
Episode Description
In this episode, Julie Bacchini Friedman, Andrew Lolk, and Frederick Vallaeys share their honest advice about which old practices to drop, what new strategies to try, and which current approaches need more attention in 2025. They cover why some popular tactics like single keyword ad groups aren’t useful anymore, how to make smart bidding work better, and why video is becoming more important.
The conversation also explores how to balance automated tools with hands-on management, and ways to stand out when advertising on multiple platforms. This discussion is packed with real examples and straightforward advice for anyone running digital ads in 2025. Whether you manage big marketing campaigns or handle ads for a small business, you’ll learn which strategies are worth your time and budget.
Episode Takeaways
✅Biggest PPC mistakes to avoid in 2025:
In 2025, relying on outdated tactics like last-click attribution or static Smart Bidding targets can harm your PPC performance. Evolving with new strategies and optimizing campaigns regularly is crucial to stay competitive.
✅Why SKAGs and last-click attribution are holding you back:
SKAGs overcomplicate things, and last-click attribution doesn’t tell the full story. Use more flexible campaign structures and better attribution models to get clearer insights and improve performance.
✅Smart bidding strategies that actually work:
Smart Bidding can bring great results, but you can’t just set it and forget it. Play around with bid strategies, adjust your targets, and make the most of portfolio bidding to drive better outcomes.
✅How to use Performance Max (PMAX) the right way:
Performance Max campaigns require careful management to deliver their best results. Regularly update your targeting settings, test creative, and monitor how Performance Max interacts with other campaigns to get the most out of the tool.
✅What’s next for Google Ads, Meta Ads, and Search Marketing: AI is changing the game. Expect smarter targeting, better personalization, and more cross-platform integration. Get ready to use these advancements to improve your PPC results.
✅Focus on competitive research & strong ad copy:
Stay on top of your competition, and don’t underestimate the power of a great ad. By crafting specific, targeted ads, you can attract the right audience and get more clicks from people who actually want what you’re offering.
✅How AI is Shaping the Future of PPC:
AI’s role in PPC is expanding rapidly, making it easier to automate and optimize campaigns. From ad copy generation to smarter bidding strategies, AI can help you better understand your audience, predict outcomes, and execute complex strategies with less manual effort. However, relying too much on AI without understanding the underlying mechanisms can lead to poor decisions. Stay informed, keep experimenting, and embrace AI as a powerful tool to complement human insights and expertise.
Episode Transcript
Frederick Vallaeys: Hello, and welcome to another episode of PPC Town Hall. My name is Fred Vallaeys. I’m your host. I’m also CEO and co-founder at Optmyzr, a PPC management software. For today’s episode, we’re going to look at 2025 and it being a new year, what are some of the things that PPC practitioners are going to stop doing and start doing?
And doubling down on. So let’s talk about strategies that we think have run their course, the new strategies that we’re excited about that are going to drive results in the year going forward. And to talk about all of this, we’ve got two great PPC experts who are going to share with us what they think in terms of starting, stopping, and doubling down.
So with that. Let’s get rolling with this episode of PPC Town Hall.
I have two great guests today, Andrew Lolk and Julie Bacchini Friedman. Welcome back to the show. You’ve been on the show multiple times. It’s great to see you again. Julie, let’s go to you first. Tell us how you’ve been and what’s new with you.
Julie Bacchini Friedman: Oh, thank you for having me. I always love coming on here and sharing my thoughts about what’s going on in the world of PPC.
Things are good. I think 2025 is going to be, I’m going to say, an interesting year. That’s going to be my catch-all for all the things that may or may not happen this year. Lots to look forward to and pay attention to.
Frederick Vallaeys: Great. And Andrew, welcome back as well. What’s new with you?
Andrew Lolk: I think the biggest thing is this: I feel like I have a New Year’s resolution to fight less on LinkedIn. And I decided to pick a fight with everybody in the first two weeks of being back from vacation. So that’s just interesting. So that’s good. Nothing changes. So, I believe everything will be the same in 2025.
Frederick Vallaeys: We’ll all fight about the same thing. So it’ll be interesting. We will. But when it comes to PPC, I think maybe there are a few things that we can all agree on. We might have to stop doing so in the format that we want to go out. We’re going to go with today for the listeners. We’re going to talk about what people will start, stop doing, start doing, and what we’re going to double down on.
So, let’s go with that first category. Is there anything you think we should no longer be doing in PPC?
Andrew Lolk: I’ll let Julie go first.
Julie Bacchini Friedman: All right, let me go first. I feel like there are some obvious things. Certainly, if you’re still doing and I can’t believe that I’m going to say this in the year 2025, but somebody just messaged me about this last week.
So, I’m going to go ahead and say it. The single keyword ad group’s time has come and gone, people. No more. No, listen, let me preface everything I’m going to say today with you. You may have something that’s working fabulously for you, and I am not going to suggest that you blow something up that’s working for you right now.
I think that’s an important sort of disclaimer that we talked about here. Sometimes, you have a legacy thing that is working in a particular account. You don’t have to blow it up; you need to be ready if it stops performing and the way that it has for you with what are you going to do next?
I’d say that when planning for both our current accounts and any new ones we take on in 2025, SKAGs should be the first thing to go. They just don’t belong anymore.
Frederick Vallaeys: Okay, so that’s one, and so that makes sense then for new accounts You probably wouldn’t introduce the SKAG before existing accounts where it’s working.
You might phase it out over time now that’s talking about account structure, right single keyword ad groups Is there something that might replace that, and are we talking here about stags and single-themed ad groups? What’s the replacement, in your opinion?
Julie Bacchini Friedman: So I feel like the platforms are trying to get us to be broader and broader in everything that we’re doing.
I prefer to try to, even if it’s just for my thought processes and managing things, sometimes, for budget allocations, I’m a fan of the stack, the single theme, trying to keep things. They are a little bit separated. So you have more flexibility in the language that you’re using in your ad copy.
You could be sending to different landing pages depending on how sophisticated you are on that side of things. So, for me, the way that I like to manage things is to keep things themed so that you don’t have everything just crushed together in one, in one giant, one giant campaign.
Frederick Vallaeys: That makes sense. Andrew, any thoughts on the structure or what’s one of your top things you will stop doing?
Andrew Lolk: I’ll do the echo. I agree on the single keyword ad groups. Also, just in general, I think one thing to just pick back off there is one of my, one of the big things for us is negative keywords.
The more you work with Smart Bidding. And again, this has to do with certain big parts of the market. Again, if you’re selling sex toys, if you’re selling posters, if you’re selling certain things, then you need to still utilize a heavy negative keyword list. But we just see in 80, 90 percent of cases actually deleting all negative keywords.
When we take over accounts from other agencies or in-house, it just blows everything out of the water. We have a case where if you open up the slides I sent you, we have one case where we actually were able to take and remove all negative keywords. And what we saw was an increase in CPC.
That’s fine. But we also saw a 6x or 5.5x increase in revenue simply by doing that. This was an account where the previous agents had overly focused on negative keywords and tried to enforce strict control. Every keyword they added as a normal keyword in the account was also added as a negative in the dynamic search ads campaign.
They were essentially trying to sculpt the account too aggressively. They also applied this control across campaigns and search, insisting that each specific search term had to appear in a particular ad group because they were using single keyword ad groups. That level of control is just pointless these days.
You can’t control it. Phrase match and close variant matching make it nearly impossible. And the more you try, the more likely you are to make mistakes and exclude things too broadly. So I’d say no more single keyword ad groups. Yes, if you’re selling mattresses, you might need multiple keyword match types in one ad group to track performance properly. Great. But as an overall account strategy, SKAGs just don’t work anymore.
And excessive use of negative keywords to control things that smart bidding should already be handling is something to rethink.
Frederick Vallaeys: And that’s interesting because Google’s matching mechanism has become more sophisticated as machine learning and artificial intelligence have improved. So resetting negative keywords makes a lot of sense.
But we’re certainly not saying don’t use negative keywords. It’s more about doing a reset. We need to see what Google is actually matching us to and identify where it still doesn’t make sense. Those should be excluded as negatives.
However, in many cases, something may have needed to be a negative in the past because we were spending too much on it due to less sophisticated bid automation. That’s where losses were happening. But now that bidding has improved, some of those negatives might actually be okay to let trigger ads.
Andrew Lolk: Yeah, the way to look at it is that negatives are a black-and-white, on-or-off switch, while smart bidding should have the flexibility to adjust bids for a search term or product based on device, audience targeting, and seasonality.
For example, if we’re selling swimwear, a keyword like “swim trunks with legs inside” might not perform well in September but could be great in May and June. Yet, we often see advertisers overusing negative keywords.
If you have a lot of data in your account, whether you’re working in eCommerce or a broad lead generation space, negative keywords are often overused. Now, if you’re a lawyer in a very specific niche, then yes, heavy use of negatives makes sense. But in 80% of accounts we see today, negatives are just way overutilized.
Frederick Vallaeys: Yeah, so the tip here is to use negative keywords for black-and-white decisions where it’s absolutely clear that something has no chance of being relevant. But don’t use them when managing bids for terms that could be relevant seasonally.
As I’ve always said, there’s no such thing as a bad keyword, only a bad bid. Sometimes, though, you can’t bid low enough to make a keyword worthwhile, and in that case, negatives are necessary.
That’s the idea here. Julie, what’s your take on negative keywords?
Julie Bacchini Friedman: I think one of the things we’re going to have to do in 2025 is find that sweet spot, that balance for each account we work on, between not letting machine learning and smart bidding go off on tangents we know will waste money.
I think it’s about finding that balance between stopping that from happening without completely cutting off any opportunity the system might be able to find. And I think it’s a bit of a balancing act, especially for budget-conscious advertisers.
If you have a lot of budget and plenty of runway to let the algorithms do their thing, you have more flexibility in how you approach negative keywords. If you’re more budget-limited or if your conversions take longer, you might need to be more precise in your strategy when deciding what to eliminate and what to experiment with.
So, I think it depends on the parameters of your account, including spend and volume.
Frederick Vallaeys: Yeah, and one thing I heard you mention ties into attribution models. You didn’t say it directly, but it’s essentially about where the value is coming from.
Are there any attribution models that we think are outdated and should no longer be used?
Andrew, I’ll start with you on that one.
Andrew Lolk: Ooh, I haven’t talked about attribution models in a long time. I just don’t believe in any of them anymore. Last-click attribution is absolutely dead. It should never be used for anyone outside the baby stage of an account. Even then, Google will give 50% of the attribution to the last or first click in their own DDA, which shows the weighting is completely wrong no matter how you look at it.
I don’t trust any of it. Last-click is 100% wrong, and first-click doesn’t work either. The closest thing we’ve got now is the DDA model. It is what it is—we can’t do much about it anymore. So within a Google Ads account, it’s not even a conversation we have anymore. But we definitely don’t use last-click.
Frederick Vallaeys: Julie, where do you stand on attribution models?
Julie Bacchini Friedman: If we’re being honest, attribution has always been a bit of fairy dust and wishes rather than hard data, right? And it’s even more so now, I think, with less data being available due to privacy and consent mode, and all these other changes happening. There’s less data for the platforms to work with, right?
So, getting more comfortable with the fact that things are modeled is something we’ll all need to really wrap our heads around. I think heading into 2025, if you haven’t already, this will be essential. Even though attribution could be argued as never being that precise or great, we as an industry marketed it that way.
We sold it as a strong, reliable factor: this is why you should do digital advertising instead of other types. But I think that’s starting to unravel as we move from 2024 into 2025. So, I think we’ll need to change the way we talk about attribution; more as a weighting or contributing factor rather than some absolute piece of data.
This is something we’re going to have to adjust in how we communicate with clients and stakeholders, just because of how it works now.
Frederick Vallaeys: And that’s a really interesting point because I think there are two dimensions to this. One is what we report to stakeholders and clients, and you’re right, it’s fairy dust.
So, in that attribution model, is that really the amount of revenue we produced? No, that’s just how we attributed it. There will be double counting and under-accounting from the reporting perspective, so we need to be really careful there.
But the other dimension is that these attribution models are using the same machine learning systems that also handle the bidding. So, if you’re going to trust your bidding to these systems, it probably makes sense to have a similarly run system to assign value and help decide where the bids should go.
Pick the fairy dust that moves you in the right direction, and then tie those two together.
Yes, good. Let’s move on to the next one. Is there anything else we’re going to stop doing in 2025?
Andrew Lolk: So one thing that I’ve seen a lot throughout 2024 and 2023, and I still see people doing, is splitting campaigns. I’ll take a twofold approach here. First, feed-only Performance Max campaigns. I’ve been the biggest spokesperson for the idea that if you want to run PMax, you should do feed-only, because delving into display and video at the same time as social shopping is a horrible idea.
I still stand by that, but today I don’t see a reason why you would run a feed-only Performance Max campaign over a standard shopping campaign. So, I just don’t see the need for it anymore.
Alongside that, if you are running PMax and you are running PMax feed-only (meaning you’re only focused on the shopping app component of PMax), for goodness’ sake, stop splitting up your campaigns. There is no data sharing between the two. The conversions are not shared. You cannot do shared budgets. You cannot combine or consolidate data in a bidding strategy.
Every time you split up a PMax campaign, that individual PMax campaign needs to hit a certain volume of conversions to be effective. And again, that’s 30 conversions a month. Yes, that enables target ROAS or something like that, but for it to be effective, it needs 300 conversions a month. To be excellent, it needs 3,000.
I’m exaggerating a bit, but the point is splitting up your products and categories into several PMax campaigns is a horrible idea. It never works, and consolidation is such an important factor for PMax these days. You’re just shooting yourself in the foot with this approach.
So, if anybody’s doing that, I know Julie said you shouldn’t blow your accounts up, but that’s the one thing I’d say blow your accounts up if you’re doing that today. If you’re not having different ROAS targets per category, then there’s no need to have a different campaign. That’s basically it.
Frederick Vallaeys: Okay, so you just said if you do have different ROAS targets, then maybe you should have different PMax campaigns, but you’re still making the point that you’re splitting your data and just making it harder for the system to reach its goals.
Andrew Lolk: If you’re selling shoes and you’re selling jeans, and your jeans can run at 500 percent ROAS and your shoes can run at 300 percent ROAS, be my guest- split it up. That’s the right way to do it because you’re splitting it up and using the target ROAS lever to target different levels and maximize revenue within each category. Best practice, beautiful scenario.
80 percent of the time, I open up something and it’s the same target ROAS. One campaign has 700, another has 725, and the other has 715. It doesn’t make a difference whatsoever.
Frederick Vallaeys: Now, you’re still a fan of shopping campaigns versus PMax for feeds only. What are some of the benefits you see there, and why do you still like using those shopping campaigns?
Andrew Lolk: So, it’s actually one of our double downs this year to double down on standard shopping. There are a couple of things. One is the knowledge that we’re only buying shopping ad clicks. Once we decrease the ROAS target or increase the budget, we’re not sending a signal to Google to say, “Try to go into display, try to go into video, try to go into search.” We’re only saying, “Maximize shopping.”
What we’re doing, smartly, is that we usually take over accounts where the goal isn’t to get us some revenue through Google Ads.
Our goal is to maximize Google Ads as a channel. By doing that, we need to find out where the diminishing returns are on shopping, where they are on search, and where we have YouTube or display. Whenever touch is a bad idea, but with shopping ads, we know exactly what we’re getting. So if we lower from 300 percent ROAS to 250, we know exactly what we’re getting more of.
When we know what that incremental extra value is we’re getting out of shopping ads, it becomes a component that’s much harder to control with PMax. The fact that we’re only getting shopping ad clicks is really the biggest advantage.
You can create different campaigns, structures where you move products from one campaign to another, use bid portfolios, set different targets, and have access to minimum and maximum bids. It just means we have different components we can use or levers to pull throughout the year. We don’t just sit and wait for certain things to happen.
It’s really one of our core beliefs: things that you can proactively do to help smart bidding should be done. That’s the only way to really scale an account beyond just grabbing what smart bidding will give you out of the box.
Frederick Vallaeys: This makes sense Julie, what’s your take on PMAX and campaign structures?
Julie Bacchini Friedman: I don’t do e-commerce, so shopping is not something that’s in my wheelhouse whatsoever.
Performance Max on the lead gen side can be a little tricky, right? Because for a lot of lead gen accounts, that threshold Andrew was talking about with conversions really comes into play. Lead gen accounts often don’t reach a point where the machine learning and smart bidding hit the level of efficiency they can on a high-volume e-commerce account.
We find ourselves trying to figure out how to take advantage of the things that will work when you’re below those thresholds where things really start humming. How do you figure out which of those things might work for you? What you want to test, and which things will just feel like they’re stuck, running laps without taking off? If you don’t have that level of data flowing in, some things just don’t get off the ground.
There are things you can do to help this if you’re in an account with lower conversions. For example, you can use portfolio bidding strategies. That helps pool more of your data, which I think is a theme we have going today. How can you pool more of your data to provide it to the platform so it can best utilize it?
Instead of having all these little pieces where it’s looking at everything, you’re not setting it up in a way that makes it easy for the system to make connections and improve performance. If you’re not at that high threshold where things will just kick into high gear, you need to figure out how to utilize those pieces. How can you put it all together and try to simulate, as closely as possible, that situation with the totality of the volume you’re working with?
Frederick Vallaeys: Makes sense. Do either of you still have things you want to talk about as far as stopping? Or should we move on to what will you start doing?
Andrew Lolk: Let’s shift gears and talk about what people should be doing instead of what not to do.
Frederick Vallaeys: Alright, let’s focus on some of the things people should be doing. Andrew, let’s go back to you on this one.
Andrew Lolk: One thing I think is key, especially from an e-commerce perspective, is that almost every single e-commerce advertiser out there has an 80/20 split between shopping and search. In some industries, that makes sense, but in a lot of industries, it means you’re not utilizing search enough.
That could mean you’re relying too heavily on dynamic search ads, especially if you have a multi-brand store with thousands of products. I still see people focusing primarily on dynamic search ads for their main search campaign. The more we’ve worked on search, the more we’ve seen that spending some effort on search really works.
Going back to the swimwear example, if you’re running a dynamic search ads campaign and comparing year-over-year performance for your swimwear keyword, you might be able to see some trends by pulling reports or going through search terms. However, if you want to push your swimwear keyword before peak season, during peak season, or even when the weather is really nice, you need more control.
So, don’t just rely on dynamic search ads. Start pushing your search campaigns more. It’s important to take control of the keywords you want to focus on, especially in seasonal cases like swimwear. Being able to adjust your ads and bid strategies allows you to respond to trends like weather changes or peak seasons.
Even in industries with a large product catalog, you need to put more effort into search. It’s not enough to let shopping ads dominate. Search is a powerful tool that lets you capture all the potential customers that might not be reached through shopping alone. The data we’ve seen supports the idea that search campaigns can make a significant difference when properly managed.
Frederick Vallaeys: That makes sense. And so search, obviously, is not a new thing, but if you’ve taken the lazy approach, which I think a lot of people listening have by nature, and especially in PPC, we’re all kind of control freaks and we like the campaigns where we can manage the negative keywords, where we can see the data, where we can set the bids.
DSA, I think what you’re saying is basically that should be your catch-all campaign that you maintain for the stuff you didn’t think about, the stuff that maybe is suddenly trending, but you still need to monitor it and then bring it back into your search campaigns where you can actually take ownership, send it to the right landing page, craft the messaging, and do all these other things that you’re saying. Hands-on management does drive better results, and none of us here strive to be average. We strive to be the best in class, right? So that’s how you get there.
Andrew Lolk: The one note I’d say on dynamic search ads is that my threshold for when to move a keyword out of dynamic search ads and into an actual search campaign has gone much higher. Where it used to be, if you got like five clicks in a month, I was like, “Oh yeah, let’s move it.” Now it’s much closer to 50 or 100.
What dynamic search ads do incredibly well is they take a very unique long-tail keyword and match it to a product or service page. That’s what you want out of dynamic search ads. You just don’t want swimwear to be a keyword that’s lying inside dynamic search ads and shown to your generic homepage or category page title, or God forbid, a product long-tail keyword like “swimwear for swimmers black XXL.” It’s just a bad match. It’s a bad way to run a search campaign, and that’s where we see people get too reliant on DSA these days.
Frederick Vallaeys: Julie, not a shopping example, but what is something you’re going to start doing in 2025?
Julie Bacchini Friedman: One of the things I’ve been talking a lot about is making sure that you’re well-positioned competitively. I think Fred, your book talks all about this, but with more leveling of what happens before the click, like within the systems, with automation, and with machines making decisions on our behalf, I think it’s even more important now to be aware of what other options someone might have when they’re searching for what you’re trying to sell.
It’s incredible to me. After 20 years in this business, I still work with clients who honestly don’t have a true sense of what those other options are. How are they positioning themselves? What are they putting forward? It’s easy for brands to get tunnel-visioned and think, “This is what we do, this is why we’re awesome, and this is why someone should buy from us.” That’s important, but it’s just as important to make sure that what you’re putting out there, the copy in your ads, what’s showing up on your landing pages, stands up against what a searcher is going to see.
Your ad is hardly ever the only one that shows up in search results. There’s almost always at least one to two more ads next to yours. Smart advertisers are going to make sure that their ads are competitive and compelling, and create a reason for someone to choose them. It’s not that we haven’t always needed to do this, but I feel like it’s even more important now with less control over what’s happening before the click. That’s a piece I think is really important for advertisers in 2025.
Frederick Vallaeys: Yeah, so we have a partnership with Adthena, which provides some of that competitive data set. How do you go about that with your clients, Julie? Is that a manual process? Do you have some tools that you like? Talk a bit more about that.
Julie Bacchini Friedman: I’m old, so I like to do things manually. Part of the reason I like doing some of these tasks manually is that it helps me really put myself into the shoes of the person we’re trying to target, even if I’m not the target demographic. I’m a middle-aged white lady, so I might not be the right fit for what you’re selling. But I want to understand what someone in that demographic might see.
Tools are great, they can provide large data sets quickly. But to me, the most important thing is seeing things for myself. I make sure I search for some of the targeted keywords and experience it firsthand. I want to see what’s showing up in the search results and then click through to see where those links lead. I don’t think there’s any substitute for that kind of personal experience, something you can’t always pull from a tool.
Frederick Vallaeys: Yeah, and this goes beyond just the ads. It’s about the overall experience on the search page. What’s showing up organically? Are there AI-generated results? Is there a map result? All of this gives you more context. There’s one example I really like, which I often share in my presentations. One of the things I really appreciate about GPT and generative AI is that it’s great at putting itself in the shoes of different types of users.
For example, I’ll give it a business and ask, “Who might be looking to purchase this product or service?” Then, you get a list. The next prompt could be, “Tell me what value propositions might resonate with that audience,” or “What kind of keywords might they be searching for?” Now, you can start building custom audiences and better ads.
One specific example I had was asking GPT to write an ad for a florist service delivering in Paris. It suggested using the word “elegance” because, apparently, people in Paris and France like elegant things. I thought, “Yeah, that sounds reasonable, I’ll go with it.” The next day, I saw an Air France ad on TV, with four words in the ad: the brand and “redefining elegance.” It was amazing—Air France was using the same word GPT had recommended.
That’s where GPT can really shine. It helps you see things that might not be immediately obvious, even if the data is there. As a middle-aged white man, I can’t always see things from other perspectives, but GPT can. It’s incredibly useful in that way.
Julie Bacchini Friedman: Yeah, I think that’s true. And I think you can also use it to get a snapshot of your competition. Whether through auction insights or just doing searches and actually seeing the competition, you can ask GPT to summarize what these competitors are about, what they’re selling, and why people should buy from them. It’s an interesting exercise because a machine won’t have the same reactions as a human. It will base its responses on the data it has, not emotions or biases. That gives us a unique perspective on the competition.
There’s a lot of interesting ways to integrate and use this type of information in ongoing strategy management. I think people are just starting to figure out how to fully use these tools for this purpose.
Frederick Vallaeys: Andrew, what are your thoughts on things we should start doing, or how we can use generative AI in this context?
Andrew Lolk: I’ll skip the AI stuff just because we talk about it so much in our space these days. People are aware of it, but I still feel like it’s not quite there yet. It definitely does things. We use it to get different perspectives, act like a brand consultant, and help with ideas. But it’s still more of a one-on-one relationship where you get ideas in a chat. It’s tough to scale it across a large account and implement it effectively. So it’s still in the interesting realm, not yet fully actionable.
That said, I think the one thing everyone should start doing more of is understanding ad spend versus demand. This applies to how we manage channel efficiency across platforms like Meta and Google. One is for demand generation, the other for demand capture. We’ve seen interesting things where, especially during peak seasons like Christmas or Black Friday, Google could actually spend more efficiently than Meta, which is generally more effective for generating scale throughout most of the year.
Understanding when to shift spending across channels is key. For example, if you see more searches on Google during certain times of the year, like holidays, your spend on Google should increase. At the same time, it might make sense to lower your spend on Meta. This dynamic, like with swimwear ads in the summer or shopping campaigns near Christmas, shows that Google can often outperform Meta during certain peak times.
The real difference comes in how we position our businesses and use our budgets effectively. With access to the same smart bidding tools, we need to focus on smart budgeting decisions to stay ahead of the competition. That’s how we differentiate.
Frederick Vallaeys: That’s interesting because you’re talking about bidding across different channels. Let’s talk about channels a little bit more. Obviously, Google’s search market share in the United States has consistently dipped below 90% for the first time in a long time. Its hold on the global ad market has dipped below 50%.
So Google isn’t growing as fast anymore. It’s still the biggest player, but the landscape is changing. What do you think we should focus on for the future? Where should we start investing in 2025?
Andrew Lolk: I think there are two types of businesses out there: ones that are heavy on search and ones that are heavy on social. Most small to mid-sized businesses fall into one of these two categories. Of course, we have the big players like Expedia, Uber, or Coca-Cola, but they’re a different story. For businesses that focus on search or social, the key is integration. The days of seeing huge growth from Meta are behind us. Google’s CPCs are higher than ever, and the skill set required for Google doesn’t translate well to Meta, and vice versa. So it’s tough for businesses trying to manage both. But expanding into multiple channels is crucial, especially as the landscape shifts and these platforms become more challenging to navigate.
And there’s this huge opportunity with YouTube that people are completely missing. If I can find people on Google, I tell them every time – you’re missing out. You’ve totally screwed up YouTube. YouTube is the best ad channel, the best ad inventory in the world. And yet, you’re selling me clicks on things I can barely even understand. It’s painful to go through the spend. The threshold to get into YouTube is so high. We won’t even touch anything below $10,000 a month—$10,000, $20,000, $30,000 is where you really start seeing results from YouTube. It shouldn’t be this difficult to make YouTube work.
Frederick Vallaeys: And Andrew, let’s dive deeper because before we started recording, we talked a bit about your newsletter. You mentioned wanting to focus more on video sites. Now, that’s not necessarily PPC, but it’s still part of marketing. I think it’s relevant for all of us as business owners.
How do we make sure people know we’re there and ready to help? Blogs have been around for ages, so have newsletters and email, but video is relatively newer. How do you see that shift happening?
Andrew Lolk: I’ve been writing articles and blogs for as long as I can remember. I’m a good writer, and it’s always come naturally to me. But in the last year, with ChatGPT coming up, it’s getting harder and harder for me to justify spending a lot of time writing articles. A lot of the people who read my stuff have been following me for years, but when I look at new subscribers or people discovering who I am or what Savvy does, articles just don’t seem to be the way to reach new people.
With ChatGPT making it easier to write decent articles, it’s getting harder to stand out. The key now is to have personality, to show up, and to be seen. It’s about being consumable across different platforms. With a blog, you can read it on your phone or laptop, but with YouTube or video, it’s a whole different level.
I took my dog for a walk this morning and ended up listening to a story about why Manchester United is trying to buy a Danish left back because I thought it was interesting. But I’d never sit at my computer for 20 minutes just to read about that. It’s the same with getting a quick update on something, like what I think about Performance Max these days. You can just click on a quick video to get my thoughts without the commitment of reading a long article. Being able to share my opinions in a video rather than writing them out is a huge differentiator.
The real “wow” moment for us came when we switched all of our agency communication from written messages to bullet points plus Loom videos. The bullet points give a quick overview, but the Loom video, where I can show the data and explain what’s going on while smiling and interacting, adds so much more depth. It’s way more collaborative. I realized that just shooting articles out into the ether hoping people will read them isn’t as impactful as this more personal, engaging video approach. That’s why I think video will be even more important going forward.
This is how it is moving- our interpretation of the data, explained in a five to eight-minute Loom video with a screen share where they can see me smiling. It’s much better, much more collaborative. I realized that we were just shooting articles out into the internet, hoping people would read them, but this video approach gives so much more depth.
People get to know me much better this way, compared to just reading a written article. That’s why I believe video will be even more important going forward.
Frederick Vallaeys: Julie, I’d love to hear from you. And you’ve obviously made a large shift with PPC chat in recent history and having really been on Twitter before it was X and now it shifted to, I believe, slack. So do you talk to us about what you’re seeing in the social space?
Again, it doesn’t have to be PPC, but where are people these days, and how’s that causing you to shift your strategies in 2025?
Julie Bacchini Friedman: We just talked about this topic in today’s chat. If anyone wants to look at the archived version.
Frederick Vallaeys: Okay, we’ll put the link in the right area.
Julie Bacchini Friedman: It’s a really interesting time when it comes to developing and maintaining a community. This applies whether you’re building a community of customers for advertisers or trying to foster interaction with people. It feels more fragmented than ever. PPC Chat, for example, existed almost exclusively on Twitter for 10 to 11 years. When I took it over, it was primarily Twitter-based, but over the last few years, the environment has changed. People’s desire and hunger to use Twitter have shifted, so I’ve diversified PPC Chat into as many channels as people want to participate in. And it’s tiring as the person running it. I used to manage just two channels — LinkedIn and Twitter — but now I’m managing LinkedIn, Twitter, Threads, Blue Sky, Discord, and Slack. It’s a lot.
One thing we need to figure out, both individually and collectively, is how we want to interact with each other. How do we want people to interact with us? I think a big part of this depends on the demographics of who we’re engaging with. There’s a significant generational divide when it comes to preferences for video versus written content. If you’re aiming to engage with people who are around 40 and under, video is highly preferred — especially for those aged 30 to 35 and younger. But if you’re trying to reach people who are 45, 50, and older, video may not have the same impact as it does with the younger demographic.
I think it’s important to tune in and consider who we’re trying to reach and then develop strategies to effectively engage with those different groups. Video can play a major role in this, and if you’re sleeping on video in 2025, you’re likely going to regret it. But it’s also critical to have a strategy in place.
Frederick Vallaeys: I think that’s a great answer. You’re spot on with the different audience segments. But I also think technology has really come to the rescue here. I would put myself in the camp of, “Listen, I don’t want to watch a five-minute video. I can read three bullet points and get my answer right there.” However, as a creator, it’s really difficult to take those three bullet points and turn them into a video that satisfies people who want to consume video content.
On the other hand, if you start with a video, generative AI can create a beautiful transcript for you. It can also break it down into bullet points, making it consumable for everyone. So, the way I think about it is, what’s the most difficult content I can create once that then becomes the foundation for everything else, which can be somewhat automatically created? And that’s where I think focusing on video makes a lot more sense because you’re not really cutting off the older audience. You’re just starting at a different point that eventually leads to assets that the older audience will also prefer.
Andrew Lolk: That’s exactly what happened with us. I tried in the last six months of 2024 to go from, “Okay, I’ll write an article and use that as a script to create a video.” But I never found the time to actually create that video based on it. Like you said, though, if you create the video first, you can easily have ChatGPT use your voice to write an article based on it. That’s easy peasy.
I can’t get ChatGPT to make a video with me, though. Not even the notebook LLM is quite there yet. It’s very interesting, but I cringe on another level when I hear someone say, “Wow, was that the case with Google Quality Score? I didn’t know that.” It’s just not quite there. But yeah, I like the video-first approach.
And just as a note, the Missouri Star Quilt Company, which makes seven-figure and even nine-figure annual revenue, does it mainly through organic YouTube. They’re just teaching people how to quilt.
Julie Bacchini Friedman: It’s possible. You just have to find what works. Flexibility has always been important in this industry. You can’t be too rigid or married to, “This is how we’re going to do things, and this is how we’ll always do them.” We all know that doesn’t work. You have to be flexible in your thinking and approach and be willing to try new things.
Try stuff, right? Try it for your business first and see if it works. If you haven’t done much video, give it a try. Make some short videos. Jill Sask and Gales are my content heroes. She makes so much content; I don’t know how she does it. I’m convinced she doesn’t sleep.
Andrew Lolk: She’s incredible. She just does it. She’s incredible with the content.
Julie Bacchini Friedman: One of the big reasons she’s able to do that is exactly what you guys are talking about: starting with one piece of content and then slicing and dicing it into other pieces for different uses. I think thinking about that for ourselves, our businesses, and our clients is key. Having that mindset will be really helpful as we head into 2025.
Frederick Vallaeys: All right, let’s talk about the things you’re going to be doubling down on for 2025 and Julie, I’d love to start with you on this one.
Julie Bacchini Friedman: So, we’ve talked about video, and I think it’s an area I personally haven’t pushed hard enough. I work with lead generation clients, and they tend to be a little more video-averse compared to eCommerce folks. But I think it’s really important to get more buy-in and create more opportunities to work with video assets across different platforms.
Especially as AI makes it easier to generate written content, as Andrew mentioned earlier, creating AI-generated videos is much harder. So, even if they’re short, even if they’re just little snippets, getting some video content going is crucial.
Frederick Vallaeys: I agree, but to the point of it being harder to do video let me introduce you to my AI double.
The video avatar that talks in my voice has my face we’ll put a link up to this in the description so people can take a look and tell me how he’s called Little Fred.
Andrew Lolk: He’s called Little Fred. That’s what we call him at conferences.
Julie Bacchini Friedman: It’s scary. I’m scared. That part is a little intimidating, so I try not to dwell on it too much. You don’t want to overthink where that technology could go. But for brands, if we’re talking about the people we work with, the time for simply saying, “No, we don’t do video,” is probably not going to be sustainable much longer.
So, if you’re not already doing it, at the very least, you need to be talking about it, putting resources in place, and planning for it. It’s time to start integrating video into the mix.
Frederick Vallaeys: Makes sense. Andrew, let’s go to you.
Andrew Lolk: So I have three things I’d say. The number one thing is smart bidding and understanding playtime with it. Way too many PPC managers today just set a target, increase or decrease it, and don’t really know what’s happening. It’s crazy to me—and I think it’s crazy to those who’ve done manual bidding—to just let smart bidding run its course and see what happens.
There are so many cool mechanisms in the way smart bidding operates. As you increase or decrease targets, whether it’s by 10%, 20%, or 30%, it does different things. And the way it expands when you lower your ROAS during certain times of the year and adjusts bids is mind-blowing. You can achieve incredible results by experimenting with it.
So, I encourage everyone to dive into smart bidding and really experiment with it. At Savvy, we do a lot of audits, and it’s become a huge part of what we do for in-house teams. One of the most common conversations we have is about a “ROAS target” that’s been left untouched for ages. The response often is, “Why do you still run this? Oh, we’ve always run it.” That’s not a reason to keep it.
You have to rock the boat. Smart bidding relies on historical performance and keeps using it over and over. But if things change in your business, say, a brand or category has grown significantly in the past few months—it won’t magically catch up. If it learned months ago that a category didn’t perform well, it won’t just update unless you guide it.
You have to guide it. By decreasing the ROAS or grouping ROAS targets together through bidding portfolio strategies, it has to go out and chase things. So I would highly encourage playtime with smart bidding as a key skill that every PPC manager should master today.
The second thing is marketing efficiency ratio (MER). MER, which looks at your overall revenue versus marketing spend, has become extremely popular, and I, for one, love it. It helps overcome many tracking issues and bypasses the usual attribution conversation. I recommend doubling down on that. Go one step beyond simply saying, “Our revenue is a million, and we can spend 15%.” That conversation is too one-sided. It focuses too much on ads and doesn’t consider SEO or other factors.
You don’t always take affiliate, email, or direct traffic into account, which can be a benefit, but also its biggest downside. You’re not thinking about those other channels anymore. The entire conversation in 2024, especially in savvy marketing circles, was: “What are we spending on Meta? What are we spending on Google? What’s our blended ROAS?” But it’s too simple. You also need to consider net profit, contribution margin, new customers versus existing customers, and lifetime value optimization. You don’t have to do it all, just pick one of these aspects to add to the blended ROAS conversation, and you’ll find you have more levers to pull than you might have thought.
The third thing I would recommend is writing better ads. Every single person listening should be writing better ads. If you search for a keyword, see what everyone else is writing, and compare. The reason shopping ads perform better than search ads is that you can see the price and the product. For example, if I’m looking for a winter jacket for Denmark to survive the winter snowstorms (and not freeze in my San Diego spring jacket), I’ll see a bright rainbow-colored $3,000 jacket and immediately pass. But if I see a $400 technical jacket in olive green, from a trusted menswear brand, I’m more likely to click. I’ve already qualified myself as a buyer for that ad.
In search, you can create the same effect as shopping ads by including the price and specific qualifiers. If you have an ad that just says “menswear on sale,” I don’t know what I’m clicking on, which is why search doesn’t work as well for eCommerce. Everyone uses generic terms. Make it specific, like “street-inspired menswear” or “stretchy menswear”—or even niche options like “comfortable menswear.” Or offer package deals. Anything that will help the right person click on your ad. You’ll see much better results. Just double down on search and write better ads. Ask ChatGPT what could be written differently, it’s not that difficult.
Frederick Vallaeys: And this brings us back to blending social and search. Social is so good at understanding exactly what kind of menswear you prefer. Why can’t that exist in search, too? That’s one of my premises I’ve been talking about a lot. Large language models are basically starting to serve as our memories. By the way, I’m wearing this yellow B computer on my wrist, and it records everything that happens throughout my day.
Later, this conversation will turn into bullet points for me. If I go into a store and ask about the availability of SodaStream refill canisters, it will remember that too. It’ll tell me, “Hey, Fred, go online and order that.” As these models become more like our memories, I can go to a search engine, type something as simple as “refill,” and it will know exactly what I mean. If I’d just visited the store and mentioned a SodaStream refill, it’d know that’s what I’m asking for.
That’s how I think search will change due to large language models.
Andrew Lolk: I’m going to get divorced! If I could pull up a computer to show my wife that I was right, I’d do it, but I wouldn’t advise that.
Julie Bacchini Friedman: I would not do that.
Frederick Vallaeys: My wife insists that I push the little button, and then it turns red and stops listening.
Andrew Lolk: I love it. I can’t wait for the day when I don’t have to remember anything or explicitly tell AI something. I can’t wait. My wife and I were talking about food yesterday. She was making a bowl of great fried rice, and we were out of toasted sesame oil. Google is listening to everything we say, so it should just add it to the shopping list. We’ve bought it before, and it’s been on the list before. It’s always listening anyway, so just add it. Don’t make me add it myself. Don’t make me tell it to add it. But I’m stuck right now because it doesn’t recognize my voice. Anyway, just add it to the list.
Frederick Vallaeys: It’s fascinating. I wear this B. computer device on my wrist, and at the end of the day, I get my reports. I never realized how much something can know about you just by being present throughout the day. That example you just gave is brilliant, but wait until you see what else it starts to learn about you by simply being on your wrist. It’s both scary and exciting. I want to try it out and see where it leads. 2025 is going to be very exciting, and Julie, I think you said it— we don’t know what’s coming, but whatever it is, it’s going to bring change, for better or worse. It’s definitely a time of change, but let’s hope it’s for the better.
Thank you both for sharing your perspectives on what to stop, start, and double down on. This has been a great episode. Julie, people can find you at Neptune Moon and also on PPC Chat.
Julie Bacchini Friedman: Yes. Anything else from usual on, I’m on all the platforms so you can find me Neptune Moon on all the socials.
You can find me on LinkedIn, Neptune Moon ,
PPC Chat Logo . I’m not hard to find if you wanna track me down and talk about any of this stuff,
Frederick Vallaeys: And we’ll include those links in the description as well. Awesome. When I asked my executive assistant for Julie’s email, she said, “Wait, is this the person on LinkedIn?” I said, “Yes, that’s the one! You can’t miss her.” Andrew, where can people get in touch with you?
Andrew Lolk: The number one thing is to just follow me on LinkedIn. I post a lot of content there. Right now, I enjoy scrolling through LinkedIn daily and calling out the ridiculous things people are saying, especially when they’re just blatantly wrong.
A lot of it is from ChatGPT, and I like to point out where they’re missing the mark. If you’re looking for some clarity in Google Ads, LinkedIn is where you’ll get the most from me.
Frederick Vallaeys: Andrew, I’m going to hire you as my fact checker. It sounds like there’s a big future in that.
Andrew Lolk: There’s a YouTube video series that comes up Andrew debunks stuff
Frederick Vallaeys: I love it. We’ll put that in the notes as well.
Thanks to both of you and thanks to everyone for watching. We hope you enjoyed this episode. Please subscribe, leave a comment on what you liked or what you’d like to see more of, and we’ll see you next time for another episode of PPC Town Hall. Take care!