State of PPC 2026: Live from SMX Munich w/ Wijnand Meijer(TrueClicks) & Frederick Vallaeys(Optmyzr)

Episode Transcript

00:00:00 - 00:00:23
Mike: Welcome to another episode of Growing Ecommerce. I’m one of your hosts, Mike Ryan. Chris isn’t with us today; instead, I’m joined by two very special guests here at SMX Munich. And we’re here for a special reason: we are all partners working on something called the PPC Survey, which you can find at ppcsurvey.com.

00:00:24 - 00:00:50
Mike: The others are on the screen behind me, but for listeners, I’ll read them out for you. We partnered with DataFeedWatch, God Tier Ads, Hero Conf, Optmyzer, ProductHero, smec, smec Advanced, and TrueClicks. That’s an awful lot of brainpower, and we all worked really hard on this report together. But Wijnand, I’m going to hand it over to you. Maybe you could tell us a little bit more about the PPC Survey.

00:00:50 - 00:01:09
Wijnand: Yeah, sure. Thanks for having me, Mike. We started this about four years ago. It was born out of a need to know the stuff we’re going to cover today: what are the challenges, what are the adoptions, and all that stuff that PPC people do.

00:01:09 - 00:01:29
Wijnand: We are with a group of partners which has expanded over the years. We run a survey usually in Q4. We did a lite version last year—no big report like we have this year—but two years ago was a big one. I think we had about 1,100 people filling it out, which was already the largest survey that I know of in our industry.

00:01:29 - 00:01:48
Wijnand: Now it was 1,306 to be exact, which makes the data significant. As people will see when they read the report, most questions are specific to an audience: agency questions for agencies, and specialist questions for specialists. That way we don’t bother people with questions they cannot answer or are not relevant to them.

00:01:48 - 00:02:06
Wijnand: We can compare segments geographically—how people in the US are doing things compared to Europe—or big spend versus small spend. I think that makes it extra interesting instead of just looking at high-level numbers. Part of this is also the Top 50, which creates a lot of buzz.

00:02:06 - 00:02:23
Wijnand: This podcast should come out before the Top 50, which lists the most influential PPC experts. That’s the final question on purpose; people have to finish the survey to be able to vote on their favorite experts.

00:02:24 - 00:02:37
Wijnand: PPC Hero used to do that, and I thought it would be cool to use the same audience. Let’s let people fill out a survey and vote on their favorite experts so we have one project to do both. This year’s survey is the biggest: 14,000 words and 60-plus charts.

00:02:37 - 00:02:57
Wijnand: If you want to read it in one sitting, you’ll spend around 45 minutes, depending on how fast you read. It’s 72 pages. There are a lot of images, so don’t worry, it’s not just 72 pages of text. The table of contents is quite detailed, so you can pick and choose what’s relevant.

00:02:57 - 00:03:14
Mike: I’ve got it here in front of me, and you can definitely pick and choose. You don’t have to read the whole thing. You can read one chapter a day with a coffee every morning. You got the AI summary, right? How many of those words were written by AI, by the way?

00:03:14 - 00:03:33
Wijnand: Well, it went to the partners who divided up the questions. I can tell you that I actually used AI to make sure it doesn’t look too much like AI, which is very ironic.

00:03:33 - 00:03:56
Mike: Mine was definitely all written by me because I got some feedback that certain sections were too academic and hedging our bets on everything. Routine listeners know I love making qualified statements, but I also love flying off the handle and making overly bold statements sometimes. But a lot of human brainpower went into this for sure.

00:03:56 - 00:04:12
Mike: Credit where it’s due, Wijnand is the main driver of the report. This company sponsored it and we filled in, but if it wasn’t for Wijnand herding the cats, this wouldn’t have happened.

00:04:12 - 00:04:26
Wijnand: It’s a good thing we did it multiple times so I can just reuse the playbook. Two years ago, it was a lot of work to create the survey for the first time. Now we could reuse some questions, although this year we have many new questions around AI. Chapter eight is dedicated to AI, based on more than ten questions we didn’t have last time.

00:04:27 - 00:04:52
Mike: It’s definitely not a repetition of last time. Before we jump in, we’re going to talk through a few other data points. Just to highlight, some of the chapters include ad platform adoption, spend goals and challenges, agency-client relationships, AI in PPC, and scripts and software. It’s quite comprehensive. Are you guys ready to chat about the charts?

00:04:52 - 00:05:13
Frederick: Yeah, what has the data told us?

00:05:13 - 00:05:42
Mike: I’ll kick it right off. This is one of my favorite questions. We saw that 53% of practitioners—over half—say that managing PPC is harder than it was two years ago. Meanwhile, Google and other platforms have these highly automated strategies. They say the goal of automation is to make advertising easier, but at least half of the advertisers don’t feel that way. Only about 16% say it is easier. What do you think is going on here, Frederick?

00:05:42 - 00:05:58
Frederick: I think it goes back to the fact that while Google has more than 1,300 advertisers, there is certainly a bias here. The people who filled out the survey are practitioners and experts—the people who do this for a living.

00:05:58 - 00:06:15
Frederick: Senior management got on stage a couple of years ago at Google Marketing Live and said advertising needs to be for everyone. To make it for everyone, they had to take away the complexity of manual bid management and figuring out every single keyword.

00:06:15 - 00:06:35
Frederick: We’ve seen all this automation creep in, culminating in PMax and IMax, which is basically, “Write us a check and we’ll do the rest.” That’s great for many advertisers who’ve never advertised before, but it’s bad for the experts who want control and expect things to be a certain way.

00:06:35 - 00:07:02
Mike: I agree. Google’s opportunity has always been that long tail of SMB advertisers. Most large spenders are already on Google, so what can they do to make it easier for smaller advertisers to spend more? Wijnand, what’s your take?

00:07:02 - 00:07:22
Wijnand: Actually, it was your quote at the end of that “Why it’s getting harder” paragraph that made me think. Maybe we should ask it differently next time. When people think about “harder” or “easier,” what are they actually thinking about? Are they thinking about the practical management of campaigns, or how hard it is to reach their goals?

00:07:22 - 00:07:40
Wijnand: Reaching goals is getting harder due to competition and rising CPCs, which has less to do with how easy it is to get a campaign running. Maybe we should split the question: is it harder to do the practical job, or is it just harder to make the client happy?

00:07:40 - 00:07:57
Wijnand: I think that’s part of the story.

00:07:57 - 00:08:16
Mike: For sure. If we think of the hyper-granular campaigns and SKAGs we used to do, it’s hard to say objectively that things aren’t functionally easier now. But I love the phrase “leveling the playing field.” Google’s technology takes away specific tactics and promotional controls, which lifts up SMBs but “cuts down the tall poppies.”

00:08:16 - 00:08:42
Frederick: Exactly. I think it’s the “tall poppies” who feel the frustration. Business conditions outside of PPC also matter. The first report was around COVID, and this one comes during economic uncertainty. No matter what we do in PPC, outside factors make the economy tough. Marketing teams are told to drive more business, but there’s only so much you can do in this climate.

00:08:42 - 00:08:58
Mike: Let’s move on. Wijnand, you analyzed a section about how agencies operate. We found a contradiction: teams managing over $1 million a month are working with 20% fewer people, yet 1 in 5 agencies still charge by billable hours, and 39% think that’s a future-proof model.

00:08:58 - 00:09:23
Mike: You said this is basically giving agencies a pay cut for getting better at their jobs.

00:09:23 - 00:09:42
Wijnand: In general, that’s the case. The billable hour persists for certain reasons, but it has a big disadvantage. If you do your job faster and you’re transparent about it, you end up charging the client less because you became more efficient.

00:09:42 - 00:10:13
Wijnand: Senior people are also faster, so the incentives with billable hours are totally misaligned. It’s becoming less common, but it’s geographically different. In the US and Australia, it’s between 4% and 12%, but in Europe, it’s around 25%. You just have to replace it with something else, like a flat fee.

00:10:13 - 00:10:34
Wijnand: The question specified “ongoing PPC management,” which should be predictable work. I understand using an hourly rate for one-off, custom projects where you don’t know how long they will take.

00:10:35 - 00:10:51
Wijnand: But if you’ve done something a hundred times, you don’t need to reinvent the wheel. You need a tiered model or something performance-based. Agencies often use hours to protect against “scope creep” when clients ask for extra calls or tasks.

00:10:51 - 00:11:10
Wijnand: If you act like a lawyer and charge for every extra hour, you’re protecting your margins. But in my agency years, we often just did the work without complaining, so the hourly model wasn’t really protecting us anyway. Agency leadership needs to think about alternatives.

00:11:10 - 00:11:29
Wijnand: I love the Peter Drucker quote: “All profit is derived from risk.” The more risk you take as an entrepreneur, the more profit you can make. The hourly rate puts all the risk on the client side because the agency gets paid regardless of the result.

00:11:29 - 00:11:50
Wijnand: The opposite is being paid only for a certain result. Injury lawyers in the US take a third of the settlement if they win, which is a big portion of the value captured. Agencies should decide when to take more risk on a client-by-client basis.

00:11:50 - 00:12:05
Wijnand: Especially with AI speeding everything up, hourly billing makes even less sense.

00:12:05 - 00:12:19
Mike: It’s striking that anyone considers it a future-proof model. Frederick, anything to add?

00:12:19 - 00:12:34
Frederick: I share the same view. I don’t understand how this fits in a landscape with so many tools and so much automation. If I use a script to manage search terms, why would I do that if it gets me paid less?

00:12:34 - 00:12:56
Frederick: In-house teams are often more innovative because many agencies aren’t financially incentivized to take risks or work faster. Why do more work if you’re not rewarded for efficiency?

00:12:56 - 00:13:11
Mike: We’ll talk more about AI adoption later. Let’s move to PMax performance. 48% of respondents complained about a lack of granular control. It’s natural that people are “hacking” the system.

00:13:11 - 00:13:27
Mike: 37% are running PMax alongside standard shopping, and 17% are forcing “feed-only” builds to push PMax toward shopping placements. Are these hybrid workarounds a permanent reality or just a Band-Aid until Google shuts them down?

00:13:27 - 00:13:49
Frederick: Anecdotally, we’re seeing more people shift back to old-school campaign types. We’ve all experimented with PMax, but now there’s a realization that we need more control and tactics. As long as Google supports those older campaign types, people will use them.

00:13:49 - 00:14:08
Frederick: Google’s philosophy has always been incremental. Have your search and shopping campaigns, then put PMax on top to capture what they didn’t. I don’t think this is a Band-Aid; it’s an ongoing strategy.

00:14:08 - 00:14:22
Mike: We’ve seen PMax stagnating a bit as advertisers shift back toward standard shopping. The percentage of advertisers running PMax and shopping side-by-side is increasing.

00:14:22 - 00:14:39
Mike: There are risks, like bid escalation if products appear in both, though Google says not to worry. Cannibalization is clearly happening, and PMax often swoops in to take queries you expected your other campaigns to handle.

00:14:39 - 00:14:59
Mike: I recently presented data showing evidence of cannibalization and CPC hikes. Even so, 65% of PMax users are satisfied with the results. It’s on par with standard shopping in terms of satisfaction.

00:14:59 - 00:15:22
Mike: Google improved transparency over the last couple of years. PMax used to be just “budget and landing pages,” but now there are brand exclusions, search themes, and more. All these decisions add to the mental load.

00:15:22 - 00:15:46
Frederick: Frederick, you wrote that Google always tries to reach high automation, finds they have to walk it back, and then tries again. It’s a cycle. Something is supposed to be the future, but then advertisers ask for controls to avoid outcomes they don’t like. Eventually, the automated system becomes just as difficult to manage as what it replaced.

00:15:46 - 00:15:59
Mike: Speaking of control, exact match is still way more popular than Broad Match (IMax). Advertisers also prefer efficiency constraints like setting a ROAS target rather than just “Maximize Conversion Value.” Why keep the guardrails on?

00:15:59 - 00:16:26
Wijnand: Predictability is nice. People are used to exact match and trust it most for starting campaigns. Most practitioners get a target ROAS or CPA from their client, so they use that as a guardrail. Other options feel risky and unpredictable.

00:16:26 - 00:16:43
Wijnand: IMax (Demand Gen) scored low, but it was new when the survey was taken. People might still be learning best practices. Sometimes IMax traffic goes to the Search Partner Network, which isn’t always high quality.

00:16:43 - 00:17:01
Frederick: If you take away control, you’re playing in a bubble where anything can happen. Guardrails provide sensibility. If you set a tROAS, your CPC bid could technically be higher than the cost of the product you’re selling, and you could still hit your average target. We want guardrails to prevent that one click from being completely unprofitable.

00:17:01 - 00:17:35
Mike: Let’s talk about AI adoption. Practitioners self-reported saving only 5.2 hours a week using AI. That’s not the “transformative 20-plus hours” promised by people like Sam Altman. Is it overhyped, or are we just not using it correctly?

00:17:35 - 00:18:05
Frederick: It’s crazy that people save less time with AI than with scripts. A well-designed script takes your process and applies it deterministically. AI has promise, but it’s currently like a Rube Goldberg device—you have to walk behind the ball and push it back on course every time it goes off.

00:18:05 - 00:18:30
Frederick: However, every six months the models get mind-blowingly better. The risk is that people try it once, fail, and give up. I’m strict with my company about talking about where AI fails, because it’s usually because the user didn’t give enough context or instruction.

00:18:30 - 00:18:54
Wijnand: Five hours is a good starting point if you just started with basic use cases. That’s 10% of your work week. I don’t think you can fire 10% of your team yet, but 24% of people say they save 6 to 10 hours, which is significant.

00:18:54 - 00:19:13
Wijnand: Most people are just doing basic stuff like generating headlines. Since the survey, powerful new models like Claude have gained traction. I think the numbers will go up as people get better at training them.

00:19:13 - 00:19:35
Frederick: Meeting recordings are a big one. For agencies, a meeting often results in more work. A meeting recorder can actually be your prompt generation system. People are figuring out how to evolve their workflows to save time.

00:19:35 - 00:19:55
Mike: Even if the tech is ten times better tomorrow, it takes time for people to understand and trust it. The world only changes so fast.

00:19:55 - 00:20:13
Frederick: It also depends on whether a company has someone in charge of figuring this out, or if they’re stuck in a billable hour model where there’s no incentive to be more effective.

00:20:13 - 00:20:33
Wijnand: Clients are already asking agencies if AI is making them work faster. There’s an expectation there.

00:20:33 - 00:20:49
Mike: 70% of respondents say quality and accuracy is their biggest AI challenge. Hallucinations and bad creative are real issues. What should a QA process look like? Babysitting AI can be more frustrating than just doing the work yourself.

00:20:49 - 00:21:18
Frederick: You have to figure out why it got it wrong. Twelve months ago, the model might not have been capable, but now they are good at following instructions. It’s often human error. It’s a struggle because you just want to get your work done and go home, not spend two hours figuring out why a prompt broke.

00:21:18 - 00:22:15
Wijnand: You can use the same prompt twice and get different answers, which is counter-intuitive for technology. You should still use predictable, rule-based systems for parts of your work. AI isn’t always the solution; sometimes a script is better. If the AI doesn’t know something, it should just tell you.

00:22:15 - 00:22:42
Frederick: Models were trained like a multiple-choice test where guessing has a 20% chance of being right. We need to retrain them or tell them in the prompt, “If you’re not sure, tell me.” You can even tell the AI, “My boss will make me cry if you get this wrong,” and it sometimes works harder.

00:22:42 - 00:23:06
Mike: Humans can be “confidently wrong” too, but machines don’t have body language to tip you off. We need deterministic systems for things like budgets where a 20% error could cost $200,000.

00:23:06 - 00:23:22
Wijnand: In e-commerce, that’s why there’s a focus on AI product feeds. Feeds provide the structured data needed to reduce bad outcomes. Interestingly, 62% of survey respondents aren’t using any third-party tools for analysis or optimization, relying entirely on the Google Ads UI or Ads Editor.

00:23:22 - 00:23:47
Wijnand: Ads Editor is great, but the UI is slow. I was surprised that Google Sheets and Looker Studio score the highest for reporting. They are free and customizable, so freelancers and small agencies start there. But even for larger spends, the usage of paid tools isn’t as high as I expected.

00:23:47 - 00:24:09
Frederick: One-third of people managing Google Shopping don’t use any feed management solution; they do everything in Merchant Center. It’s tough to get people to pay for a tool.

00:24:09 - 00:24:26
Frederick: Jeff Allen once spent six hours doing analysis in the UI. The client asked for one more level of detail—the mobile segment—and he had to spend another six hours redoing the work. Tool vendors exist to make you efficient so that same task is just a one-click button.

00:24:26 - 00:24:44
Wijnand: Look at your processes and see where you’re wasting time on repetitive tasks. Start with a script, and if you hit limits, try a tool.

00:24:44 - 00:25:01
Frederick: Be careful of “snake oil.” Look at the people behind the tool and their advertising experience. Changing a budget isn’t hard, but the “scar tissue” of ten years of experience helps you consider the edge cases. That’s the difference between a tool built in a few days versus one built over a decade.

00:25:01 - 00:25:25
Mike: Renting a tool means the maintenance and infrastructure are someone else’s problem. Meanwhile, 22% of professionals are using AI to build their own no-code apps and scripts. Frederick, you mentioned that AI won’t replace specialists, but a specialist using AI will replace one who isn’t.

00:25:25 - 00:25:46
Frederick: “Vibe coding” is a broad term, but people are now successful at getting a working script on the first attempt. You can turn your optimization process into a script or a custom report without doing it manually every time.

00:25:46 - 00:26:08
Wijnand: Miles and the PPC Mastery guys are already offering courses on how to build your own agents and tools for internal use. I encourage that.

00:26:08 - 00:26:36
Mike: 80% of respondents worry about users shifting to tools like Perplexity or ChatGPT for search, which could damage ad visibility. I’m not as worried. Google is increasingly dominant and they are an advertising company. They have the power to dictate the pace of how AI services roll out to ensure ads continue to exist.

00:26:36 - 00:26:53
Wijnand: These platforms will find ways to monetize. If ChatGPT ads take off, that becomes a new platform for PPC specialists. Advertising budget goes where the eyeballs are. We’ve been used to a monopoly for too long; maybe we just need to diversify.

00:26:53 - 00:27:18
Frederick: Gemini is doing much better now. The bigger risk is moving further away from keywords and landing pages toward a world where you just say, “Buy it for me.” Demand will still exist—plumbers will still be needed and people will still buy clothes—but how we connect to that demand will shift.

00:27:18 - 00:27:37
Mike: Let’s wrap up with one bold prediction each. Where will the PPC industry be when we run the survey again?

00:27:37 - 00:27:53
Frederick: I think we’ll see a new segment of “AI-Amplified Marketers” doing things they couldn’t do before because they didn’t have the time or the specific skill. Those restrictions will be gone.

00:27:53 - 00:28:10
Wijnand: Most people wish they could clone themselves to get more work done. I think in a year, every professional will have built a “coach” or “clone” in an AI-based tool to speed up everything they do.

00:28:10 - 00:28:23
Mike: I’m an identical twin, so I already have a natural clone—it’s overrated! But check out ppcsurvey.com to see the full report. It’s free and contains a wealth of information we barely scratched the surface of.

00:28:23 - 00:28:34
Mike: Thanks to all the partners, especially Wijnand for being the main driver. Thanks also to SMX Munich for letting us record here. We’ll see you next time on Growing Ecommerce. Visit smarter-ecommerce.com for more.