And you know, in this example, in this demo data, I have a lot of terms. I have a lot of grams spend with no Sales. So to me in this demo, I can be pretty aggressive with the way that I'm adding negatives. So there's a couple of things to do. In my head, one is sort of an advanced or nuanced view, which is I have the word lid as the root word. Everywhere where it showed up over the last 191 days did not get an order. I kind of like the word lid. Maybe I'm selling lids. So one thing I can do is I'm just going to add everything that appeared as a negative exact and I'll just like shrink the amount of stuff that I trigger for relating to lid. So I'll still get some traffic related to lid because it's being triggered, but I won't get this kind of traffic for the word lid. So that is where if I'm in an account that's struggling with ACOS, It's an acceptable move to just be like, every single thing with the word lid, you're getting at it as negative exact. Goodbye. I'll still get some traffic for the word lid, you know, with my expanded ASIN targeting, with my auto targeting, with maybe broadened phrase, I'll get some lid traffic, but I'm going to get less traffic and I'm not going to get this traffic anymore. So that's where you can add some negative exacts. Or it could have been a little bit more discerning there. Got rid of some negative exacts. You also mentioned I can just reduce spending. Maybe I go into the campaigns, reduce the budget. Maybe I go into the campaigns or ad groups and just wholesale reduce those bids. That's an option. I could potentially just pause the keywords. That's triggering it. That's an option. And you are absolutely right. You want to be really thoughtful and careful when you add a negative phrase because if I were to just add negative phrase lid, I'd be blocking all the existing traffic and all future traffic with the word lid in it. And that might be totally good. Again, if the account's struggling with ACOS, I would say in this case, I haven't converted on LID over a hundred bucks of spend. It's just not going to happen. If the account's struggling with ACOS, we're trying to get this down, I'd probably add it as a negative phrase and just get rid of the word LID and all of its phrases in one swoop. Michael Tejeda: And there's one thing that we need to take into account also, too, is kind of like where we're at. So we're kind of almost framing this as like a evergreen type product, like something that we've optimized already, something that we've taken several stabs at, and we're trying to convert on particular areas of focus. Now, if this was a new launch product, then obviously things are going to be a little bit different for us. But that's not what we're talking about in this specific case. So just wanted to make sure that we were clear on that too. Michael Erickson Facchin: Great point. So much of Amazon is situational of like, what stage of the game is the business in and like, what are we doing with our, you know, how do we consider bad ACoS right now? For sure. That's a great point. So yeah. So, so far we've got two activities, same actions on both steps. Find things that are irrelevant. I'm doing a visual scan. Find things without any orders. I'm doing a visual scan. And for activity two of like finding orders equals zero, sort by most spend, I generally do that on one grams. That's where you find the beefiest thing. And later, I'll show you a way to jump from Ngram to Ngram in just a second. But let's do activity three. Very, very similar. All of these first three activities have the exact same action steps of trimming the fat, of getting rid of things that are underperforming. But so the second one was orders equals zero, sort by most spend. Now we're going to do things that have converted, but Everybody's favorite, high ACOS. So we're going to go back here in our example and get rid of the orders equals zero. And we're going to sort by ACOS high to low. And when we sort that high to low, oh boy, we've got some high A-costs. We've got a 1,268% A-cost. We've got an 1,150% A-cost. We've got a 500, 600% A-cost. Let's go with the word white. Let's go with the word white, a nice little good generic one gram term. 436% ACOS is the one gram. So again, just recapping what it means. It means that the word white in every single search term that the word white is found in, it ultimately summed up to a 436% ACOS percentage. I'm going to dig in. I'm going to see those individual search terms that have the word white in it. And I've got, you know, plastic forks and spoon, white, white forks, so on and so forth in this demo account. So I've got a whole bunch of data related to the word white. I did get one order. So protocol is the same here. Protocol is exactly the same. Repeat it for the good people in the back. Michael Tejeda: So add native execs. That's our first swath at it. And the next one is to reduce your bids. So make sure the spinning is down a little bit on those keywords, and then potentially pausing particular keywords that we're bidding on. And the last point, number four, is adding that phrase, but we want to be very careful with that. There's actually a fifth step, and I think that we noted it way at the top, but it was one of the obvious steps too. So if we go back to where we saw the tool, We can see on the screen here, and for the people on the podcast that can't see this, there's one that is talking about a coffee cup. So white paper coffee cup is kind of very irrelevant to the rest of the search terms that we see here on the screen. And it's that visual marker too. And I know that we had said the word visual a couple of times also, but it's finding those things that kind of are out of sorts with the rest of the search terms. And those are things that need to be separated and isolated into their own sort of campaign. So this is irrelevant for us right now. The Ngram's identifying it. So we definitely negate that there. Michael Erickson Facchin: Yeah, exactly. Work that system. Find the things that are orders equals zero, A costs very high, jump in, see if there's any negative exacts. It could be some or all of them. Reduce the spending on the keywords. Potentially stop bidding on these keywords. And then last thing, if you really want to terminate an entire Ngram, you would just add it as a negative phrase. So that's trimming the fat, which I think is the most popular way to use Ngrams. Now let's jump into a way to find new positive keywords to bid on and rank for by using Ngram analysis. So we often talk about Ngrams as a way to trim the fat, but you can actually mine for gold here because it's the same thing. You might not notice it if it's like one search term with one click and one order, but it's possible that that gram is was, when you take it in aggregate, actually ended up with, you know, 20 clicks and three orders. Maybe it's a great search term. So that's great. So this filter, well, it's a sort. We're going to sort our one grams from high to low and do something that I like to call Ngram graduation, where we work our way up the Ngram ladder. If I take this and I sort by sales high to low, I will find, let's pick a nice generic one, maybe I would have mixed. Should we go with Forks? Michael Tejeda: Sure. Michael Erickson Facchin: So Forks has a good amount of sales. And then what I'm going to do is I'm going to go next level up. I'm going to go to 2 grams and I'm going to type in the word Forks and find the terms with the word Forks in it. Now I see 2 grams and all these 2 grams have the word Forks in it. And I'm going to scan for something that is just absolutely crushing it. And I've got, yeah, I feel pretty good about Let's say forks bulk. That's pretty interesting. And I'll go up to three grams. And then when I'm at three grams, I will type in forks bulk and I will find things that are bulk here related. And I find some interesting things. I see something with a lot of orders. I see plastic forks bulk. And right there, I can see that, hmm, this gram converts very well for me. I'm going to investigate this in a little bit more detail and I'm going to scan through and I'm going to find the term that just is running away with it. And it is actually that gram. And it allows me to find a term and then verify, do I have this as an exact match? Do I have this in a top of search percentage high campaign? So, I find this and then I know that this is a search term that does well. I should probably go after it. So, I'm sort of looking at the gram and I'm making sure that I'm actively bidding on the thing that I want. And in this case, I sort of worked my way up from forks to forks in bulk and it allowed me to identify a term that maybe I wasn't aware of that was doing so well, a term that I might want to optimize for, a term that I might want to You know, be sure that's in my product page somewhere and be sure that I have as an as an exact match with a strong top of search somewhere because it's probably worth ranking for as well. So that sort of process helps me mine for the best terms. So like the one gram put into the two gram put into the three gram until you sort of find A word that either you're not bidding on, was being triggered by auto campaigns, or is that a variant of one of the terms that you're appearing for, and it allows you to identify trends that you would have missed because they're spread out over 50 different search terms. So I absolutely love this sort of Ngram graduation technique to find new terms to bid on. Michael Tejeda: Yeah, I actually like the other term that you had there too, which is a further extension of it that had the number 1000 in it. 1000 count. So that was looking pretty good as well too. Actually, take those two guys out and push them a little bit further. Right. Michael Erickson Facchin: Amazing conversion rates, good ACoS. I want to lean into this and hopefully rank for this organically as well. So, perfect observation there. And that is Ngram analysis and optimization. So, we spent the bulk of this episode talking about Ngram optimization, actually doing work after you run your Ngram. And I love it. It's something to do, I mean, for any account, do this twice a month at least. It's a very useful activity to help you mine your search terms in ways that you simply cannot if you're looking at a default search term report. Go, people, and get this done, whether you're doing it in a spreadsheet or your favorite Amazon PPC tool. Anything else to say about Ngram as we wrap here? Michael Tejeda: I love it. I'm a fan. One of the things that I like is just finding that drippy, leaky faucet and being able to turn it off because there's a lot of accounts that just kind of drip away with the one-click, two-click type situations and just being able to find that. Michael Erickson Facchin: I mean, three out of four activities are about trimming the fat. Michael Tejeda: Exactly. Michael Erickson Facchin: And I don't think people realize how to find the good parts of their Ngram. Sort it from sales, highs to low, and begin to dig in. And I promise you, especially if you've never done this before, you will find tons of gold there. I was talking to someone the other day and they noticed that they found the word Diwali. Their product was very popular for Diwali celebrations that they had no idea of. Because like maybe they see it like a onesie, twosie, doesn't have much data, but they found an aggregate. It had like, you know, 13, 14 orders in a month when you summed up all the Diwali words, making it really powerful. So they put it in an exact match, top of search, high percentage and went after it and are trying to rank organically for it. So there's a lot of cool stuff with Ngrams that you just can't access because search terms are listing things one by one, you know, one click, two click. Yeah, I love it. This is like the 10th episode I've done on Ngrams, but really the first one where we've actually like done the optimization and hopefully inspires people to go take action. Thanks so much for coming back on the show, Michael. I'll leave you to it to get back to these campaigns and optimize them. Everyone else, I'll see you next week here on The PPC Den Podcast. Michael Tejeda: Bye guys.