The Most Dangerous Word in Your Marketing Stack Isn’t “AI” — It’s “Optimization”
Every creator I know has a love-hate relationship with paid social. You can post organically until your fingers cramp, but the moment you want predictable growth — the kind that compounds into actual revenue — you have to open the advertising spigot. And that’s where the romance dies.
I’ve managed campaigns for my own newsletter, for client accounts, and for a short-lived e-commerce experiment that I still have nightmares about. The pattern is always the same: you set up a campaign, Google or Meta’s algorithm takes over, and suddenly you’re staring at a dashboard that tells you everything is “learning” and nothing is performing. Your cost per acquisition creeps up. Your creative fatigue sets in. And the platform’s automated recommendations? They’re designed to make them money, not you.
So when I saw a Product Hunt launch for something called GoodLads, an “AI growth manager for your Google Ads account,” my first instinct was skepticism. Another tool promising to “10x your results”? I’ve seen a hundred of those. But then I read the thread, and something caught my attention — not the product itself, but the philosophy underneath it. The makers aren’t claiming to be smarter than Google’s algorithms. They’re claiming to be different from them. And in a world where every platform is pushing you toward full automation, that distinction matters more than you might think.
This isn’t just a story about Google Ads. It’s a story about the broader tension every social media operator faces: how much control do you hand over to the algorithm, and how much do you keep for yourself? The answer GoodLads is betting on — that human oversight plus AI-generated hypotheses beats either one alone — has implications far beyond search advertising.
The Problem Nobody Wants to Admit: Your “Strategy” Is Just Platform Recommendations
Let me paint a picture that might feel uncomfortably familiar. You’ve got a Google Ads account running for your content business or your client’s e-commerce store. You set it up weeks ago, maybe months. You check it every few days, nod at the metrics, and wonder why your cost per acquisition hasn’t improved since the first week.
The platform tells you your campaign is “limited by budget.” The obvious fix, according to Google’s own recommendations, is to increase your budget. So you do. And your CPA gets worse. So you increase it again. And somehow, you’re spending more money than ever while your results stay flat or decline.
This is the trap that Pavel Kucherbaev, one of GoodLads’ makers, describes in the launch thread. He worked on launching and optimizing Google Ads for his own product, TimeTuna, and found the experience more complicated than filing a tax return in the Netherlands — which, for anyone who’s dealt with Dutch bureaucracy, is saying something. His cofounder, Yannick Veys, has 15+ years of experience optimizing Google Ads for various companies, and he’s developed know-hows that often contradict Google’s own recommendations.
Here’s the specific example that made me stop scrolling: if your campaign is limited by budget, Google recommends increasing the budget. Yannick’s counter-intuitive insight? Reducing your target cost per acquisition could improve performance with the same budget. It’s a fundamentally different diagnosis of the same symptom. Google says “you need more fuel.” Yannick says “your engine is inefficient — fix that first.”
This resonates with me deeply because I’ve seen the same dynamic play out across every platform. TikTok’s algorithm tells you to post more frequently. Instagram tells you to use more Reels. LinkedIn tells you to engage with comments within the first hour. All of these recommendations might be true in aggregate, but they’re not necessarily true for you. Platforms optimize for their own engagement metrics. You optimize for your business outcomes. Those two things are not always aligned.
The deeper problem, though, isn’t that platform recommendations are wrong. It’s that they’re lazy. They’re the default answer, the one-size-fits-all suggestion that requires zero understanding of your specific account, audience, or economics. When you rely on them, you’re not running a strategy — you’re just following instructions.
What GoodLads Actually Does (and Why It’s Different From What You’d Expect)
Let me break down what this product actually is, based on the launch thread and the makers’ comments. GoodLads is positioned as an AI growth manager for Google Ads accounts. The team claims it’s specifically scoped for accounts spending between $10,000 and $250,000 per month — which tells you immediately that this isn’t a tool for hobbyists or small creators testing their first campaign. This is for people who are already spending real money and need help getting better returns.
The workflow, as described in the thread, works like this: the AI analyzes your account, generates hypotheses for experiments, and presents them on a kanban board. You review each hypothesis, decide which ones to run, and approve any changes before they go live. The system tracks what was tested and what happened, giving you a clear record of your experimentation history.
That kanban board is the detail that caught Evan Taft’s eye in the comments, and honestly, it caught mine too. The makers admit it wasn’t their first choice, but it felt natural for tracking what’s going on. And I think that’s actually the smartest design decision in the whole product.
Here’s why: the biggest problem with paid advertising isn’t that you don’t know what to do — it’s that you don’t remember what you’ve already tried. I’ve been guilty of this myself. I’ll run a campaign variation, see mediocre results, and then three months later, I’ll have the “brilliant” idea to try the exact same variation again. Without a system for tracking your experiments, you’re doomed to repeat your failures. The kanban board isn’t just a project management feature — it’s a memory system for your marketing brain.
The other design decision worth noting is the approval requirement. Dylan Friddle praised this in the comments, saying he wouldn’t want anything touching an ad account on its own. The makers confirmed that some users actually prefer a one-click apply button, but they’ve deliberately kept human approval in the loop. Pavel explains: “We wanted to make sure that AI helps with the feedback, summary, and coming up with new experiment ideas, but we wanted to make sure that the actual changes in Google Ads only happen with a user approval.”
This is the right call, and I’ll explain why in the next section.
Why the “Human-in-the-Loop” Approach Is the Only AI Strategy That Makes Sense
There’s a lot of hype right now about fully autonomous marketing. Let the AI run your ad accounts, let it write your captions, let it schedule your posts, let it respond to comments. Just set it and forget it. And look, I get the appeal — I’m tired too, and the idea of handing my content calendar to a bot sounds like a vacation.
But here’s the thing I’ve learned from actually testing these tools: full automation fails not because the AI is dumb, but because it lacks context. An AI can analyze your Google Ads data and tell you that your search terms are too broad. It cannot tell you that your best-selling product is seasonal, or that your audience responds better to emotional storytelling than feature lists, or that your competitor just launched a price war that changes everything.
The makers of GoodLads seem to understand this. They’re not trying to replace the marketer — they’re trying to augment them. The AI generates hypotheses and summarizes feedback, but the human makes the final call. This is what Grace Gui appreciated in the thread: “I like the idea of always having something new to test instead of wondering what to change next.” The AI isn’t making decisions for you. It’s making sure you never run out of ideas.
This is the same philosophy I’d apply to any AI content tool, whether it’s Canva for design, CapCut for video editing, or Metricool for analytics. The AI should be your brainstorming partner, your first draft generator, your data analyst — not your replacement. When you hand over full control, you lose the ability to course-correct based on the intangibles that only a human can perceive.
There’s also a trust issue that goes beyond strategy. If an AI makes a change to your ad account and that change fails, who’s responsible? If an AI posts something offensive on your brand’s Twitter account, who apologizes? The answer is always you. So it makes sense to keep a human in the loop for anything that touches your money or your reputation.
Where the Math Breaks: Statistical Significance in a Fast-Moving World
One of the most interesting exchanges in the launch thread was about how long experiments take to reach a verdict. Nivy asked a sharp question: “How long does a hypothesis sit on the kanban before it gets a verdict?”
Pavel’s answer reveals the statistical reality underneath the product: “We follow a statistic testing, so it depends on the number of clicks, and the size of the impact. If the impact is huge (e.g. 25% improvement) — it could be visible within a week, if the impact is modest (e.g. 5%), it might take 4 weeks, and very small impact might not be detectable at all. The recommendation is to make sure the experiment runs for at least 2 weeks.”
This is honest, and I appreciate that. But it also reveals a tension that every advertiser faces: statistical significance takes time, but markets move fast. If your experiment runs for four weeks to detect a 5% improvement, the competitive landscape might have changed entirely by the time you get your answer. Your competitor might have launched a new product. A viral trend might have shifted consumer behavior. The algorithm might have updated three times.
This is why I’d argue that the kanban board is actually more valuable than the AI itself. The AI generates hypotheses, but the board gives you a historical record that lets you spot patterns over time. You might not be able to act on a 5% improvement within the two-week window, but you can look back at six months of experiments and see which types of changes consistently moved the needle. That meta-learning is worth more than any single optimization.
How This Compares to the Incumbents (and What They’re Missing)
The Google Ads optimization space isn’t empty. There are established players like Optmyzr, Kenshoo, and WordStream that have been doing automated optimization for years. There are also agency-focused tools that offer similar features. So what makes GoodLads different?
Based on what I can see from the launch, the difference is in the framing. Most optimization tools approach this as a data problem: they analyze your account, find inefficiencies, and either recommend or automatically apply fixes. GoodLads approaches it as an experimentation problem: it’s not just finding what’s wrong — it’s generating a continuous pipeline of things to test.
This distinction matters because it changes the user’s relationship with the tool. With a traditional optimizer, you’re waiting for the tool to tell you something is broken. With GoodLads, you’re being fed a steady stream of hypotheses, which keeps you in an active testing mindset. Ethan Blake captured this in his comment: “I can see this being useful when you’re managing an account and just don’t have the time to keep coming up with new things to test.”
The makers also claim that Yannick’s know-how often differs from Google’s recommendations, which positions the product as an independent voice rather than a Google sycophant. In an era where Google’s own recommendations are increasingly self-serving — pushing you toward automated bidding, broad match, and higher budgets — having a tool that challenges the platform’s default advice is genuinely valuable.
But here’s where I’d push back: the launch thread doesn’t disclose much about the actual AI models or how they’re trained. The makers mention combining AI models with the Google Ads API and Yannick’s expertise, but the specifics are vague. How does the AI generate hypotheses? Is it pattern-matching across many accounts, or is it using rule-based logic derived from Yannick’s playbook? The answer would tell me a lot about whether this tool will be genuinely useful or just another layer of abstraction on top of Google’s own suggestions.
Why TikTok Creators Should Care More Than LinkedIn Ones
You might be thinking: “I’m a content creator, not a Google Ads advertiser. Why should I care about this?”
Here’s my answer: the underlying philosophy applies to every platform where you’re trying to grow. Take TikTok, for example. The algorithm is notoriously opaque, and creators often feel like they’re guessing in the dark. You post a video, it gets 200 views, and you have no idea why. The platform’s advice is always the same: post more, use trending sounds, hook viewers in the first three seconds.
A tool like GoodLads, applied to TikTok, would approach it differently. Instead of giving you generic advice, it would generate specific hypotheses: “Your last five videos all used the same hook structure, and the ones that varied performed 20% better. Try testing three different hook styles this week.” It would track your experiments on a kanban board, showing you what you’ve tested and what the results were. And it would require your approval before implementing any changes to your content strategy.
LinkedIn creators, on the other hand, might find this approach less useful. LinkedIn’s algorithm is more predictable, the audience is more professional, and the content formats are more constrained. The experimentation space is narrower — you’re mostly testing different angles on professional insights, not radically different content structures. For LinkedIn, a simpler tool that tracks engagement metrics and suggests posting times might be sufficient.
The point is that the methodology of continuous experimentation applies everywhere, but the tooling needs to be platform-specific. GoodLads is built for Google Ads because that’s where the makers have their expertise. But the lesson — that you should be running a structured experimentation process rather than relying on platform recommendations — applies to every creator and marketer.
What Creators and Social Media Teams Can Borrow From This (Even If You Never Touch Google Ads)
Let me give you some practical takeaways that you can apply this week, regardless of whether you ever sign up for GoodLads.
First, build your own experimentation kanban. You don’t need fancy software — a Trello board or even a spreadsheet will do. The structure is simple: one column for hypotheses, one for active experiments, one for completed tests, and one for results. Every time you try something new — a different content format, a new posting time, a different hook style — add it to the board. Write down what you expected to happen and what actually happened. After a month, you’ll have a record that shows you what’s working and what isn’t, based on your own data rather than platform recommendations.
Second, question platform recommendations. When Google tells you to increase your budget, ask yourself if that’s actually solving your problem or just making Google more money. When TikTok tells you to post more frequently, ask yourself if that’s based on your account data or on aggregate trends. When Instagram tells you to use more Reels, ask yourself if Reels actually drives the outcomes you care about. The platforms are not your enemy, but they’re also not your business partner. Their recommendations serve their interests first.
Third, adopt the “small impact might not be detectable” mindset. Pavel’s comment about statistical significance is a lesson for all of us. Most of the changes you make to your content strategy will have small effects — a 2% improvement in engagement, a 1% increase in click-through rate. These are real but hard to measure, and they’re easy to miss if you’re not tracking systematically. Instead of chasing dramatic wins, focus on building a consistent testing cadence. The compounding effect of many small improvements over time is where the real growth happens.
Fourth, insist on human approval for AI-driven changes. This is non-negotiable. Whether you’re using AI to generate content, schedule posts, or optimize ads, you need to review everything before it goes live. The AI might be brilliant at generating ideas, but it doesn’t understand your brand voice, your audience’s sensitivities, or your business goals. Keep a human in the loop, even if it slows you down.
Where I’d Push Back: Limitations and Open Questions
For all the things GoodLads gets right, there are some limitations I’d want to explore before committing my ad budget to it.
The pricing is not disclosed. The launch thread mentions a 50% promo code (“MAKEITEASY”) for the Product Hunt launch, but the actual price isn’t listed. For a tool scoped to accounts spending $10,000-$250,000 per month, the pricing could be anywhere from a few hundred to several thousand dollars per month. That’s a significant investment, and I’d want to know what the ROI looks like before signing up.
The AI’s recommendation quality is unproven. Yannick has 15+ years of experience, and that’s genuinely valuable. But translating that experience into AI models is a different challenge. How many accounts has the AI been trained on? What’s the track record of its hypotheses versus Google’s recommendations? The launch thread doesn’t provide any case studies or performance data, so I’d be going in on faith.
The statistical testing approach has inherent limitations. As I mentioned earlier, detecting small improvements requires long experiment windows, and markets move fast. The tool might be great for accounts with high traffic volumes where statistical significance is achievable quickly, but less useful for smaller accounts where you’d need weeks or months to get reliable data.
The scope is narrow. This is a Google Ads tool, period. If you’re running campaigns across Google, Meta, TikTok, and LinkedIn — which most serious advertisers are — you’d need separate tools for each platform. The lack of cross-platform integration is a significant gap.
Who is this NOT for? If you’re spending less than $10,000 per month on Google Ads, this tool is probably overkill. You’d be better served by learning the fundamentals yourself or working with a freelancer. If you’re running campaigns on platforms other than Google, this tool won’t help you. And if you’re looking for a fully autonomous solution where you set it and forget it, the human-approval requirement will feel like a burden rather than a feature.
What I’d Watch / Test Next
If you’re intrigued by the philosophy behind GoodLads — even if you never use the product itself — here’s what I’d do this week:
Set up your own experimentation tracker. Start simple: a spreadsheet with columns for date, platform, hypothesis, action taken, and result. Commit to testing at least one new thing per week. It doesn’t have to be a big change — a different hook, a new posting time, a revised thumbnail. The goal is to build the habit of systematic experimentation.
Audit your platform recommendations with fresh eyes. Go through your Google Ads, Meta Ads Manager, or TikTok analytics and look at the platform’s suggestions. For each one, ask yourself: is this recommendation serving my goals or the platform’s goals? Write down your answer. You might be surprised by how many recommendations you’ve been following without questioning.
If you do manage Google Ads accounts, consider testing GoodLads. The 50% promo code “MAKEITEASY” makes it a lower-risk experiment, and the kanban board alone might be worth the price of admission. But go in with clear expectations: track your own metrics, compare the AI’s hypotheses against your own judgment, and be willing to abandon it if the results don’t justify the cost.
Watch the comments section. The Product Hunt thread has some genuinely useful feedback from people who are clearly experienced in paid advertising. The makers are responsive and seem genuinely interested in building something useful rather than just cashing in on the AI hype. That’s a good sign, but it’s not a guarantee of success. Give it six months and see if the product evolves in response to user feedback or if it stagnates.
The bottom line is this: GoodLads is solving a real problem, but the specific solution matters less than the underlying philosophy. Continuous experimentation, human oversight of AI, and skepticism toward platform recommendations — those are the principles that will serve you well no matter what tools you use. Whether GoodLads becomes a staple in your ad stack or fades into the Product Hunt graveyard, the lessons from its approach are worth taking to heart.
