How Can I Be The Most Efficient With My N-Gram Analysis? (Classic)

What's up Badger Nation? Hold on to your hats because we're about to unleash a ferocious classic episode from the deep dark archives of the PPC DEN podcast. This episode is guaranteed to set your Amazon PPC instincts on fire. Get ready to navigate the chaos madness and mayhem of Amazon advertising and unleash your inner badger. Everyone in Amazon advertising knows you need to scan through your search to report. And when you get your hands on data like this, you can really take your Amazon PPC to the next level. What's going on Badger Nation? My name is Michael Erickson-Pachine and welcome to the PPC DEN podcast. The world's first and longest running podcast all about Amazon advertising to make your Amazon PPC life a little bit easier and a little bit more profitable. Today we're going to be talking about one of my favorite topics. It is N-gram analysis. There's been some advancements that we've made over the last few months about N-gram analysis. So I'm going to run through it on how to do N-gram analysis the fastest and best way I know how in 2024. In case you're unfamiliar, we're going to be walking through what N-gram analysis is so that you can get advanced analysis from your search terms. So without further ado, let's jump in. I've launched campaigns and picked keywords. I've got my bits, some place monks too. Now badby's late, I've made a few. I've had my share of rock keywords, but I've got through. We have a PPC in my friends. And we'll keep up the rated to the end. You two are the PPC DEN. We're talking about Amazon. No time for wedding calls, cause we fix the gambles of war. So as you know, I've been talking about N-grams for over five years now here at AdBadger and on the PPC DEN podcast. And there's six advancements that I've made over my previous work on N-gram analysis. N-gram analysis continues to be one of the most popular and magnetic topics that I seem to talk about. I recently gave a presentation to the e-commerce community, core community, and people seem to really enjoy it. So we're going to be going over how to manage, navigate this in 2024. Now let's say you're unfamiliar with N-gram analysis. I'm going to walk you through it really quickly. So for sponsor products, sponsor brands, and basically any kind of search advertising, you have the keywords that you bid on, and then you have the search term that the people actually search to trigger your ads. So that list of search terms isn't always exactly the same as the keywords. So for example, I can do broad match keyword, Badger gear, and I might appear for Badger blanket. And Badger blanket would be the search term, and Badger gear would be the keyword. So search term analysis, actually seeing what triggered your ads, what actually got clicks on your ads, is an incredibly important process. So if we use the example of running shoes, I can have running shoes as my target, my keyword, and then I can have the search term running shoes for men or men's running shoes actually trigger my ads. And this is of course really important because maybe I'm selling women's shoes. I'm bidding on running shoes and I appear for men's running shoes. And I don't sell men's running shoes. I would want to find words like that and block them from my account and turn them into negative keywords. Now what's really important to do is to download your search report and basically run a filter where you look at your orders equals zero and then everything else with one order or more. And generally when you do this, you find out that you spend too much on things with no orders. And you spend not enough on your search terms with orders. This is just basic spread. You know, we don't have 100% conversion rates. So therefore we're going to end up with a lot of clicks, a lot of search terms that don't actually give us sales. Now this is from a real account over a 30 day window. This is a big beefy account. Many of us would love metrics like this once you look at the overall a cost. But when you dig in, you know, this account spends about $15,000 a month and it generates about $61,000 a month. So it actually has a good overall a cost. But then when you dig in and you look at the spend distribution of things with no orders versus things with orders, you end up realizing like, hey, wait a second, I spent $10,000 on things that did not convert. You know, so this is a real account. They spent $9,900 on search terms with zero orders. But things with orders, they spend only $4,000 and those $4,000 generated $61,000 revenue. So they generate a lot of revenue from the things that have orders, but they ended up wasting a lot of, wasting a lot of spend on search terms that they did not convert. So you should download your search to report. We have lots of content talking about where to find that and how to download your search to report. And you want to do this for sponsored products and sponsor brands. But what's the classic advice to do is, oh, you want to save some money on your search to report. You want to prevent wasted ad spend, all those good things. So why don't you just go into your search to report and download it and then negate everything with over 20 spend and no orders. That's the advice, right? And you should do this except a big issue. Most search terms only have one to two clicks. That's just the truth of it. So again, I might see running shoes for men. I got one click. No sales men's running shoes to clicks. No sales down press. Ringer only one click. No sales hand closed ringer. One click. No sales hand crank laundry ringer. Four clicks. No sales. So all these are like ones. E two Z's. I'm spending 90 cents. I'm spending 60 cents. I'm spending 47 cents. I'm spending almost nothing. But don't forget this added up to nine thousand dollars of things with no sales. So you can put that search to report into a pivot table and you can run that analysis and you can actually see. If I were to actually look at for an account that spends 14, 15 thousand dollars. If you actually see, is there any search term that gets over 100 clicks? Well, there's a couple of them. Well, there's one, two, three, four, five, six, seven, eight, nine, ten, eleven, twelve, thirteen, fourteen, fifteen. There's about maybe let's just say 20 search terms that actually get over 100 clicks. The rest of it is all of these tiny ones and two Z search terms. Where individually, you can't really make a decision on the search term that gets one click. That's really difficult. So in fact, if I were to look into this account, how many search terms get five clicks or less? It was 6,293. That's 94% of the total search terms are getting one to five clicks. And then only 5% of the search terms are getting over five clicks. So it's an insane amount of individual search terms that get almost no traffic. That get almost no data that are really difficult to make decisions on. So it's very rare actually that you'll end up with search terms that have 20 dollars in spend with no orders. 30 clicks and no orders. Those are actually relatively rare. What's way more common are search terms that get one click, two click, three click, no sales. And you end up with thousands of them. So with this particular account, there was $9,900 of spend with zero orders. All of these things have very, very few clicks, one click, two click, and you know, they're kind of relevant. And like, what do you do with this information? Well, you begin to find trends. You begin to see that list search term and this other search term, which each have only one click, have a common word in both of them. So up on the screen, I have hand crank laundry ringer and laundry squeeze ringer. Both of these have the word laundry. And I can get thinking they both also have the word ringer. I can start thinking, what's the performance of all of my search terms with the word laundry in it? And then I can run analysis and maybe I might find that everywhere where the search term includes the word laundry, maybe all of those search terms I end up spending $50 a spend without an order. So I can just get rid of every search term with the word laundry in it and be way better for it. This kind of analysis is known as N-gram analysis. N-gram analysis. It's an N-gram. N-gram analysis is a thing that exists outside of the world of Amazon PBC. But since we're analyzing a list of words and performance around those words, we can use N-gram analysis to zero in and better understand our search terms. So the N to take you back to math class is the variable. So it's one-gram. That's a one-word trend like. laundry in that last example. Two gram, three gram, four gram, we can go all the way up to five gram. So what we can begin to do is we can take our search term report and we can begin to break it out. So we can say running shoes for men, we can turn into running and then separately shoes for men. And we can sort of break this out. Now years ago I created a spreadsheet, I share with everyone, it's one of the most popular spreadsheet tools that we've released. I highly suggest you go get it, play with it. It's awesome. It's fantastic. And what it does is it basically does that for you. It lists everything out for you. So you know, Jim bucket with the Spencer, it broke all of these things out into individual terms. And then I can see the pooled data every search term with the word Jim in it. What was the performance every search term with the word bucket? What was the performance every search term with the word dispenser? What was the performance? And it really allows me to dig in and find search terms that perform really well and search terms that do not perform well. So for example, it's going to allow me to find new negative keywords. In this example, I did an N gram analysis, a one gram analysis. And what you see here is the word metal everywhere where the word metal is. Maybe it appeared like one time, you know, metal dispenser, one click, metal barbecue, one click, almost nothing. But then what this allows me to do is it allows me to find every single search term with the word metal in it. And boom, tell me that I got 146 clicks for all of them. It's been $140 on all of them. And it was out of 320% a cost. So I can go in there and take action. And this is one of my favorite ways to analyze search terms. In fact, I would say your search term analysis is incomplete if you do not do this. So yeah, there were 97 search terms with the word metal in it. And this is actually real example. Somebody was searching metal gatorade dispenser. Okay, maybe people are out there drinking their gatorade on a hot summer day out of a metal, out of a metal dispenser. But either way, that got searched one time. I spent 4th, three cents on it. And I ended up spending a bunch of money on things with the word metal in it. Just these onesy, twosy clicks. But when they combine that, I end up with 140 spend through 120% a cost. I can just negate every term with the word metal in it. And this was a really interesting example. In a for a two gram, I found the trend coffee earned, which is apparently a thing. So again, just onesy, twosy clicks with the word coffee and earn in it. Didn't really know much of it. But I ended up spending 44 clicks, $27. And it was at a 9% a cost. This is something for me to lean into. I probably have an unfair advantage compared to my competition on this. So awesome stuff. So we have this spreadsheet. You can go and find this spreadsheet. If you just go to Google and type in, you know, let's say Amazon and gram badger. You'll go find that we have an end gram analysis tool for Amazon advertising. It's a spreadsheet. It's pretty cool. Works great. I did a cool episode with Elizabeth Green way back when, when did I do this? I did it way back in just one year ago, 2022. So two years ago. It was really cool. Since then, we've made some advancements on this really speeding up this process. So there's been a couple things before we get into that. Really how often should you do this? I would say when you're doing it with a spreadsheet method, maybe once a quarter, maybe if you have an issue, like, oh, my a cost is high or like my I'm spending too much, I think on waste of terms. Now we can do it as frequently as we want to. So this is a little bit about the end gram tool and how I do it now. So the first thing is combining sponsored products and sponsored brands. Previously, you had to do this one at a time. And now we sort of combine both of them, which is really, really nice. So that way it pulls your data. You have a bigger data set, which is really nice. The second thing too, the spreadsheet did not list the campaign name and the portfolio name or the ad group name. And what we've done now is allow that. When you're inside the end gram analyzer and you find a one gram, for example, you can click on it and it will actually tell you what triggered it. So you see all the search terms with the particular word in it. This is not in English, but you can see the keyword ET, the phrase ET is in everywhere. And then I can see the search term next to what triggered it as well as the ad group and the campaign and portfolio, if there was a portfolio right in there. So that was a major benefit now that I can actually see where these things triggered. And then of course, from there, I can actually go in and I can say, hey, you know what, I'm going to block this search term. I'm going to turn into a negative phrase or negative exact and where am I going to put it? And I can put in specific places, which is really, really nice. So that's been really nice to be able to go from my analysis, see where it's triggering, see what triggered it, and then make decisions about if I want to promote it to a positive keyword or negative keyword right from the analysis screen. The second thing too is this, you know, using a spreadsheet is super awesome. It's a great way to access this end gram data. But what's really cool and really convenient is just being able to have it right there where I can say, okay, show me the last 30 days, you know, go back longer, show me the last 60 days, show me the entire last year. So I can pull this data from really long periods and I can begin to scan and see how things are trending over time. I can get a good sense of everything. It's awesome. And then of course, dig in a little bit deeper and I can say, okay, well, I have the end gram ET. I can click on it and I can see what the composition of that is. I can see where it triggered. I can go back and forth really quickly, which is really nice. And then of course, going all the way up to five grams, you can see that, so you can get more data, you know, the spreadsheet that I made way back when, it goes up to three grams. So yeah, so that's sort of how I'm navigating through it now. So this is probably one of my favorite, favorite things that we do here. Just analyzing end gram, analyzing the search term, reports, a lot, lot faster. So hopefully if you haven't gotten that spreadsheet way back when you want to play around, play around with end grams, I highly suggest you make that search. We'll probably put it in the description here. So you can go get that spreadsheet and you can begin to do search term analysis. You build off of it, have fun. But as of right now, I'm sort of really enjoying that we have this in our tool now. So as far as I know, this is the fastest, easiest, best way to analyze end grams to go back and forth between one to four to five grams and see what triggered it and make decisions about adding positives and negatives. So I absolutely love that and end grams. If you're not using them, be sure to use them. It's a major, major part of analyzing your search terms at a really high advanced level so that you can get insight from all these low-click search terms. I hope you enjoyed this episode of the PVC then podcast. Hopefully this has given you some homework, some things to do. At the very least, go get that spreadsheet for sure. I will include a link to it. And if you have any questions about an app based end gram analyzer, hit me up. We'd love to go through things with you. So have a good one and I'll see you next week here on the PVC then podcast. [Music]