YT YouTube2Text

YouTube2Text → Transcripts → How Experienced Marketers Work Differently With PMax and AI Max | PPC TH 124

How Experienced Marketers Work Differently With PMax and AI Max | PPC TH 124 — Transcript

by PPC Town Hall · 8,009 words · 1,164 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00Hello and welcome to another episode of
0:02PPC Town Hall. My name is Fred Vales.
0:04I'm your host and I'm also the CEO and
0:06co-founder at Optimizer, a PPC
0:09management tool. For today's episode, we
0:13have one of the most influential voices
0:15in PPC, someone who's got a big opinion
0:19on Performance Max, and he has a big
0:21opinion because he's been using it very
0:23extensively. He knows all the ins and
0:25outs. And so today we're going to talk
0:28to Manahim Ani from JXT Group and he's
0:32going to share how he's been using
0:34performance max as well as using other
0:36campaign types and basically getting to
0:39the roots of how to use all of these
0:40Google ad systems to get the most out of
0:43it for your business and to make sure
0:45we're not just shooting for vanity
0:47metrics or looking at inflated rorowass
0:50but actually getting results that makes
0:52clients happy and makes them uh keep
0:54coming back for more from agencies. So
0:58with that, let's get rolling with this
1:00episode of PPC Town Hall.
1:06>> Hey, Manam, welcome back to the show.
1:08It's great to have you on.
1:09>> It's great to be on. It's always good to
1:11talk.
1:12>> Yeah, last time we talked, I was in
1:15person for the the Google Marketing
1:16Live, I believe.
1:18>> Yeah, that that's always a fun time.
1:20Always exciting to hear the new things
1:22and try them out.
1:24>> Yeah. Well, hey, let's uh let's jump
1:26right into that. So, I know we're going
1:28to talk a lot about performance max, but
1:30the new thing since uh last year is AI
1:34Max. Is that something you've had a
1:36chance to play with?
1:38>> Yeah, we've been we've been playing
1:39around with it since it's been in beta.
1:41I kind of like it. Um, it really does
1:44help you expand your search campaigns,
1:47but I like my recommendation is use it
1:49when you would use Pmax because it it's
1:51a I think a lot of the same
1:52underpinnings, you know, taking the
1:54pieces of search and shopping and
1:55bringing it to your search campaigns
1:57where it makes sense,
1:59>> right? And so, um, you know, to to maybe
2:02give people a little bit more context
2:04for those of the listeners that haven't
2:06met you before, but uh, where do you get
2:09all your experience? What does JXT Group
2:11do? Sure. We're we're a digital
2:13marketing agency. We manage campaigns
2:15for midsize clients. Typically, it's a
2:17mix of lead genen and e-commerce. So,
2:19you know, like we we learn the best of
2:21each discipline and kind of apply it to
2:23each other. So, a lot of times we'll try
2:25things from e-commerce for lead genen
2:27and lead genen and e-commerce, you know,
2:28try different strategies and, you know,
2:30really figure out what works for our
2:32clients.
2:33>> Okay, that's great. And so, uh, one of
2:36the things I know you've been writing
2:38and talking about quite a bit is when
2:41clients come along, maybe start with a
2:43new agency, maybe start advertising for
2:45the first time,
2:46>> uh, the expectations can be quite high,
2:48right? And I think part of that maybe
2:50comes from the fact that everybody's
2:52doing AI searches now
2:54>> and AI likes to stroke people's egos and
2:57maybe make it sound a little bit easier
2:58than things really are in life and uh,
3:00and so they come in with huge
3:02expectations. But what's the reality
3:04that you see when people start with a
3:07campaign say like Pmax?
3:09>> The reality that we're seeing is that
3:12today the ad platforms are a lot more
3:14algorithmic than they've ever been,
3:15which just means that it takes longer to
3:17get started because you used to be able
3:19to set up a Google search campaign and
3:21just capture high intent traffic and go.
3:23when and you can still do that, you
3:24know, with with a very granular setup,
3:26but if you really want to build a strong
3:28foundation for growth using the AI tools
3:31and be able to scale past like you
3:33really got to show the algorithm what's
3:35valuable for your business. So, it can
3:37take a bit of time to set up the right
3:39tracking, the right keywords, the right
3:41ad copy, get the conversions rolling,
3:43and then really build on it. So, I think
3:45I just tell people patience more than
3:47anything.
3:49>> And and let's go a little bit deeper on
3:51that, right? So, is that patience
3:53related to a specific amount of time? Is
3:56that related to a specific number of
3:58clicks, impressions, conversions?
4:00Where's that sweet spot that you see?
4:02>> It's it's probably a mix of all the
4:04above. Like, you're starting out, you're
4:06looking at the soft metrics in the
4:07account. You're looking at at your
4:09impressions, your clickthroughs, your
4:10cost per click, you know, all that
4:11stuff. You want to make sure that you're
4:12spending your budget, you're getting the
4:14right traffic, and you know, as it as it
4:17starts to scale up, obviously you need
4:18to make sure leads come in or sales come
4:20in and and you're hitting your your
4:21KPIs. Um, but like you know, the first
4:24thing is is traffic. A lot of times if
4:26you start a campaign, jump in too early
4:29to max conversions or conversion value.
4:32The system doesn't know where to go.
4:34Sometimes it works well out of the gate,
4:36but a lot of times we find start off in
4:38a more manual way, forced data through
4:40the system and and it'll it'll learn
4:43faster from that.
4:45>> Right. And you're specifically referring
4:47there to the bid strategies that you
4:49would select.
4:50>> Yeah.
4:51>> So so you're saying basically would you
4:52go manual CPC maximize conversions
4:55before switching to more of a rorowass
4:58or CPA limited model?
5:00>> Yeah. So, I think depending on the
5:01campaign type, like for lead genen, I
5:03like to start out with manual CPC and
5:06and phrase match, just let it get like
5:08very good quality traffic through,
5:10measure some good conversions. With
5:12e-commerce, I'm a little less cautious
5:13just because this system will get
5:15revenue and it'll know. So, I'll start
5:16out with like a max conversion values or
5:18something like that. Just let it jump
5:20right to it and it'll learn a lot faster
5:23that way.
5:24Hey, [snorts] and uh I'm I'm happy to
5:26hear someone say start with phrase match
5:28because I'm kind of a fan of phrase
5:29match too and I take a lot of flack for
5:30that, but like what's your thought
5:32process around liking phrase match?
5:35>> So, it does depend on the industry and
5:37the volume. Um if if it's a lowbudget
5:40campaign, definitely start with phrase
5:42match because you want the best quality
5:44clicks, but you know, as you scale it,
5:47it'll restrict you. it'll hold you back,
5:49you know, and so you have to layer on
5:51broad match, AI max, pmax, all that
5:53stuff if you want to scale beyond a
5:55certain point, which is important for
5:56many businesses,
5:59>> right? And so I I think the nuance there
6:01is that exact match is obviously too
6:04specific for a new campaign. You can put
6:07in exact matches once you see the search
6:10terms data coming in and you find your
6:12winners and losers and you use exact as
6:14negatives or positives to to you know
6:16maybe bit differently maybe different
6:18ads for them.
6:19>> Uh but but broad match I suppose can
6:22also be a little bit overly aggressive
6:25right in the beginning. So it's it's
6:26basically telling Google, hey, go and do
6:28anything you want. Whereas phrase match
6:30is still kind of like that sweet middle
6:32spot,
6:33>> which gives them a little flexibility,
6:35but not too much.
6:36>> Yeah. And especially in industries where
6:38you have very nuanced keywords like for
6:41newer accounts as well, like I
6:42mentioned, if the if the algorithm
6:44doesn't really know what a good
6:45conversion is for you, the broad match
6:47with max conversions can really go very
6:50wide. Whereas if you start with, let's
6:51say, phrase match, it won't go as wide.
6:53let the system understand what's what's
6:55a good conversion and then add broad
6:57match later. I find that that that does
6:59better.
7:00>> Yeah. Now you you said good conversions.
7:05Um
7:07explain a little bit what constitutes a
7:10good conversion and to what degree uh
7:12would you focus on quality versus
7:14quantity?
7:16>> Sure. So for any any lead genen
7:18campaign, the number of leads is is less
7:22important as the number of leads I can
7:24sell a service to. So if if you're
7:26getting, you know, a ton of leads, but
7:28your client is not able to close any
7:30deals, that's really what we're trying
7:32to stay away from with all of this. You
7:34know, when it comes to e-commerce,
7:35again, because we we're tracking
7:37revenue, I'm less worried about all this
7:38stuff, and I'm much more willing to rely
7:41on the algorithm and the broad targeting
7:43to to get where I want to go.
7:46Now, I I think it was you, but we were
7:48at Google giving feedback to a bunch of
7:51the product managers at Google Marketing
7:53Live, and the point of
7:57understanding the business that we're
7:59working with came up. Um, and so we as
8:03marketers, we can only do so much,
8:04right? We can basically direct someone
8:06to your landing page in most cases, but
8:09what happens at that point is really up
8:11to the business. So, so to what degree
8:12do you think businesses are doing a good
8:15job with the traffic that we send them?
8:17Um, and what's maybe the biggest mistake
8:19that you see them make that they could
8:21improve on?
8:22>> So, that's something I think about a lot
8:24because, you know, we'll we'll launch
8:26two campaigns that are very similar for
8:28two different companies and one will do
8:29a lot better in that regard. And like we
8:31do see a huge variance in close rates
8:34just based on the intake process or how
8:36fast they they're calling leads. Like
8:38you, you know, we have one client, he'll
8:40call leads within 30 seconds any time of
8:42day and his close rate is a lot higher
8:44than anybody else because he's
8:46aggressive and he runs after it whereas,
8:49you know, so that it's definitely a big
8:51piece and and it plays a huge part of
8:53the success of the campaigns, but it is
8:55out of our control. I try to give, you
8:57know, guidance and advice to people to
8:59make sure they're following a good path.
9:01But I think the biggest piece we could
9:02do as marketers is make sure that the
9:04leads we're delivering are sellable that
9:06they they can turn into a paid customer
9:08for the client,
9:10>> right? And then what tips and tricks do
9:13you have as far as setting up the
9:15conversion data pipeline so that it
9:17feeds back to us? Because ultimately, I
9:19mean, we we could be sending them leads
9:21and
9:21>> have a very poor understanding of what
9:23actually constitutes quality. And so, so
9:26we could even look at um a lack of a
9:29close rate and then start to assuming it
9:31is due to us having poorly targeted when
9:34in fact it is maybe the fact that they
9:36didn't pick up the phone to call that
9:38lead
9:38>> for two or three days and that's why
9:40they didn't close because they'd already
9:41bought from someone else. So, so how do
9:43you make sure that that data quality is
9:45high?
9:45>> So, it definitely comes from both
9:47places. The first part is obviously the
9:49search term reports, making sure that
9:50you understand the client's service,
9:52what what they're selling and that your
9:53keywords really line up, but then you
9:56know setting up offline tracking where
9:57it makes sense. I think for smaller
9:59budgets, it's not really worth the time
10:01and effort because you need I think at
10:02least 30 over the course of a month. Not
10:05sure if they changed that, but you know,
10:07unless you're getting 30 qualified leads
10:08piping back in, it's not worth the
10:10setup. But in an ideal world with larger
10:12budgets, we'll we'll connect Salesforce,
10:15HubSpot, you know, whatever the client's
10:16CRM is, and send a conversion action
10:19whenever there's a qualified lead or a
10:21converted lead. And the goal would
10:23really be to use that especially in in
10:26Pmax for lead genen to use only that as
10:28opposed to just tracking leads, but tell
10:31the system which ones are the good
10:32leads. So go after more instead of
10:34looking at your cost per lead, look at
10:35your cost per qualified lead. And that
10:37that could make a big difference. But
10:38you need enough data. you need, you
10:40know, probably at least 100 leads coming
10:42through on a monthly basis to get a
10:44decent number of qualified leads.
10:47>> And like what systems do your typical
10:49clients use to pipe that back into to
10:51you and then you into the Google ad
10:53system?
10:54>> So, if they're using Salesforce or
10:55HubSpot, there's a direct integration. I
10:57think Zoho also has a direct
10:59integration. If they're using something
11:00else, which a lot of our clients are,
11:02we'll typically use something like
11:04make.com or Zapier to just send the data
11:07back and forth. It's fairly easy to set
11:09up. You got to be a bit technical, but
11:11once you do it a couple of times, it's
11:12it's easy.
11:14>> Yeah. I'm curious. So, make Zapier Nadn.
11:18>> Yeah.
11:20>> What's sort of the difference that
11:21you've seen between them? Because I'm
11:23certainly looking at some of these newer
11:25tools like NAD.
11:26>> Yeah.
11:27>> Zapier's always done a really good job
11:28for what we needed it for, but like what
11:31would be a good use case to maybe go
11:32with something a little bit newer?
11:34>> Yeah. I mean, I find a lot of our
11:36clients are gravitating towards make
11:38just purely based on price. It seems
11:40their pricing is a lot more competitive
11:41than some of the others and it ending to
11:44play around with. So, I'm not as
11:45familiar with it, but it seems to be,
11:47you know, you know, it's a lot more
11:48advanced, a lot more AI forward. So, you
11:51know, we'll see. But at the end of the
11:52day, you just need something to kind of
11:54pull that data back and forth so that
11:55you're not relying on, you know, manual
11:58uploads.
11:59>> Yeah. And to what degree do you think
12:01agencies should step into business
12:04consulting? Because I know that's been a
12:05bit of a debate as well. Um, and kind of
12:08the framing is that as more of PPC tasks
12:12can be done automatically or through AI,
12:16what is the value that you as an agency
12:18bring? And so some people have argued it
12:20is just becoming a business consultant,
12:21but but then others have argued that no,
12:24the client hires the agency to do things
12:26in the ads account. they don't hire you
12:28to be a business consultant. So, where
12:30do you draw that line?
12:32>> So, I I think for me, like we've always
12:34been a bit more of like the boutique
12:37approach, trying to help wherever we
12:38can, even if it's outside the account,
12:40just to provide extra value. But I think
12:42that at the end of the day, if I'm
12:44spending somebody's money, I want to
12:45make sure that it's going to work for
12:47them. And so if there are things that
12:49need to get done, whether it's, you
12:51know, setting up a better landing page
12:52or or more deeper endto-end tracking,
12:55you know, we always try to jump in and
12:56get it done so that we can help them
12:58grow their business and and grow our
13:00business alongside them.
13:03And now with the whole consumer shift
13:05towards doing generative AI searches and
13:08using chat GPT
13:10uh using Gemini and Google, have you
13:13seen that change the volume first of all
13:16that you get on more traditional
13:18campaign types? And how do you and your
13:20clients think about making sure you play
13:24in these new surfaces that are taking
13:27more and more volume?
13:29>> It's it's definitely something we're
13:31thinking about a lot. I mean, the whole
13:32world is changing. You know, in a year
13:34from now, it's going to be even vastly
13:36more different than it is today. And so,
13:39we're just trying to stay ahead of
13:40everything. Like, you know, Chat GPT
13:42just announced that they're going to be
13:43putting ads in inside generative AI and
13:47and Google's been talking about ads in
13:49AI mode for a while. So, it's it's
13:51starting to get there, but it's still
13:52very early stages. I think one of the
13:54biggest things I would tell people is
13:56take advantage of it as soon as the
13:58option is available to you because it'll
14:00likely be, you know, you'll get that
14:02first mover's advantage. Traffic will be
14:04much cheaper. Um, but I haven't really
14:07seen any meaningful declines in any of
14:09our client campaigns in terms of search
14:11volume. I think Google, you know, is
14:13very smart about how they roll these
14:15things out. I don't think it's going to
14:16impact their, you know, the revenue,
14:18their bottom line. They'll figure out a
14:19way to make it work. And so, you know,
14:22we're just paying close attention and
14:24pushing forward.
14:25>> Yeah. Yeah. And if anyone's interested
14:27in getting ads on OpenAI, uh, please
14:30leave a comment or reach out to
14:31Optimizer. We have access to the
14:33inventory.
14:34>> Okay.
14:34>> Um, so we can get you in there.
14:37So, um, yeah. And then speaking of AI
14:40and like obviously there's this huge
14:41shift that you're alluding to where we
14:43don't know what it's going to look like
14:44in a year from now. uh but talk to me
14:46about what is your perspective on the
14:50future first of all of agencies.
14:53>> So it's definitely shifting but I think
14:55it's it's always been shifting. It's
14:58just uh this is going to be a bigger
14:59shift than what we've seen in the past.
15:01Um you know like I've I've been doing
15:04this for 20 plus years and like I've
15:05seen a lot of different iterations of of
15:07the Google Ads app platform, seen the
15:10roll out of of meta ads, Tik Tok ads and
15:12and a lot of the other ones. I think
15:14that the agency just you know needs to
15:16provide more value in the overall
15:19strategic marketing and and you know
15:21developing the strategy more endtoend
15:23like I mentioned instead of just sending
15:25traffic and reporting on what you see in
15:27Google ads to make sure that you're
15:29actually driving the business forward
15:31generating net new revenue for the
15:33client and you know really being more of
15:35a partner and less of like a a siloed
15:38agency um you know and then obviously
15:40leaning into all these different AI
15:42tools to figure out how to make the work
15:43more efficient and and provide more
15:45value for customers. It's not even about
15:47like yes some tasks are are being
15:50handled by the automation by AI but
15:53maybe now that gives us more time to do
15:55something even better and drive more
15:57value than we did before. So I don't
15:59know to me change is is exciting. It's
16:01daunting but it's exciting and we'll
16:02we'll find the best way forward.
16:05>> Yeah. Um, and as far as like career
16:07advice, uh, I don't know if you're
16:09hiring at the moment, but if you were to
16:11bring on new people onto the team, like
16:13what kind of skills would you look for
16:15that are maybe different than what you
16:17traditionally looked for?
16:19>> So, I I think the I've always liked the
16:22attitude in someone where it's like, let
16:24me just figure this out instead of
16:26asking a lot of questions or how do I do
16:28this or waiting for instruction. the
16:30type of person who just sees something
16:33and goes, "Huh, let me try to figure
16:35this out." That's that's probably going
16:37to be more valuable now than ever
16:39because there is so much more to figure
16:40out with AI. You know, the curious
16:42person who who asks a lot of questions
16:44and tinkers and just tries to get it
16:46done really just make it happen.
16:49>> Yeah, for sure. And the pace of
16:52improvement or change is so rapid right
16:54now. Um, I mean, I don't know if you
16:58have a perspective on
17:00picking a horse in the AI race. Have you
17:03picked a horse or are you kind of
17:05playing between Claude Open AI?
17:07>> So, I'm kind of playing I'm kind of
17:09bouncing around a lot. I find myself
17:10literally going from Claude to Open AI
17:13to to Grock to Gemini. I think at the
17:15end of the day, Google has a lot on
17:18their side between baking AI mode into
17:20search and then, you know, obviously
17:22Google Workspace, they have paying
17:24subscribers who don't need to pay extra
17:26for it. I think like there there's a lot
17:28that they can really push a lot faster,
17:31I guess, than than somebody like Open
17:34AAI who has to build the whole
17:35functioning business underneath it. So,
17:38I kind of think Google will land ahead,
17:40but time will tell.
17:43>> Yeah. No, I I agree with what you're
17:45saying there. And for us,
17:47it was a subscription to OpenAI for the
17:49team, but then you know when Gemini adds
17:52that to the existing workspace plans,
17:54>> it does cost a little bit extra. You
17:56have to upgrade, right? But it's it
17:57makes sense because you also then look
17:59at, oh, do we need a Zoom license,
18:01right,
18:02>> in addition or does that higher plan
18:03then have like the advanced Google Meet
18:05capabilities?
18:07>> Um, and so they certainly have an
18:09advantage there where it just
18:10financially makes sense.
18:11>> Yeah. start using that one and so long
18:14as their tool is good enough on par with
18:17all the others and right now you might
18:18say in some ways it's the best that's
18:20out there so then it's a no-brainer to
18:22go that direction.
18:24Um
18:25yeah so so that's certainly an advant
18:27and I think the other advantage for
18:28Google is the fact that they have so
18:30much data right it's
18:31>> they have all of my many many years of
18:33Gmail of maps data where have I been so
18:36if I ask a question to their assistant
18:40um yeah it knows me better it kind of
18:42knows what I'm likely going to respond
18:44to and that's helpful to me
18:47>> yeah Google does have a tremendous
18:49amount of data across many different
18:51areas you know it's video it's text it's
18:53email It's meetings, it's it's ads. So I
18:57I think that Gemini will will land
18:59further ahead than people think. Um even
19:02though chatbt right now seems to be be
19:04the big one that's in like the general
19:06consumer's mind. Like I think the kids
19:08they call it chat. So
19:10>> yeah.
19:12Well, and I don't know if you watched
19:13the Super Bowl, but what I found
19:15interesting between the ads for the AI
19:17companies was that Google's Gemini ad
19:21wasn't super explicit about it being AI.
19:25People often probably don't really
19:26equate Google's capabilities with AI.
19:29>> Uh, but it was a brilliant illustration
19:31of how it could be useful in everyday
19:33life.
19:34>> Um, and it was also not very chat
19:35focused. it was image generation focused
19:37and sort of um you know ways that the
19:41technology will improve your life in a
19:42way and I thought that was great because
19:44it did speak I think to the average
19:46person who doesn't really care about
19:48Gemini versus cloud versus GPT they just
19:51want to get stuff done and it's like oh
19:53here's a cool way to do something that I
19:55couldn't do before let me try it out and
19:57who cares what back end it runs on
20:00>> yeah I would agree with you there for
20:02sure
20:03>> yeah
20:05Um, all right. So, let's go back to some
20:09of the hands-on PPC stuff. Uh, but I I
20:12know that you've also talked about sort
20:15of the fundamentals
20:17and the timeless principles of, uh,
20:20advertising. I think that's super
20:22relevant because as we see so much
20:24evolution in campaign types and how you
20:27set these things up, what are sort of
20:29those fundamentals that you think people
20:32really need to pay more attention to?
20:35So with with everything, you know,
20:38keeping in mind those fundamentals is
20:39important. So Google ads the platform
20:41primarily is based on capturing search
20:44intent. Somebody's searching for
20:45something that you're selling, put your
20:47ad, let's go. I think all the recent
20:49iterations are are Google trying to move
20:52beyond the intent, you know, because if
20:55there's a limited amount of searches
20:57happening for your keyword, all these
20:58new algorithmic features can take you
21:00beyond it. But I think like the
21:02important piece I tell people starting
21:03out is just start with the basics. So if
21:06it's e-commerce, you're going to, you
21:07know, you might want to start with a
21:09shopping campaign, pull in search
21:11volume,
21:12collect data, what's working, what's not
21:14working. And for lead genen, it's just a
21:17basic search, you know, get traffic, get
21:19leads, get phone calls, close deals. You
21:21know, once you have some data, then you
21:23start to layer on some of the more
21:25advanced things like performance max or
21:27demand gen or AI max and try to try to
21:31go beyond it. But ultimately, you know,
21:33you can't sell something to somebody
21:34who's not your customer or has no
21:38interest in buying it. Like you do have
21:39to remember the stage of the funnel that
21:42the campaigns sit and and the purpose of
21:44each campaign type, you know. So start
21:46with the basics and then layer on
21:48additional as you're looking to expand
21:50beyond it.
21:52>> Yes. And so I hear you advocating for
21:56some of the basic campaign types. And
21:59it's interesting because you're saying
22:00Pmax is the more advanced campaign type,
22:03whereas I think Google would position it
22:05as it's the campaign type for everyone.
22:08You don't really need to have done much
22:09advertising in the past. You don't need
22:11to figure out things like keywords. just
22:13tell us your landing page, your budget
22:14and uh we'll figure this thing out,
22:16>> right? But but obviously the audience
22:17listening here today is a bit more
22:19advanced. Uh then that's for us, how
22:22would you frame this?
22:23>> So I think it's important to remember
22:25like how each of the campaigns work like
22:28shopping is basically looking at your
22:30product titles, descriptions, finding
22:31keywords, and matching it to to people's
22:33search queries. So it's going to be a
22:35little bit um I'd say more towards the
22:37bottom of the funnel. PMAX it's it's
22:40more simple from you know what it might
22:43need from someone to set it up but it's
22:45a lot more advanced in what it's capable
22:47of and how it works. It'll besides for
22:50running those shopping listings it'll do
22:52retargeting. It'll do standard search
22:54and shopping. It'll do, you know,
22:56display video. And so it can it can find
23:00customers in a lot of different ways.
23:02And if it's not set up the right way,
23:05it'll just kind of recycle your very
23:07bottomfunnel traffic. And if you really
23:09want it to like, you know, drive new
23:11business, you have to think about
23:12everything else you're doing and, you
23:15know, how you're setting it up. Because
23:17if you set it up one way, it might drive
23:19a ton of brand search. If you set it up
23:21another way, it might drive a ton of
23:23generic search. Um, you know, if you're
23:25running advertising on Meta with large
23:28budgets or any paid social, PMAX will
23:31largely recycle that traffic and you
23:34won't see like a meaningful lift in your
23:35overall metrics. But if you set it up in
23:38a in a different way, it can really push
23:40new traffic. And I think that those are
23:42the some some of the things I try to
23:44tell people to think about is like how
23:46everything is configured. Um, are you
23:48including brand, excluding brand, are
23:50you adding search themes? Are you
23:52putting creative or are you just relying
23:54on your product data feed? Um, you know,
23:56so there there's, you know, are you
23:57targeting new customers or or new
24:00customers and old customers. So there's
24:02a lot of different things it can do. And
24:04so it's simple, but it's complicated,
24:08>> right? And so yeah, you mentioned a lot
24:10of factors there that you need to
24:12consider. Um, if people want to learn
24:14more about that and get some of these
24:16specifics, where should they look? Have
24:18you blogged about that?
24:19>> Yeah, I have on our website, you can go
24:22jxtroup.com. We have fairly
24:24comprehensive article on on JX on PMAX
24:27and and how it works. And, you know,
24:29there's there's definitely a lot more
24:31than people think about when it comes to
24:33PMAX.
24:35>> Yeah. Um and so one frustration with
24:39PMAX especially very early on was that
24:42it was taking credit and inflating its
24:45rorowass from sources like remarketing
24:49brand. Now Google has added a lot of
24:51measurement capabilities and they've
24:53added uh targeting capabilities.
24:56Do you feel satisfied at this point or
24:58how do you look at uh PMAX and how it's
25:02measuring itself?
25:04So to me, I think the biggest piece of
25:06Pmax and how it's different than a
25:08regular campaign is that you can't make
25:10changes directly. Your changes that
25:13you're making are basically pulling the
25:14algorithm in a different direction. And
25:17so they've uncovered a lot more data.
25:19You could see like the channel
25:20performance breakdown. You can see
25:22breakdowns of of whether your spend is
25:24going to display video um whether it's
25:27coming from creative assets or from a
25:28data feed. So there's a lot you can see
25:30there. I'd love to see more um from
25:33Google in relation to is the traffic
25:36coming from which specific audience?
25:39Like is it new customers versus old
25:41customers? You can't see as much as
25:42you'd like to there. Um but also like
25:45how much of it is coming from
25:46retargeting versus new tra like
25:49completely new traffic because even if
25:52you connect your customer list and tell
25:54it only go after new customers, it's
25:56still going to be retargeting website
25:57visitors. And if you're spending a ton
25:59of money on meta, it's largely going to
26:02be just, you know, retargeting to meta
26:04traffic. So there there I I wish there
26:07was a bit more data around that kind of
26:09stuff. But, you know, I'm definitely
26:10pleased that, you know, with the amount
26:12of of data they've unlocked for us over
26:14the last year or so.
26:16>> Yeah. And hopefully more to come in
26:19terms of uh what they add to it. And
26:21it's always funny, too, right? I mean it
26:23had so few options in the beginning and
26:25that made it somewhat easy but then now
26:27there's so many options that like you
26:28said it's there's a lot of
26:30considerations
26:32>> decisions to make so it does become
26:33quite difficult in a way.
26:36>> Now um let's talk specifically about
26:38e-commerce for a minute. Advertisers
26:41here have the option of running a feed
26:44only PMAX campaign or they can have a
26:46traditional shopping campaign.
26:49When do you use which and do you ever
26:51use a combination of both?
26:53>> Sure. So, I think the first rule of any
26:57commerce setup, whether it's shopping or
26:59PMAX, is make sure that each product is
27:02only in one campaign. If you have it in
27:04more than one campaign, it kind of
27:05fights with each other and it just
27:06reduces efficiency. Um, but then how do
27:10you know like should you use standard
27:12shopping, should you use Pmax feed only,
27:14should you use Pmax with creative
27:16assets? I think for me it starts with
27:19what type of product you're selling. If
27:21you're selling something that's highly
27:23commoditized, like think about like
27:25digital cameras or appliances, you
27:27probably want to stick with shopping.
27:29There's a ton of search volume around
27:31those types of products. People are
27:32searching, they're ready to buy. Um,
27:35those are the kinds of products that
27:36don't really succeed, let's say, with
27:37meta ads. So, they probably won't do
27:39well with creative assets in PMAX. Um,
27:42so either just shopping or like a data
27:43feed only PMAX. But if you're selling a
27:46product that's very visual, maybe like
27:48fashion and apparel, um, or even, you
27:51know, cheaper items where you do have
27:55success on meta, then I tend to push
27:57more towards PMAX and likely with
27:59creative assets, especially if you're
28:02seeing success on paid social. So that's
28:05kind of like the the main consideration.
28:07The other thing I would think about is
28:08also the length of of your purchase
28:11journey. So, for items that take like a
28:13day or two to buy, you don't really need
28:16the the retargeting component as much.
28:18But if you're selling a product that
28:19might have like a 30-day consideration
28:21cycle from when somebody first looks at
28:24it until they buy it, you might want
28:26that retargeting piece of PMAX to just
28:28stay on top of the consumer's mind. So,
28:31I think that those are like the main
28:33things I I'll think about is like what
28:34kind of product, what stage of the
28:36funnel the consumer is, and and the how
28:39long the purchase journey takes.
28:41Yeah, makes sense. Um,
28:45we're now 2026.
28:47We've been talking about PMAX for a long
28:49time.
28:51What is it that you would say people
28:52need to stop doing in 2026 related to
28:55PMAX? And what is it they should really
28:56double down on?
28:59>> So, Pmax specifically, I think the
29:01biggest thing is don't just set it up,
29:04try it, and then be like, "Ah, it didn't
29:06do anything. Forget it." Like like I
29:08mentioned before, there's a lot of
29:10different ways you can set it up and try
29:14different things. Lead it in different
29:15directions, you know, see what works,
29:17what doesn't work, look where the
29:19traffic is coming from, where the
29:21traffic is going to, what landing pages,
29:24um, you know, and and just at the end of
29:26the day, it's like everything else. You
29:28got to test it and test different
29:30strategies within it and see what works.
29:32>> Yep. Always be testing. Um,
29:38>> here's uh one of from my this is called
29:42the spicy questions. I don't know how
29:43spicy these are, but do you think
29:46agencies are overrelying on PMAX
29:49because it maybe in a sense that it can
29:52be easier to manage? I know you say it's
29:53harder to manage and I know you're one
29:56of the people who does it right,
29:59>> but what bad agency behaviors have you
30:01noticed that kind of piss you off
30:03because maybe they give agencies a bad
30:04name?
30:07So, I I don't like to talk bad about,
30:10you know, other people, but I'll tell
30:11you that like when you see a PMAX that's
30:13just like all products lumped into one
30:16set up with very basic configuration,
30:19missing like a lot of the granularity
30:20that's available to you in the sense of
30:22maybe missing search themes or
30:24first-party audiences or it's like why?
30:27Um, another one is when it's set up to
30:30inflate the revenue by giving like new
30:32customers a higher value. So, if you set
30:34it to target uh to bid higher for new
30:37customers, you can say like a new
30:38customer is worth $100 extra dollars.
30:40That like inflates the reporting by
30:42$100. It's so there there's things like
30:45that where it just looks better than it
30:47is. Um or if you have a PMAX where it's
30:50like, you know, the whole budget is
30:51going towards the consumer's the
30:53client's brand and then like it's not
30:56really driving any new business for
30:57them, you know. So, it's those kind of
31:00things where you're just using it to
31:01sort of pad the numbers. um you know
31:03that that's frustrating but a lot of
31:05times we'll revamp it you know 30 days
31:0860 days in and all of a sudden you know
31:10you're seeing a big lift in in Shopify
31:13revenue as you know as opposed to just
31:14in platform and that that's really nice
31:16to see.
31:18>> Yeah that's amazing. So, okay, you said
31:21kind of pisses you off when you see all
31:23the products being put into a single
31:24PMAX campaign. And I know this is a big
31:27question, right? Because you also
31:28mentioned you need 30 conversions over a
31:3130-day period for the system to really
31:33be able to do a good job. Now, I've
31:35always said this is 30 conversions of a
31:38certain conversion type. And I think you
31:40can split it between campaigns and the
31:43system can learn. Not everyone agrees
31:44with that. Um, so if you disagree to
31:47tell me your opinion, but
31:52>> I believe,
31:54>> sorry, we're cutting out, but I believe
31:55you're right on that. It's 30 of a
31:57specific conversion type over over a
32:0030-day period. Um, but I think like like
32:04everything else, it depends. You know,
32:05if you only have 20 conversions and a
32:08really low budget, definitely keep all
32:09your products in one campaign. Um, but
32:11as you grow and scale, it makes sense to
32:14segment out, you know, better performing
32:16products, seasonal products, different
32:17categories with different goals.
32:19>> And and so then break it down for us
32:21between additional campaigns, additional
32:26uh asset groups, additional listing
32:28groups like what does the actual
32:30structure look like? And then just like
32:32let's have a an imaginary client here,
32:35but what if you have that decent enough
32:38conversion volume? Say you have a couple
32:40hundred products, you have a couple
32:41hundred conversions per month, like how
32:43would you structure that?
32:45>> So I think the way to think about it is
32:47is try to mirror like the website
32:49navigation in as much, you know, as much
32:54as you can because if the client has
32:56three top level categories on the
32:58website, you know, that's a good
33:00starting point. Three campaigns with all
33:02those products underneath it. Typically
33:04the reason to do that is so now you have
33:07budget going to each of the categories
33:08that are important to the client.
33:10Whereas if you put them all in one, the
33:12spend might be skewed just to one
33:14specific category or you know like it
33:16ends up PMAX ends up pushing a lot of
33:18the spend towards the best performing
33:20areas. And so by by segmenting that you
33:23can give them different budgets,
33:24different target rows, different sales
33:26goals and it's good to have diversity
33:30because if one you know if something
33:32goes wrong in one or a product runs out
33:34of inventory at least you have something
33:35else to rely on. Um, the other thing
33:38also is a lot of times even if you do
33:40have something like that set up where
33:41you have three campaigns, one for each
33:43top level category, you might find that
33:45as a whole subset of products that are
33:47not getting touched. And so then we'll
33:49block those out from those campaigns and
33:51put them in a new one. You know,
33:53obviously making sure you have enough
33:55budget to run each one successfully.
33:59>> Okay. And then one level deeper from the
34:01campaign, where do you use asset groups?
34:03How what do you use them for?
34:05So asset groups are very good for
34:07basically product themes. So let's say
34:09you have a separate campaign for each
34:11highle category, but there's, you know,
34:13different subcategories. There's
34:14different visuals that you might want to
34:16put with each one. There's different
34:18search themes, different audience
34:19signals. Um, I think one thing I don't
34:22like when people do is put the same
34:24products in multiple asset groups with
34:26different audience signals because that
34:27doesn't really do anything. An audience
34:29signal is not like real targeting. It's
34:31just telling the system, "These are
34:32people who might buy from me. you know,
34:35use that as a suggestion. And so you
34:37want to make sure that it's really
34:39segmented around product themes or or
34:41category themes as opposed to, you know,
34:44audiences or first party data like that.
34:47>> So that's interesting. So say that you
34:49have a running gear company
34:51um and their audience splits between
34:54trail runners, marathon runners, and
34:57casual weekend joggers. Um they might
35:00all buy the same shoe, the same shirts,
35:02shorts, etc.
35:04Your advice then is not to have
35:07different asset groups even though
35:08there's I mean I could clearly imagine
35:11they would resonate with different
35:12visuals in the ads and different core
35:14messaging in the ads.
35:17So if you have different landing pages
35:19and completely different messaging, yes.
35:22But I've seen so many accounts where
35:24it's basically like the same asset group
35:26duplicated with different audience
35:29signals like in effect trying to like
35:32imagine it was demand genen campaign
35:34where you have different audiences with
35:36the same ads that that would actually
35:39work because it's targeting your
35:42audience. But in in PMAX, the asset
35:44group audience signals aren't really
35:46targeting. So, it's only worth making
35:48them separate asset groups when the
35:50messaging is different and when the
35:51landing page is different.
35:53>> Interesting. So, that kind of hearkens
35:54back to the days when people would put
35:56all three match types of a keyword in
35:58the same ad group
36:00>> and they would claim that it was because
36:01they got more traffic that way. uh when
36:04in fact the one broad match would have
36:06covered all match types equally
36:08>> and you could have used the segment to
36:10get your reports to understand what was
36:11working and what wasn't.
36:13>> Um so so you're saying a lot of you
36:15advertisers now kind of make that same
36:17mistake with audiences
36:19>> in asset groups when it's not really
36:21necessary unless you have different
36:23landing pages,
36:24>> right? Landing pages and creative.
36:27>> Yeah. Um okay. Okay. And then do you go
36:30one level below that in terms of like
36:32grouping your products in those asset
36:33groups?
36:35>> So I typically do just because I want to
36:37see how everything's performing. But
36:39from a um targeting and performance
36:41perspective, it's not really needed for
36:43PMAX, it's more of for us to, you know,
36:46watch how everything's performing. And a
36:48lot of times we'll go through the the
36:50more granular um listing groups and
36:52exclude maybe a specific product or a
36:54subcategory that's just not performing
36:56well um and either drop it into, you
36:59know, maybe a standard shopping campaign
37:00with much lower bids or just exclude it
37:03completely.
37:05>> Great. Okay, that's awesome specific
37:07advice. Uh people hopefully love this.
37:08So, thank you Manakim for that. Um now,
37:11let's talk about what is your biggest
37:13frustration? Um I'm a big vibe coder.
37:17Uh, I love using technology to solve
37:19problems. And um, like what is one thing
37:21that frustrates you in the industry
37:22right now? I don't know if I can help
37:24with it, but like what should somebody
37:26go out there and fix?
37:29>> What should somebody go out there and
37:30fix? That's that's a good one. Um, I
37:33think AI is got to slow down. It's a
37:35little little too fast in in the ad
37:36platforms. Like real businesses rely on
37:38this to provide, you know, food for
37:40their families and it's just AI AI AI.
37:43Like we still got to prove it out. It
37:45it's definitely very valuable and and I
37:47you know see the value of of AI and
37:50smart bidding targeting and and
37:51performance but like
37:53>> I almost feel like we need to slow down
37:54and take a deep breath until we figure
37:56it out. And
37:58>> well where does it freak you out?
37:59Because like smart bidding has been
38:01around for a long time and generally
38:02gets really high marks.
38:04>> Um Right. It's it's it's learned a lot
38:06along the way and it can it can fail if
38:08you don't have enough data like you
38:09talked about. But by and large smart
38:11bidding I think can be trusted quite
38:13well. Yeah.
38:14>> So, so when you say like slow down, is
38:16this about creative generation? Is it
38:19about broad match always becoming even
38:21more and more aggressive or where is
38:23your main frustration?
38:25>> So, smart bidding, I agree with you.
38:27It's it's a definitely a very uh very
38:30wellperforming product. We use it pretty
38:33much everywhere except for very few
38:35scenarios like brand search or or low
38:38volume lead genen. Um, so I agree with
38:41you on that one. I think like where it
38:43pushes too far is like on on some of the
38:46the lead genen search campaigns the
38:49targeting is like way off base
38:50especially like we have a lot of clients
38:52in healthcare and is we're looking for a
38:54very specific kind of traffic and I'd
38:56love to use broad match more but it's if
38:59it brings a um a search in that's
39:02completely unrelated to what the client
39:04is trying to sell I just I can't do
39:05anything with it and so it pushes a
39:08little bit too far in in certain things
39:09like that. Um, also in in like I think
39:13the the AI for for creative it works it
39:17doesn't work like the ads advisor it
39:19works sometimes it doesn't work
39:20sometimes but then a lot of times it it
39:22gets pushed into the workflow and it's
39:24like I just want to build a campaign but
39:26at the same time I should say I am very
39:28um very forward thinking when it comes
39:31to these things. I'm I'm not like stuck
39:32in the past, but I just want to make
39:35sure it doesn't slow down, you know, the
39:36the processes and progress that we have
39:38in place already.
39:40>> So, a measured level of AI is really
39:43what you're advocating for, and that
39:44makes a lot of sense. Um, ultimately, I
39:46think we all want the guard rails, and
39:47we want to understand why the AI is
39:50making decisions and what it's doing for
39:51us and make sure it's doing the best for
39:53our clients.
39:55>> 100% with you on that.
39:58>> Well, good. Um, final question.
40:01How was the uh Tesla tequila that I see
40:04the bottle behind you? [snorts]
40:06>> I bought the empty bottle just for
40:08display purposes.
40:10>> I have an empty bottle, too. I bought
40:11the empty bottle from eBay
40:13>> because I couldn't get a full bottle. I
40:15was too late to get that one.
40:17>> I've been told I have too much Google
40:18stuff, so I was just trying to add
40:20variety.
40:21>> Yeah. Well, very good. Hey, Benam, uh
40:24thank you for joining us today. if
40:26people do want to get a hold of you,
40:28like what's uh where do you hang out
40:30online and uh what can you help them
40:32with?
40:33>> Sure. Sure. First of all, it's been a
40:34pleasure talking with you as always. If
40:36anybody wants to reach out, I'm I'm on
40:38Twitter, I'm on LinkedIn, or you can
40:40visit our website, jxtgroup.com. Happy
40:42to provide a free audit if anybody needs
40:44help with their Google ads and we're
40:46here to help.
40:47>> Okay, that's a generous offer. All
40:49right, everyone. Thank you for watching
40:51today's episode. If you've enjoyed it,
40:54then please do give it a like. Put in
40:56some comments if you have questions for
40:57myself or Manahim. And with that, we'll
41:01wrap it up here and we'll see you for
41:02the next PPC town hall.

About this transcript

This page contains the full transcript of How Experienced Marketers Work Differently With PMax and AI Max | PPC TH 124 by PPC Town Hall, generated from the public captions YouTube serves with the video. The transcript has 8,009 words across 1,164 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.