Lazar: (00:00)
Hey guys, welcome to the Wild PPC Bunch podcast. My name is Lazar and I’m a PPC nerd. I have over 10 years of experience in online advertising. And currently I’m the owner of the growing Amazon advertising agency called Sellers Alley.

Brent: (00:14)
And I’m Brent, the owner of AMZ Pathfinder. I started this company five years ago and we’ve been working in online advertising since 2013.

Lazar: (00:21)
Every week we will spend around 30 minutes covering one topic and it will get nerdy, I promise. We’ll prepare a topic, covering everything from PPC basics, in -depth strategy and current trends.

Brent: (00:32)
One thing’s for sure you won’t be bored and you will hear insights, tactics and ideas straight from two experienced agency owners, so strap in for the ride and enjoy,

Lazar: (00:50)
Hey guys, today, we’re talking about one interesting topic and it’s keywords – when to give up and when to reconsider keywords.

Brent: (01:00)
Right…So Lazar we should probably dive in first to what is giving up mean and what do we mean by keywords? Let’s like define everything pretty clearly before we, before we dive into the action items or, or the topics.

Lazar: (01:12)
Yeah, definitely. Where would you like to start with that? Everybody knows that keywords is the main way, how your target when you create ads. If you’re a first time in Seller Central, one of the first things that you see is, for ages, to create automatic campaigns or to create manual campaigns and in them you have some keywords. So can you tell us a bit more for people that are just starting maybe?

Brent: (01:38)
Keywords would be, uh, in different match types too. And also I think for the purposes of this discussion, we’re also covering ASINs, right? I mean, we’re not just talking solely about keywords, like blue shoes, right. We’re also talking about ASINs cause those can be set as negatives and you know, those have the same, for example, like bid adjustments that we, that we have with keyword center, uh, you know, words, you know, key words. Uh, but ASINs also count and on the same discussion.

Lazar: (02:07)
Yeah, definitely. You can use them as, as one of the targeting methods. So I would focus on ASINs as well. So basically, let’s maybe start with, how would you structure your campaign on your account when it comes to keywords? Like where first to start with everything? You, you want to advertise something and you want to make it like in the best possible way. So you optimize spend, you go to some tool where you place your ASIN, you do reverse ASIN research. There is whole bunch of tools that people can use, one of them is Helium 10, obviously. But you can use literally whichever you wish and you get whole bunch of keywords or at least some words that people are using when they’re searching for products. And one of the things that you can usually see there is some kind of search volume and competitors or stuff like that. And you end up with lhundred or 200 or 500 keywords and what to do with them?

Brent: (03:12)
At that point, I would suggest taking that list and filtering it as aggressively as possible because one of the key considerations for a keyword or a search term that’s, uh, you know, uh, based on an ASIN, I suppose, would be relevant, right? Relevance is the main thing we’re centering this discussion around because I think the assumption we’re making, going forward in the discussion is we’re talking about relevant keywords, right? So there was a time on Amazon maybe some years ago where you could dump 500 keywords in an ad group and, um, just let it run, but that’s probably not a best recommended practice anymore. So we want to make sure that the ones we’re running are as relevant as possible, you know…They’re closely connected, they’re related to the product or the category. It has to make sense.

Lazar: (03:59)
Yeah, definitely. One of the things that we currently suggest and what’s our best practice at this point is not to have more than 30, 40 keywords per ad group. And one of the most important things that you need to know is not to mix match types. That’s one of the most common mistakes that happen with whole bunch of sellers. So it doesn’t matter if they’re big or small, this is like pretty common mistake…

Brent: (04:24)
And software for that matter. As we talked about recently off the podcast, you and I had a conversation about that…

Lazar: (04:31)
The thing is like, basically.. Let’s explain why is the problem to have all the match types in one ad group? The main reason that we find is not being able to narrow down the search funnel and optimize spent, because when you have exact phrase and broad in one ad group, you end up having all of them triggering whole bunch of stuff. And when you check your search term report, when you see what triggered certain search term in ad group, you want to negate it. And when you negate it and if it’s exactly the same as the exact match type keyword, you’re going to block it. And that’s why you’re, that’s how you were basically killing the keyword. And also if you have exact, phrase and broad all in the same ad group and you want to push some variations of that main keyword and you increase the bids in phrase and broad rather than in exact form. Exact form has the lowest bid. You end up triggering that exact form from phrase and broad, just because you have higher bid there. And when you want to negate it just like not to trigger that exact keyword, you place negative exact, and you literally kill your exact keyword. And that’s one of the huge problems that, that, that happen, that people usually do when they structure campaigns in a bad way.

Brent: (05:58)
Yeah, the structure is kind of malformed like that. You do end up with some confusing logic puzzles. You have to sort out every time you want to go and optimize something. And it does make setting negatives very difficult, because you know, you might have some weird overlap and block things out by accident. Um, and then I think bidding is probably another major reason for that too, if you want to use some kind of tiered bidding system based on match types. Um, but also just for like easier legibility into the account. And optimization through software or through bulk files and manual, uh, through the Seller Central console interface, whichever one you’re using. Um, so yeah, I, I wouldn’t mix those up too much. Yeah. So I think that’s a, that’s a pretty good primer on some of the basic structure that we think is, is good, but at the end of the day, uh, you know, what is giving up mean? When, when we say like, we want to give up keywords, give up on them, what does that mean? Does that mean we, uh, we don’t call them again. Like what, we just drop them off at the bus stop and drive away? Like, what does it mean Lazar?

Lazar: (06:56)
Well, depending on who you ask, like the worst thing that you can do is archive them. I think that there is no, like the other thing that you can do as bad as that. The softest way, how to do it is basically play super low bids, 2 cents or something like that, 10 cents and like keep them that way. But one of the things what we like doing is, uh, when, when you like.. It really depends on your search terms that are triggered by that keyword. So if you’re having problems with phrase and broad for example, and not exact form, like the exact keyword is super relevant and it’s, um, really good when it comes to rate and like convergence is profitable and you have phrase and broad that are not doing that good. That’s probably because they’re triggering whole bunch of different stuff that you don’t want to show up for. And one of the first thing that you need to do is to go to your search term report, check all the search queries that people are searching for and that are triggering your ad for that specific keyword. And if you see that something is irrelevant, obviously you are going to negate it, but whole bunch of people are having problems like what to do with something that is highly relevant that is not your keyword, but it’s triggered by your phrase or broad keyword and it’s not making any money or it’s some profitable. Like, I didn’t know what you guys are doing. Maybe you can share it as well, what we usually do, we negate it there. First we check if we have that specific search term as a keyword somewhere. And if we do, we negate it in a current campaign where it was triggered, but if we don’t have it, we also negate it, but we create some campaign that we like to call non-profitable, non-converting keywords and we place those keywords there. And those search terms are becoming exact form keywords and they get way lower bid than the previous version of keyword that triggered that search term. So basically we’re trying to optimize that search term in the manner that is going to make some money, unless like we try with like higher bid. And over the time we decrease it until the point when it works. So we don’t really give up, like right away, we really try to fight for it for each search term.

Brent: (09:31)
So you isolate it and then you kind of ease off the gas slowly to see what direction it takes. Cause there may be, and this is the case for any keyword, even irrelevant ones. There’s a point where the bid will be profitable. But the only question is, is that bid still gonna make you eligible to be entered in the ad auction. Because yeah, you could say you’re selling blue shoes, you could be advertising for, um, I don’t know, like kitchen tables and yeah, for 1 cent, it might make sense, but you’re not going to be shown and you won’t be seen as relevant. Um, even if people are buying something that’s totally different, you might get a sale once every six months based on. So yeah, I’ll come back to that. But like I wanted to say for, for giving up, um, on our end, like I agree with you that the most harsh version of giving up on a keyword would be archiving it, right. That’s basically killing it. That’s saying like never again. Um, but a less harsh version would certainly be, uh, just simply pausing it or putting the bid down to some level where it’s barely going to get any exposure, which I know is a fairly popular methodology you just described. Um, and then I guess another more extreme version of giving up would be setting it as a negative. So you’re actively saying we no longer want this to appear here. You know, which is, which is even more aggressive than pausing it, I would say, because you could pause a keyword and it has a bunch of search firms behind it, but setting one of those search terms as a negative, it’s ultimately going to nip it in the bud from ever appearing anywhere. At least, at least in that ad group or campaign. Right. And I like the idea of migrating them to a non-performing campaign because, you know, they might not work now, but they may have a day in the sun later on. Um, you know, depending on seasonality competition, state of your competitors, uh, businesses…Like for instance, we saw a lot of companies that had, uh, competitors that stopped spending so heavily during COVID and the kind of like that, that period of crisis there, uh, in like May, April, Marchish period And so we are able to actually get clicks cheaper for a lot of terms because a lot of other people have dropped out of the auction. So maybe there was a time period where some of the keywords we had, you know, quote unquote given up on, actually might’ve worked again during that period because the bid level was more sustainable. However, that has seemed to have passed now, you know, our, our CPC is looking at recent data, um, just today, actually in our, in our software, like, you know, things are a fixer backup. Like where are they roughly speaking where in the U S maybe a little bit lower, but certainly click through rate and other things are back to normal, uh, pre pre COVID. So that’s what giving up would mean to me. And I think we can’t have this conversation without talking about search term duplication- deduplication, which is exactly what you mentioned. Like if you see it’s appearing somewhere else, then maybe you can set it as a negative or pause there and let those impressions drift over to that place for that search term is still appearing. Um, and so that’s effectively giving up on, keyword, but just redistributing that exposure somewhere else.

Lazar: (12:30)
Yeah. And also when it comes to all of this, one of the main things that you can discuss about is talking about, uh, what’s really unprofitable keyword. Like when do you start considering a keyword not to be profitable. It’s more or less about the strategy, how you think about, um, about as keyword, where do you use it, are you using it in your product launch or not…Is it like for, um, for cash cow product? Like where, where you’re just like milking the product and making every other thing profitable or if the product is dying or..Whole bunch of different stuff are there tobe considered, when you think about what’s the profitable keyword. So basically when you add a new keyword and when you see that you get, you got like three clicks and CPC super high, and like it’s constantly $10 to get to that point. Is it really unprofitable at that point? What usually people don’t think about is considering different time frames. People tend to stick to one time frame when they’re checking stuff, but you should always think about bigger picture..Or certain period of time what happened and like to filter out after or before change that you did to that certain keyword, because it can be affected by so, so many things. Like with negative keywords and so on. You shouldn’t consider that a keyword is going to perform the same after you negated some other keywords, added some negative keywords to the same ad group or decreased bid, or combined those two together. And you have your keyword in, let’s say phrase and broad. And, uh, it, it basically has like just a couple of clicks. So what we usually think about there is, what’s the main strategy, are trying to push that keyword for ranking or you’re willing to spend more money. But at that point, I’m sure that you’re fully aware of, and you’re also using the keywords to, to try to improve some ranking. Do you also see sometimes when you are under the niche average, when it comes to the CTR and conversion rate, do you see that those keywords can work negatively basically for the product?

Brent: (15:03)
Yeah, that’s a great question. That’s one of the things we have in our notes here, right? It’s like do push campaigns, you know, so-called push campaigns to push your product up the ranking or for a specific keyword or small set of keywords. Do they actually work? Cause I feel like there was a time on Amazon when you could throw just hundreds of dollars a day at some keyword. And just because you’re getting the sessions and clicks, you’re gonna drive up your relevance in search ranking for that. Uh, I feel like these days it’s much more focused around conversion rate, um, primarily, but also click through rate and like relevance. Um, and so, you know, if you’re trying to effectively build a sales history against a search term slash keyword, uh, and doing that through advertising, if you do have a good, uh, you know, percentage conversion, click through rate, I think that’s helping you. Right. But can it adversely affect you? Uh, I would say, I would say yes, actually, if you’re really forcing it, what you’re doing is pushing too hard on the spectrum of, uh, you know, of relevancy and costs and, and you know, it’s not as relevant as it maybe should be, but you’re throwing tons of money at it. You’re going to end up costing yourself. And with the way the ad auction works, you can actually end up spending less than someone else for the same keyword and still win a better position if your relevancy score is higher. Right. You know, we don’t have visibility into that relevancy score, at least at this point or any way that I know of, but it certainly exists. Right. It’s part of the it’s part of the, uh, the way it works. So my feeling is that, yeah, you can adversely affect yourself, not only from a bottom line pocketbook perspective, spending too much money, but actually harming your organic rank by trying to fit a square peg in a round holes so to speak with a keyword, that’s not really a good fit and relevance.

Lazar: (16:46)
Yeah, that makes a lot of sense. So one of the other things like that we usually think of is when it comes to keywords.. When you have, but it’s not always like unprofitable, but non-working keywords, one of the things that we usually see in campaigns that have more than 50 keywords in them and when you expand your lifespan of that campaign, you see, you see that, that keyword didn’t work at all during that time period, like for the lifetime. What we usually do, we take it out with separate campaign and try to treat it differently. I don’t know why, there is no logical reason for it. I’m pretty sure that you cannot read it anywhere. Like why the keyword is not working and it’s highly relevant, your indexed for it and you have like 200 keywords in one campaign and that one is not working. But when you place it in a new campaign, it starts working. Literally I have no idea why, but it’s working for us. We are not planning to change it…

Brent: (18:01)
You’re saying you basically transplanted out of where it is? This is similar to the idea you mentioned before, like a non-performing campaign, you put them in a separate area where they have a separate budget allocation, it’s easier visibility, maybe at the top level. Um, or maybe you software even to like apply a different logic to how those bids are managed and that’s where it does well, or you’re saying it’s different strategy?

Lazar: (18:22)
No, no, no. It’s just like you…Yeah, literally like, like you mentioned, it’s the opposite way. Just place it to a new company. You can even place higher bid if you want to push it a bit more and it’s going to, it’s going to work basically. As I said, I didn’t know why, but it’ s working. And also when it comes to long tail keywords, there are a whole bunch of long tail keywords that are highly relevant and they’re not having that much traffic. Like, would you include them in, in your campaigns or in your account or not?

Brent: (18:56)
Yeah. There’s two ways to think about that because you could say let’s exhaustively research, every single long tail keyword that exists and let’s put them all in a, in like a tight match play, like exact match, right. Let’s go with the more extreme scenario it’s like, all right, this has five or six words. So just based on that length, you know, it’s going to be probably a long tail keyword. That’s going to get searched not very frequently, but maybe that search term keyword is super relevant for your product and audience or whatever it may be, but you only get a couple of clicks on it a month or something like that. They result in good sales. Uh, that becomes really difficult at scale. You know, you can do that really easily for three products, three SKUs, right. But then you have a hundred SKUs that becomes immediately a much harder, and that’s why match types exist you know. You could use phrase or broad to try to predictively, you know, jockey for those impressions and for those keywords. Um, but you also run the risk of course, of catching a bunch of other stuff in that net. So really, it’s a, it’s a balance between time and having the, I guess like, uh, you know, foresight tools to predict what is going to be a long tail. And I definitely fall more in the second camp where yeah, it’s helpful to put long tails in, but this is why we have brought in phrase. This is why they exist to catch these things that we otherwise couldn’t come up with. I have no idea what the stat is for Amazon, but I know on Google, like some huge percentage of searches that happen every month or like searches that have never happened before. So, you know, we can’t predict the future, but that’s why match types exist…

Lazar: (20:29)
Yeah, that makes a lot of sense. Like we also like to play some long tail keywords, not all of the long tail keywords in the universe.

Brent: (20:38)
Right. You get diminishing returns at some point.

Lazar: (20:40)
Yeah. You can end up like creating huge accounts with whole bunch of different stuff and a whole bunch of, um, campaigns. And you can start getting lost in, in, in that kinds of campaigns, in that kind of account with, with limitless amount of campaigns and all that.. Just losing it and having like 50% of them not working, is it really worth it, I don’t really think so. Especially as you said, there are phrase and broad keywords that are basically doing all the work. So there’s no need for that many long tails,

Brent: (21:17)
They’re going to do the heavy lifting.

Lazar: (21:19)
Yeah, we call them safety nets, basically like our automatic campaigns.

Brent: (21:23)
Oh, that’s cool. I like that, you call it a safety net campaign?

Lazar: (21:27)
So basically that’s, that’s how we place it. So, um, yeah. Oh, one of the things that we mentioned at the beginning of this episode is ASINo, and that’s something that is super relevant to keywords. And basically, uh, that’s something that you use as well when it comes to targeting. Uh, Oh, there is one question that everybody wants to hear the answer and that’s like when you have automatic campaign and when you have ASIN in a search term, report, can you negate it Brent?

Brent: (22:01)
The dreaded, the dreaded ASIN in the auto campaign, the most awful thing. Yeah. So there’s no way to set those as negative or positive, right? Not that I know of.

Lazar: (22:13)
Yeah. Like when you place it as a negative, it’s not going to work, but..

Brent: (22:18)
It doesn’t do anything. Right? Has no effect. But you can do that in the product targeting. So, you know, if you have an ASIN that terms up, even if you’re using category and it like pops up in there, you can set that as a negative. No problem. So we typically will try to shift a lot of the spend for ASINs over to the manual campaigns. And in the case of the autos, we’ll turn off those targeting types. Although we have seen that, I think it’s close and loose match. Those are the ones that are the keywords and it’s the substitutes. And I’m blanking on the other one. There’s four different types right now. We sometimes see..

Lazar: (22:50)
Sorry for interrupting. Are they working?

Brent: (22:53)
Yeah. Yeah. Well, we turn off the ones ..ASIN target sometimes, not all the time, cause sometimes it works really well. But other times we just switch that targeting type off and then that eliminates most of the ASINs. And those are transitioned over to manual campaigns where we have more control.

Lazar: (23:07)
Yeah, the last theory that I heard is you cannot negate your ASIN in automatic campaign But if you create, um, manual campaign that is targeting ASINs, and if you negate that ASIN, your account is going to learn about it, and it’s going to show that ASIN less in your automatic campaign. To be honest….

Brent: (23:32)
Ooh, I like that. That’s like Amazon PPC, conspiracy theory.

Lazar: (23:36)
It’s literally conspiracy theory. I was like, I’m not really sure that it’s true since you cannot read it anywhere, I’m still going to do it. I’m going to try.

Brent: (23:48)
I know people that set negatives just so they can um, one day when Amazon enables it, they will have it in there ready to go, which is an interesting idea too. Cause I’m not sure if that’s literally ever going to happen. You know, we don’t have word from Amazon on that, but that’s an interesting tactic as well.

Lazar: (24:04)
Wait. Can you explain it a bit more?

Brent: (24:06)
Sure. So in an auto campaign specifically, they will go and set ASINs as negatives. Not because they work, but because with the hope that one day they will work. Let’s say four months from now, Amazon says, hey, this is a new feature and then, well, you’ve already done all the hard work of setting this negatives in your auto. So that’ll turn those, uh, you know, turn on so to speak.

Lazar: (24:28)
Yeah, that makes a lot of sense.

Brent: (24:28)
Speaking of negative keywords, one of the things we have in our notes here, um, setting a schedule for removing negatives or unpausing keywords. So yet again, this discussion is predicated on the idea that we’re talking about relevant ASINs. So when you finally do need to set one as a negative and say, all right, we’ve had enough of this. We’ve given it months of spend and enough clicks over the time. You know, it has enough statistical validity. We’ve looked at it through different time frames. We’re just going to set it as a negative. It’s relevant, but we got to set it as a negative. At what point do you go and review that? Cause you have to revise it and say it might work again.

Lazar: (25:09)
Yeah, definitely. Like that’s something that we do occasionally. There is no any specific rule when it comes to that, like was the best time to revisit your keyword. There, there is no manual for it. So basically what we usually do when we feel that we need to expand a bit further, that we want to try to catch more traffic or so on. We try to revisit paused keywords, and also we try to revisit negative keywords as well. Because, um, during one part of the season or like during one part of the year, it might be working and the other one it won’t.

Brent: (25:46)
Right. It depends on whether, competition, state of global pandemic, whatever, there’s all kinds of influencing factors that could affect it.

Lazar: (25:54)
You can literally name 8 or 10, like just out from your head right away. And there are like new competitor or like somebody’ willing to push a bit more when it comes to product launches and it’s affecting you on the other hand. And it’s making overall expense for, for that specific keyword a bit more. That’s something that I can compare to, to something that I did maybe almost 10 years ago when it comes to Facebook ads. I did it in Serbia where I started my marketing journey. Um, and there were elections here and there’s just one amount of space on Facebook. Like, especially at that time, 10 years ago. And like when the election started, there were like only political ads and that kind of stuff and they made all of the clicks so expensive. Like..

Brent: (26:53)
That might be the case in the US very soon on Facebook. Yeah.

Lazar: (26:56)
Yeah. So that, that’s something that, if you’re doing some external traffic that you should be aware of. I remember the last time, like we shouldn’t go political, I don’t want to go political, but politics can change the price of advertising on platforms.

Brent: (27:12)
Right, right. Yeah. Yeah. Without, without, you know, and not understanding some of these larger conflating factors can lead you to really scratch your head. So yeah. It might be that your product seasonality is obvious, like, okay. Summer is when we sell these inflatable beach toys. Right. Obviously winter, no one’s buying them. But at the same timevmaybe there’s some other competitor that’s new that jumps in the market that has deep pockets and they want to fight their way to page one for all these big keywords that you’ve already been bidding on for years. So maybe you’re going to have a rough time suddenly you could probably do a whole podcast just about these kinds of factors. But I think that’s probably good for the scope of this one.

Lazar: (27:48)
Yeah. I think like from time to time, you should have some, some budget for AB testing. And part of the AB testing should be like revisiting keywords or revisiting negative keywords, just to double check if things have changed. Because that’s more or less like shooting in the dark. You don’t know what’s the current situation until you try it. Like imagine that you’re using for the first time and basically that’s it. And like, I’m not even sure if you should unpause the same keyword in the same campaign or create a new one. I would probably create a new one just to be absolutely sure that it’s isolated. So I know where the spend is coming and all that, but I didn’t know. What, what would you do basically?

Brent: (28:33)
Well with, we do this with software specifically, but being able to see where search terms are appearing across an account and making sure that yeah, if you are a pausing it somewhere, you’re effectively, you know, we’re saying giving up, right. If you’re giving up on it, then it’s not still appearing somewhere else. Um, so you would use software or just, you know, Excel and then reports from Amazon to determine that. But yeah, you, you might be right. Maybe it’s time to put it in a totally new, uh, campaign where it’s like a fresh start on a fresh time of year, for instance. Um, and so it’s easier to keep, keep tabs on what it’s doing or maybe you change the match type, um, for that matter. So either, either approach might be valid.

Lazar: (29:12)
Yeah, makes a lot of sense. We’re almost 30 minutes talking and one of the things that I wanted to mention is day parting for campaign and keywords. And I’m not really sure that we should even scratch that topic because you can deep dive into it. And like whole bunch of people are having different opinions. Is there anything else that you would like to mention about keywords and like revisiting them?

Brent: (29:38)
Uh, one thing maybe we didn’t touch enough on is statistical validity, right? So not just making a decision based on information you got from the past five days instead saying like, all right, what’s 60 days, what’s 30 days and not only days, but amount of spend, right. The faster and more you spend, the more you’re collecting data. So it doesn’t matter if it’s 30 days, if you can only spend five hours. So you want to make sure you have enough money through it to have enough data. And, um, you know, that there hasn’t been some massive changes to the product page. Like for instance, you were out of stock or like your price went up by a bunch of dollars or something when you increased it. Um, so consider those factors too when, when assessing these keywords and search terms, um, don’t just give it like a couple of days and then call it quits. Um, you want to have it over a long enough time with enough data with enough spend and then you can make some, some high quality decisions.

Lazar: (30:29)
Yeah, exactly. Exactly. The amount of impressions is also important. Like I find it super important because if you have a lot of impressions and you don’t have enough clicks by comparing to other keywords that might be telling you something about a specific..

Brent: (30:47)
Right, what is the account benchmark there, that’s a good leading indicator. Yeah, absolutely.

Lazar: (30:52)
That’s it for today’s episode, if you guys have any questions for us or if you would like us to talk about some specific topic, or if you have any questions, feel free to send us an email on ask@wildppcbunch.com or you can reach out to Brent or to me.

Brent: (31:09)
Sounds good. I know we’ve got some emails recently, haven’t we?

Lazar: (31:12)
Yup. Okay. Well that would be it for today’s episode and talk to you in the next one. Have a good one.

Brent: (31:19)
Farewell.

Lazar: (31:19)
Bye.