Tackle 70% Wasted Spend with N-Gram Laddering in Amazon Ads

Tackle 70% Wasted Spend with N-Gram Laddering in Amazon Ads Michael Erickson Facchin: Today, we're going to be doing N-Gram Laddering. Now, if you've listened to the show before, if you've been following, this might be the sixth or seventh piece of content I've done on N-Grams. I still believe in the epidemic of non-converting spend for Amazon marketers. I still think it's one of the biggest issues in Amazon campaigns, so I'm continuing to bang on that battle drum. What I mean by that is, I open up some accounts and I see 50, 60, 70% non-converting spend, meaning you're spending 10 grand, you're not converting on 5, 6, 7. So even if you have a good ACOS, an ACOS that's in the 30s or the 40s, it is possible that you have a high percentage of non-converting spend and your ACOS could be even better or your sales volume could be even higher if you were dedicating your ad spend dollars towards things that actually converted versus non-converting. So I wanted to start with a brief reminder of what N-Grams are in case you have not been following along. In a normal Amazon PPC account, you will find that most search terms have five clicks or fewer. In this account, there were 6,662 total search terms in a given month. In them, 94.4% had one, two, three, four, or five clicks. So 6,000 had 5 clicks or fewer. Only 369 had over 5 clicks in a 30-day time frame. And I will also say that having 6 clicks, 7, 8, 9, 10 in a 30-day time frame is also not a lot as well. My point that I'm trying to make is when we think of non-converting spend, when we think of performance, Most of the time, these are search terms that have just a few clicks a month, making it really easy for non-converting spend to rack up. So all these little tiny things, one, two, three, four clicks, it's the source, I believe, of non-converting spend. Now, in practical terms, that same account had $9,900 of non-converting spend. And the reason nothing could be done about it was because the search terms, which have been anonymized, looks like this. Running shoes for men, one click. Men's running shoes, two clicks. Down pressed ringer only, one click. Hand closed ringer, one click. All these things had such few clicks, but when you rack up all the spend for them, the amount ended up being $9,000. Now, most of them were not show me all orders equals zero, clicks over 20. Most of them were not that. Now what N-Gram Analysis is, is a way for you to analyze this. So I can see in both of these search terms, they both have the word laundry in it. So I can begin to make decisions about all the search terms with the word laundry in it. So that is the essence of N-Gram Analysis. So this would be known as a one gram, where I'm looking at all of the search terms with that one gram in it, that word laundry in all of them, and I'm going and I'm taking action. At the same time, I can also see the word wringer is in both of these search terms. So like laundry wringer, I can pull in some additional analysis. That would be an example of a 2-gram. So that is N-gram analysis in a nutshell. Now, the thing that I wanted to do here with you is actually run through some examples. You can go and Do something called an N-Gram Ladder. And the reason why I had to give this a name, I had to give this something that you could wrap your head around, because a lot of times I would see people look at their N-Gram data, which we have a spreadsheet with, you can go and grab it. It's in the description of this video where we sort of catalog a lot of our spreadsheet resources that we've made. Simple things that you can paste your search term report in and get some really nice analysis outputted. So we've created one for spreadsheet analysis. And essentially, when people run the N-Gram, and I'll pull it up on screen here, just so you can sort of see what the spreadsheet does, when you open up the spreadsheet, this is what happens. So you paste it in. And what it does is it splits out the terms. So like historic Italian cream becomes three separate split terms. Simple, right? Then from there, it demonstrates to you all the 1-grams, 2-grams, and 3-grams. Makes sense, right? So I can see here all the search terms with the word cream inside of it generated $118,000 of revenue at a 33% ACOS, cost $40,000. Pretty good, right? So that's fantastic. Now, the thing that I often see people doing is like the one gram lip, 59% ACOS. So they go back to their search term report and they begin looking for all the terms with the word lip inside of it. And a lot of times people don't know what to do. So they've done this analysis. They've seen the N-Gram. They understand the concept, but the purpose of this, and this is piggybacking onto an episode that I did previously with Michael Tejeda just a couple of weeks ago. I wanted to put almost like an epilogue on it and really just round in what you actually do. So you do something called an N-Gram Ladder. So that's what we're doing today. We're doing something called an N-Gram Ladder, where basically, start at one gram. Find something interesting, and it could be interesting for a wide variety of reasons. It could be irrelevant. It could be incredibly relevant. It could be converting. It could be non-converting. It could be really low ACOS, really high ACOS. You find something interesting. And before you do anything, before you begin to take action, you want to boost your confidence. So what you do is you move to 2-gram, but you use that as a search into the 2-gram. So if I found the word lip on one gram, I'm going to go into two gram and filter for all the grams. With the word lip in it, and you'll see this all make sense over the next couple of slides, then you can go into 3Gram until you get a good instinct about what to do, looking for related ladders along the way. Because as you do this, different things will spark in you. So if you have not already, pause the video, go and get the N-Gram sheet. It's in a checklist. It's in this video description. It's also on the AdBadger website. So it's going to be easy to find in the description. Go to adbadger.com and find the N-Gram sheet. Now, The first process that we do is an irrelevant manual scan. So what does that mean? That means you look at one gram and the goal is to add new negatives, reduce spending on these keywords, potentially stop bidding for them, and worst case scenario, you add a negative phrase. If that is so bad, you would never want to appear in any variant as well. Now, when I did a one gram here, I found the word go. It had 126% ACOS, irrelevant, right? I can go and take some action there. But part of the N-Gram Ladder is taking that 1 gram, moving it to 2 gram. So when I take that word goat, and I move it up, and in this case, I moved it to 2 gram, and then I moved it up to 3 gram. Now all of a sudden, I see something interesting. Because don't forget, the 1 gram goat had like 130% ACOS. Don't make the judgment just on the 1-gram. Take it into 2-gram. Take it into 3-gram. So here I filtered for all my 3-grams with the word goat in it. And I actually see two direct hits that inspire me to bid on something. So Dionysus goat milk, Korean goat milk. You know, I see the word Korean here. So what I would do is I would find out why did I convert so well on Dionysus. And I can look at that and I can actually see that, oh, okay, that term is actually a good thing for me. So I can actually go and bid on a manual exact because don't forget the one gram I may have missed, but it just gives me some additional insight into something to go do. The other term that I see come up is the word Korean. You know, maybe my product has nothing to do with the word Korean. So what I do is I rerun that just for Korean. And what do I see when I rerun that for Korean? Well, I see that the word Korean was 84% egg cost for conversions. I go into 2-gram and I actually see some good stuff. Korean Lip is actually a great place to go. And then I jump up to 3-gram and I find that There is something to be said for this sort of Korean lip oil I converted on, Korean cream, South Korean I converted on. So again, this process of going up and going down is incredibly powerful. I started this sort of laddering by Starting at 1 gram, looking for something irrelevant, taking it into 2 gram, taking it into 3 gram, and what initially started as a way to find irrelevant things actually led me to something that I didn't even know about. So I found this competitor, Dionysus. Goat milk. Everyone's rolling their eyes. How did I mispronounce that? Dianus. Dianus? Goat milk skincare. So I found a competitor term that I actually convert well for and it just spurred some stuff in me that I can go and maybe target this competitor's ASINs. I can go and target competitor branded keywords because I had a nice conversion there. I found goat milk I actually did convert for. So what started, again, I'm just repeating this process. N-Gram is, I believe, one of the best ways to scan through your search terms. So again, I started this search by seeing something irrelevant, the word goat, and I decided to not do anything at the one gram level, but go up a notch. Again, I was scanning for irrelevant search terms and I found the word goat and it spurred on some action. So I took it. I jumped into two grams and when I jumped into two grams, I looked at all these 2-grams and I asked myself, let me just show me 2-grams with the word goat in it. And I found all these things and then I went to 3-gram. And I'm happy that I did because I found something really interesting and I was able to do something with it. So there are a bunch of direct hits when I look at it on the 3-gram level. Then I found the word Korean and I thought that was interesting. So I took it back down to the 1-gram. So I looked at all the terms with the word Korean in it. I went to two grams. I looked at all the words with the word Korean in it. I went into three grams, looked at all the words Korean in it, and it gave me such incredible insight into that Word that I never would have paid two seconds to because they all have very, very low clicks. It's a way for me to move through my campaigns really nicely. So I'm a big fan. That is the concept. Look at the one gram, take it into two gram, take it into three gram. As you're doing that, identify other things you'd want to go back and filter for at the one gram, go back into two gram, go back into three gram. So you're constantly moving up and down this N-gram ladder from one gram to two gram to three gram. In a way, that gives you a ton of information. Another thing to do is export your N-Gram list. So basically, I'm taking my one grams, and I'm using a really simple prompt in an AI of your choice, whether it be ChatGPT, or Claude, just drop it in there. And then what I see is A couple interesting things. I see exact matches, potentially new exact negative matches, potentially new phrase negative matches, potentially broad negative matches. So I can sort of just get another pulse on this. So, you know, I might see the word soap if I'm not selling soap. I have a bunch of different country-based terms. So if I'm not selling something French, I could really attack that. So I could see the word French. The word toe came up over here. So I saw these body parts. So I saw the word toe. Again, it just clued me in like, oh, I would have missed that in a really, really long list. Maybe it would have flown under the radar. I found the word toe as a one gram. But again, before I do anything, I take that word and I move it into two grams, move it into three grams until I feel really comfortable and confident with that. So that is the essence of the N-Gram Ladder. Start at the one gram. If you find something irrelevant, do not add it as a negative just yet. Take it into two grams, take it into three grams, which is great. That word toe that came up with that AI scan, I found a bunch of different things that actually did not work well for me, and I can go and take action, which is really nice. If I take it all the way up to three gram, I actually see some, it looks like I had a random conversion, maybe I wouldn't want to do anything about it, but it's just giving me insight into this that was really great. So that is the essence of the N-Gram Ladder. Start by finding irrelevant terms, do the manual scan, Do it with AI. Do another scan for orders equals zero. And again, when you run that orders equals zero, you know, I see the word vanilla, $47, no spend. Before I do anything, move it up to 2-gram. So if I were to search In this case, bum. You know, I would see all these things for bum, or I'll jump over to vanilla, and I'll jump over to vanilla, and it just sort of adds to my confidence level. Meaning, should I do anything about the word vanilla? The answer is yes. I can drill into exactly the word vanilla. I start at one gram, I see vanilla, $47, no spend. I take it into two gram, do I need to go any further? No, it just confirmed all my suspicion that vanilla is just bad, and I can then go ahead and negative I can just hit negative exacts on all the search terms that contributed to it, boom, and be done with it. And then I can meticulously add some negative phrases as well. The other thing that you might notice is when I'm scanning through this, I see something interesting. I see tobacco. Hmm, that's interesting. I haven't done a ladder on tobacco. So again, I started with vanilla. I went up to two grams. I noticed the word tobacco. I go back down and I spot another thing to negative phrase. Don't forget the whole purpose of an N-Gram analysis in the first place is to find things that were too small for you to notice with all these different search terms. It is also very likely that I would have missed the word tobacco if I wasn't doing this sort of ladder activity. Because if I did orders equal zero, this only had three dollars in spend. But by sort of doing this ladder where I'm going over here and I'm seeing everything that comes up to it, and I have so many ideas here. How does eco perform? How does How does butter perform? How does cinnamon perform? How do all these one-word roots perform? I have an endless amount of ideas just by doing this ladder activity, starting at one gram, taking that word into two grams, and moving up and down the ladder allows you to find things that it's a fantastic potential new negative phrase, which I absolutely love. Of course, sorting from high to low A costs as well. So again, starting from super high ACOS, so these are my one grams of really high ACOS. So I see the word scrub come up. I see the word tinted come up. And again, I see the word tallow came up over here, and I can then take that into the next level up and actually see something that might be relevant and things that might not be relevant. So I found the word kids, right? Completely irrelevant. Tinted, I take that and I take it back down. So you're constantly just moving up, you're moving down, To find new opportunity if something does have a really, really low A cost, as well as new opportunity for negative keywords. And that is The N-Gram Ladder. You know, I've talked a lot about N-Grams over the years here on the show. I get lots of messages almost weekly about how the N-Gram spreadsheet has helped people. So I do invite you to go and download that spreadsheet. We have it in our description. Download the spreadsheet and now hopefully this episode is giving you more oomph and confidence about how to get in there. And again, when you jump into your N-Gram analysis and you begin to sort of sift and sort around, High ACoS to low ACoS. Orders equal zero. Low ACoS to high ACoS. Irrelevant terms. When you begin to do this and you look at that and you need more confidence, you then take that information and you move to the next gram up. You take that into the next gram up until you feel very comfortable and confident with it. Sometimes you will see things just like this example where the word vanilla is just bad down the line. And when you find something bad down the line, it gives you a lot of confidence. At the same time, look at the words around that, right? So if vanilla was so bad, I actually want to research all the two grams where the word vanilla isn't. And I have a whole slew of terms that I would want to research around this. So it's such a useful tool to zoom in, zoom out, move up the ladder, move down the ladder, find something new as you're moving up, take it back down to one, Take it two, three, find something interesting, take it back down to one, and repeat that process until you have true mastery over your search terms. I think N-Gram Analysis is the way to really wrangle in non-converting spend. Hope you enjoyed this episode. Have a good one. Let me know how you're using it. I'll see you next week here on The PPC Den Podcast.