Why most people never get good at ads

The wrong mental model

Most people think advertising is a targeting problem. They spend hours obsessing over audience segments, campaign objectives, bidding strategies, and placement settings. When results disappoint, they assume the algorithm is broken, the platform has changed, or they haven't found the right audience yet.

They're not wrong that these things matter. They're wrong about how much they matter — and about what the real lever is.

The best advertisers in the world think about almost none of that. They think about human behavior. About whether their offer is actually compelling. About whether the creative earns attention in the first two seconds. About the economics: does acquiring a customer at this cost make sense given what they're worth?

How ad managers got commoditized

In 2010, running paid ads was genuinely hard. You had to know where audiences were online, negotiate CPMs with publishers, manage insertion orders, understand reach and frequency curves, and manually optimize placements based on performance data you were reading yourself.

It required skill, experience, and relationships. Agencies charged large retainers for it. The knowledge barrier was real.

Then Meta and Google spent the next fifteen years automating every single one of those decisions. Audiences? The algorithm finds them. Placements? Automated. Bidding? Smart bidding handles it. Creative rotation? Automated. Budget allocation across campaigns? Automated.

The tactical layer — the part that used to require years of training — has been almost entirely replaced by software. What's left is setup, budgets, and creative. And anyone can click "Boost Post."

The algorithm ate media buying

Meta's Advantage+ campaign type doesn't ask you to specify an audience. You give it a creative, a budget, and a conversion event — and it handles everything else. Google's Performance Max works the same way: one campaign, all placements, automated bidding, automated targeting.

This isn't a simplification for beginners. This is the recommended setup from both platforms for most advertisers. The evidence suggests that broad targeting, fed with good creative and good conversion signals, genuinely outperforms granular manual targeting for the vast majority of accounts.

What the algorithm needs to work well is:

• A large enough conversion event pool (usually 50+ conversions per week to exit the learning phase)
• Clean, accurate tracking — so it knows what a conversion actually looks like
• Good creative — so it has something worth showing people

Two of those three things are engineering problems. One of them is creative. The media buyer's traditional job is almost entirely gone.

Creative is the last bottleneck

Here's the uncomfortable truth about modern advertising: distribution is solved. The algorithm will find your audience. Bidding is solved. Placement optimization is solved. Attribution is improving every quarter.

The one thing that hasn't been automated is the creative itself — the actual message, the hook, the offer, the emotional trigger that makes someone stop scrolling and pay attention.

That's where the game is now. Not in the settings. In the creative. And creative is downstream of something even more fundamental: understanding people. Why do they buy? What do they fear? What do they aspire to? What friction stops them? What framing shifts their perception?

AI can help generate creative at scale. But it needs a strong brief. And a strong brief requires understanding the human on the other side of the screen — which the algorithm cannot give you.

Media buyer vs. growth operator

These two roles often share a job title, but they have almost nothing in common in terms of how they think about their work.

A media buyer manages spend. They monitor campaigns, adjust bids, check frequency caps, write reports. They measure success by CPM, CPC, and CTR. They optimize within the platform. Their job is to not waste budget and to follow the platform's best practices.

A growth operator thinks about the entire acquisition system. They understand the offer (is this compelling enough for someone to pay?), the creative (does this earn attention and create desire?), the landing page (does this convert the interest into action?), and the economics (does acquiring a customer at this CAC make sense given LTV?). They measure success in revenue, MER, and payback period. Their job is to find scalable, profitable growth.

• A media buyer optimizes a campaign. A growth operator builds an acquisition system.
• A media buyer reports on what happened. A growth operator diagnoses why and changes what comes next.
• A media buyer improves within constraints. A growth operator challenges the constraints.

The industry is full of media buyers. There's a permanent shortage of growth operators.

Why engineers have an unfair advantage

Something interesting has happened in growth marketing over the last five years: many of the most effective growth operators have engineering backgrounds. Not because engineering skills are required — but because the mental models transfer almost perfectly.

Engineers are comfortable with systems thinking. They don't see a campaign as a one-off execution; they see it as a component in a larger feedback loop. Engineers are comfortable with uncertainty and iteration — "ship fast, measure, fix" maps directly to "test fast, measure, iterate."

Engineers also build things. When a manual process is slow or error-prone, they automate it. When data is scattered, they build a dashboard. When a decision is made by intuition, they instrument it. Traditional marketers often lack the instinct — or the ability — to do this.

• Engineers think in feedback loops — which is how effective ad testing works
• Engineers are comfortable reading data — which is what optimization is
• Engineers build tools — which is what separates operators who scale from ones who plateau
• Engineers treat hypotheses seriously — which is the foundation of real experimentation
• Engineers have low tolerance for manual repetition — which drives the automation that creates leverage

None of this means you need to be an engineer to be a great growth operator. But adopting the mental models — systems thinking, hypothesis-driven iteration, automation instinct, data fluency — will move you far ahead of peers who haven't.

What to actually learn

If you want to get genuinely good at advertising in 2026, stop spending time on:

• Campaign structure hacks ("the right objective for each goal")
• Audience segmentation tactics that Meta's algorithm ignores anyway
• Bidding strategy optimization that smart bidding handles better than you
• Platform-specific settings that change every six months

Spend that time instead on:

• Human psychology: Why do people actually buy? What fears, desires, and identity signals drive decisions?
• Offer construction: Is the thing you're selling genuinely compelling? Is the risk-reversal strong enough? Is the value obvious?
• Creative strategy: How do you earn attention in two seconds? What hooks, angles, and emotional triggers work for this audience?
• Unit economics: What can you afford to pay to acquire a customer, given what they're worth over time?
• Attribution and measurement: Are you actually measuring what you think you are? Is your tracking accurate?
• Systems thinking: How do you build a repeatable engine that produces creative, tests it, learns, and iterates — at speed?

The platforms will keep changing. The hacks will stop working. The fundamentals — human behavior, compelling offers, strong creative, sound economics — will compound for as long as advertising exists.