When Search Engine Land celebrated its 20th anniversary, they sat down with our co-founder and CEO Frederick Vallaeys for a wide-ranging conversation on 20 years of pay-per-click (PPC). I watched it expecting a nostalgic trip through industry history. What I got instead was a clear framework for thinking about the AI transition — and a few Optmyzr origin stories I hadn’t heard told quite like this before.

If you haven’t watched it yet, here’s the full video:

But if you want the distilled version with some of my own thoughts on what it means for advertisers today keep reading.

What most people get wrong about AI in search

Most advertisers use AI tools in one move: ask a question, get an answer, decide the answer isn’t useful, move on. Fred’s read on this is blunt:

“Most people stop at step one of a ten-step process and then conclude that AI isn’t that useful.”

His own approach, shaped by 20+ years of thinking in systems, starting with his electrical engineering degree, is to treat AI the same way you’d treat any other tool: define your inputs, specify your outputs, design the process in between. When he evaluates keywords, he doesn’t ask for an opinion. He asks for a structured output (JSON, XML, a table) that he can pipe into something else, run across hundreds of accounts, and present in a format that means something.

The second thing he raised is less about tools and more about how we’ve been trained by Google to ask bad questions. A decade of one-input, one-output search has conditioned us to strip every request down to the bare minimum. AI assistants work the other way around, they need context to be useful.

Fred’s example: if you’re searching for a mattress, you’re not really searching for a mattress. You want to sleep better. Tell the AI that and it’ll give you an answer that might include the mattress or might tell you three things that will make a bigger difference first. The same logic applies to every campaign brief you write and every target you set.

“Give it that information, because then it’ll be better at what it does.”

It’s the same principle we’ve built into Optmyzr’s approach to PPC automation — giving the system enough business context to make decisions that connect to real goals, not just bid targets.

From Blockbuster cassettes to Google Ads: how PPC actually started

Most PPC origin stories start with Google. Fred’s started at a Blockbuster video store in 1998.

As a Stanford student looking for pocket money, he noticed that Blockbuster was selling used cassettes for $10 each, tapes that cost $100 new. He bought in bulk, listed them on eBay, and hit his first distribution problem: how do you find buyers at scale without a real budget?

That question led him to GoTo.com, a search engine that let advertisers bid on keywords and pay per click. He bought keywords, drove traffic, made money, the entire modern PPC funnel, before most people had heard the term.

The timing is what makes this remarkable, not the cleverness. Fred was running cost-per-click (CPC) campaigns a full year before Google even had ads. By the time he joined Google in 2002 to set up their sixth language ads team, in Dutch (because he’s originally from Belgium), he wasn’t learning a new concept. He was helping build infrastructure around something he’d already pressure-tested himself.

What “early Google Ads” actually looked like

At the time Fred joined Google’s AdWords team, a $30,000-per-month spend made you a tier-one advertiser. The upper echelon. That number would barely register as a mid-size account today.

The early AdWords ecosystem was tiny, manual, and primitive in ways that are hard to imagine now. Fred shared a story about a client whose team had to come in early each morning to press a button to turn ads on, then stay late to turn them off:

“She’s the one who’s in charge of day parting. She has to come in early in the morning to push the button to turn the ads on, and she’s the last one at night to leave to turn the ads off by pushing a button.”

No automated bidding. No conversion tracking out of the box. No search terms report. You bought a keyword, you got clicks, and what happened next was a complete black box unless you’d built your own tracking technology.

Worth sitting with, for anyone frustrated about Performance Max opacity or declining search term visibility right now: we have been here before. The industry started in a black box, opened up progressively with tools like Google Analytics and the search query report, and is now cycling back toward opacity, just with far more sophisticated machinery underneath. The pattern is cyclical, which means the current opacity isn’t permanent.

The comment that started Optmyzr

After leaving Google in 2012, following a decade on the AdWords team, Fred wrote a piece for Search Engine Land about quality score, how to think about it and how to build a Google Ads script to monitor account-level quality score. Two people, Manas and Geetanjali, commented. They’d been building API-based tools and saw exactly the problem Fred was describing. A call followed:

“Within half an hour we decided to start Optmyzr together.”

The rest, for those of us who work here, is very much felt history. And account-level quality score, the metric that sparked all of it, remains one of the most underused diagnostic tools in PPC. Google only reports at the keyword level. Aggregating that into an account-level score gives you a directional read on whether your overall account health is moving toward what Google rewards or away from it. It’s something we’ve built into Optmyzr from the beginning.

The three moments that actually reshaped PPC

Fred has lived through hundreds of Google Ads updates. Three came through clearly in the interview as genuinely pivotal, the kind that changed the game rather than tweaking it.

Conversion tracking and Google Analytics

Before Urchin became Google Analytics and before Google built conversion tracking into the platform, search advertising ran on faith. You spent money, traffic happened, you hoped for the best.

The decision to change this came from Eric Schmidt. As Fred tells it, Schmidt’s logic was simple and commercial:

“The way that Google doubles its revenue is by proving that every dollar you put in leads to more dollars coming out. So why wouldn’t we do this?”

They bought Urchin, built the infrastructure, and within a few years every serious advertiser had conversion data they could act on. It didn’t just improve campaigns, it validated the entire model of performance advertising.

Smart Bidding becoming actually good

When Smart Bidding first launched, most experienced PPC managers dismissed it. Then it got better. Then it became the dominant approach across the industry and that created an existential moment for us at Optmyzr, as it did for every third-party tool vendor.

“There was a discussion: why should we pay for a tool when Google does all of these things so well for free?”

Optmyzr’s answer then, and now: Smart Bidding handles the optimization, but it doesn’t handle the guardrails, the context, or the recovery. When automation works, great. When it breaks, you need levers. Fred described how advertisers put it themselves:

“I want PPC insurance. Give me the guardrails, give me the controls to bring it back to where it needs to be.”

That’s exactly what Optmyzr’s Rule Engine and monitoring tools are built for. Smart Bidding and human oversight aren’t in opposition, they work best together, one handling scale and the other handling judgment.

ChatGPT going public

Fred is unambiguous about this one. The day ChatGPT launched publicly was the moment that lit a fire under everything, not because Google wasn’t already doing the work, but because being the incumbent changes your constraints:

“Google has some of the smartest engineers in the world. The Google teams invented what later became the transformer, the model that OpenAI then perfected. It’s not like Google was late to this game. Google invented this game.”

But Google couldn’t have released something as rough-edged as early ChatGPT without significant regulatory and reputational consequences. OpenAI could, and did. The result is that the entire advertising ecosystem is rebuilding around a fundamentally different model. Users no longer just type keywords, they prompt, they converse, they expect context.

Google Ads is, at its core, a keyword advertising system. Nobody has a clean answer yet to how you connect ads to users who no longer think in keywords. But it’s the question that will matter most over the next decade.

How Google changed the definition of success in paid search

Success in paid search used to mean cheap clicks. Now it means profit on ad spend. That shift didn’t happen overnight:

• Early days: Cost per click (CPC) was low, volume was the goal.
• Mid era: Conversions entered the picture. Then cost per acquisition (CPA). Then return on ad spend (ROAS).
• Now: The conversation is shifting to incrementality and profit on ad spend — metrics that connect directly to business outcomes, not advertising intermediaries.

Fred’s framing on this was direct:

“The metrics that advertisers look at have just come closer and closer to the actual business metrics that drive a business and not just some intermediate thing.”

Each step brought advertisers closer to measuring what actually matters. The frustrating part is that each step also came with more complexity, more data requirements, and more ways for something to go wrong.

Optmyzr’s Q1 2026 State of Google Ads Benchmark Report found the same thing playing out in the data: impressions falling as AI Overviews reshape the search engine results page (SERP), clicks holding. Each impression is carrying more weight than it used to. If you’re still optimizing purely for volume, you’re measuring the wrong thing.

What the next 20 years of search actually rewards

Fred’s answer to this cuts through a lot of the noise:

“Were you a keyword manager? Was that your ability? Or were you trying to find new customers for that company? Focus on what you were trying to achieve and figure out the new means of doing that.”

The channel is a means to an end. In the end, connecting people with products and services that solve their problems, doesn’t change. The means will keep evolving, from keywords to prompts, from static search to personalized AI conversations that remember who you are and what you’ve asked before. The advertiser who holds their ground is the one who stays focused on the customer and adapts the channel strategy around that, not the other way around.

Fred also credited the community, Search Engine Land, SMX, the broader PPC practitioner network with compounding his thinking and his career in ways he didn’t plan for. Optmyzr itself started in a comment thread. His early AI writing found its audience through publication reader data. The feedback loops from publishing, attending events, and having honest conversations with people dealing with the same problems, those matter more than most people admit.

“At the happy hour, people tell me after a drink: Fred, here’s what really frustrates me. They don’t tell you that any other way. And when you get that, it’s like I can actually do something about that.”

What to take into your next campaign

1. The black box isn’t new and it won’t last. PPC started without transparency, opened up, and is cycling back. Understanding the pattern makes the current frustration easier to navigate and harder to overreact to.

2. Set guardrails alongside your automation. Smart Bidding, Performance Max, and whatever comes next all work better with controls alongside them. Set a performance alert for any Smart Bidding campaign where CPA moves more than 20% in a seven-day window.

3. Prompt with context, not just queries. Tell the AI (or your campaign tool) what you’re actually trying to achieve. More context produces better outputs.

4. Move your success metrics closer to the business. Cheap clicks were never the goal. Neither is ROAS in isolation. What business outcome does that ROAS actually connect to?

5. Stay inside a community that gives you real feedback. The advertiser who compounds is the one who stays connected to ideas, practitioners, and publications where honest reaction happens, not just polished case studies.

Take the next step

At Optmyzr, we’ve built everything around one principle Fred has held since day one: advertisers should be in control of their automation, not the other way around. If the last 20 years of PPC have proven anything, it’s that the tools change constantly but the need for visibility, guardrails, and intelligent management never goes away.

Whether you’re running Smart Bidding across dozens of accounts, managing Performance Max with limited controls, or trying to make sense of an increasingly AI-shaped SERP, Optmyzr gives you the oversight layer that keeps everything working the way you intend.

Start your free 14-day trial → No credit card required. Cancel anytime.