Ask Skip: Automation Made Advertising Faster. Not Necessarily Smarter

Why does every ad platform talk about automation like it automatically makes advertising better?

— Operating Partner

Because automation is a hell of a sales pitch.

It sounds efficient. It sounds modern. It sounds like progress. It lets every platform say some version of the same thing: give us more inputs, more budget, more creative, more permission, and more control, and our system will find the performance you were too manually exhausted to find yourself.

That pitch works because part of it is true.

A lot of advertising should be automated. Nobody should be nostalgic for manual trafficking hell, spreadsheet archaeology, 47 naming conventions, or junior media buyers spending half their week adjusting bids. Advertising has too much repetition, too many surfaces, too many formats, too many signals, and too many workflows for humans to hand-crank the whole machine.

Automation can absolutely make advertising better. It just doesn’t do that automatically.

That’s the part the platforms tend to skip.

Automation doesn’t remove judgment from advertising. It moves judgment upstream. Someone still has to decide what the system is optimizing toward, what data it can trust, what creative it’s allowed to use, what tradeoffs it can make, what counts as success, and when the machine is producing performance versus just producing a cleaner excuse.

More automation can mean better advertising. And it can also mean bad assumptions at industrial speed.

Automation Still Needs Someone to Own the Outcome

The useful version of automation is boring, which is usually why it works.

It can help with pacing. It can clean up trafficking. It can manage bid adjustments. It can resize assets. It can rotate creative. It can reduce repetition. It can handle reporting grunt work. It can surface anomalies faster than a human staring at a dashboard while slowly losing the will to live.

The problem starts when the industry takes automation that’s good at tasks and starts treating it like it’s good at responsibility.

Those are different things.

A system can allocate spend. It can’t tell you whether the strategy deserves the budget. A system can generate more creative variations. It can’t tell you whether the idea was any good before it got multiplied into 4,000 slightly different disappointments. A system can optimize toward conversions. It can’t automatically know whether those conversions are incremental, valuable, loyal, profitable, or just the same customers the brand would’ve reached anyway with fewer fees and less theater.

Automation can execute a plan. It can also hide the fact that the plan was mediocre.

That’s why “automated” and “better” should never be used like synonyms. Automation is a capability. Better advertising is an outcome. Many people confuse the two because capabilities are easier to sell than outcomes are to prove.

The Platform’s AI Has a Boss

Every automated ad system has an incentive structure.

It may be dressed up in clean product language, but the system still lives inside a business. That business has revenue targets, margin goals, auction dynamics, inventory to move, measurement products to defend, and shareholders who did not invest because they enjoy advertiser transparency as a philosophical concept.

That doesn’t make the platform evil.

It means the platform’s automation isn’t neutral just because the interface looks clean.

When a platform says its AI will optimize performance, the first question should be: performance for whom, measured how, and according to whose definition?

If the platform controls the targeting, bidding, placement, creative assembly, reporting, and attribution, it has a lot of room to define success in a way that keeps the spend flowing. The system may genuinely improve outcomes. It may also move budget into inventory the platform wants to monetize, claim credit through measurement the platform controls, and present the whole thing as optimization.

The best marketers understand this. They use platform automation, but they don’t confuse the platform’s recommendation engine with their own strategy. They know the machine has a job. Their job is to make sure that job is aligned with the business, not just with the dashboard.

Bad Inputs Don’t Become Strategy Because AI Touched Them

Automation has a way of making weak thinking look sophisticated.

Give a system bad conversion signals, and it’ll optimize toward bad conversion signals. Give it lazy creative, and it’ll produce more ways to serve lazy creative. Give it a shallow audience definition, and it’ll chase shallow proxies. Give it a business objective that nobody bothered to translate into media terms, and it’ll happily optimize around whatever was easiest to measure.

Automation will optimize against whatever inputs and incentives you give it. If those inputs are weak, the system won’t pause to ask whether the goal makes sense. It’ll just pursue the wrong thing faster and with more confidence than a human team could manage manually.

That’s useful when the inputs are disciplined. It’s dangerous when the inputs are garbage. And let’s be real, advertising has a long and proud history of feeding garbage into expensive systems, then being amazed when the output smells familiar.

Automation can find patterns. It can test combinations. It can process signals faster than any human team. But it doesn’t magically know that the promo offer is training customers to wait for discounts. It doesn’t know that the creative has flattened the brand into beige paste. It doesn’t know that the landing page is doing more damage than the media plan. It doesn’t know that a cheap conversion is cheap for a reason.

Humans still have to ask those questions.

That’s the part some organizations are trying to automate away because it’s uncomfortable. It requires taste, judgment, commercial discipline, and enough backbone to challenge a platform recommendation even when the chart looks happy.

More Versions Can Mean More Average

Creative automation is where this gets especially slippery.

On paper, it sounds great. More assets. More variations. More formats. More personalization. More testing. More relevance. And more speed.

In practice, more versions often means more average.

The industry already had a problem with creative being overprocessed before it ever reached the audience. Automation can make that worse by multiplying weak ideas before anyone has stopped to decide whether the idea is worth scaling.

More assets don’t automatically mean more relevance. More variations don’t automatically mean more learning. More testing doesn’t automatically mean the work has a stronger point of view.

Sometimes automation just helps average work travel farther, faster, and with better reporting.

Automation can help creative teams when it removes repetitive labor around resizing, formatting, localization, tagging, assembly, and trafficking. It can help brands learn which messages work in which contexts. It can help teams move faster without asking humans to spend their lives making platform-specific cutdowns until their souls exit through the shared drive.

That’s good.

But automation shouldn’t become a substitute for a point of view.

The danger is that brands start optimizing creative around what systems can easily generate, test, and score. Clear product shot. Familiar phrase. Safe offer. Template-friendly layout. Platform-approved call to action. Slightly different background. Repeat until the dashboard appears satisfied.

The work may become more efficient while also becoming forgettable enough to qualify as visual white noise.

Advertising still has to make someone notice, care, remember, laugh, feel, trust, understand, or act. If automation helps do that, great. If it just produces more assets for people to ignore, the machine is busy, not useful.

The Human Didn’t Disappear. The Job Moved

Automation doesn’t eliminate the need for people as much as it changes where human judgment has to insert itself.

The work moves away from constant manual adjustment and toward setting the right objectives, protecting the inputs, defining the guardrails, judging the creative, questioning the measurement, and knowing when the system is optimizing toward something that looks good in a dashboard but doesn’t actually help the business.

The old media-buying job involved a lot of manual control: settings, bids, budgets, placements, reports, adjustments, approvals, screenshots, exports, and endless platform maintenance that mattered less than the time it consumed.

Automation reduces some of that, but raises the value of people who understand systems.

People who know how to define the right objective. People who can separate platform-reported performance from actual business impact. People who can interrogate incrementality. People who can recognize when the system is exploiting a loophole. People who can build guardrails, control groups, exclusions, measurement frameworks, creative standards, and escalation paths before the machine starts improvising with real money.

That’s not less expertise. That’s different expertise.

The best operators in an automated environment aren’t button-pushers. They’re translators between business strategy and machine execution. They understand enough about the platform to use it, enough about the brand to protect it, enough about the customer to respect the context, and enough about measurement to avoid clapping every time a dashboard says “efficient.”

That’s a rarer skill than the industry wants to admit.

It’s also why automation doesn’t make weak teams strong. It often exposes them faster.

CTV Is Where Automation Can Help, and Where It Can Get Dumb Fast

Streaming advertising is a perfect test case.

CTV has too much fragmentation for purely manual management. Different apps, platforms, devices, data sets, measurement layers, ad servers, programmatic pipes, content contexts, and audience states all collide inside what the viewer experiences as a simple ad break.

Automation can help make that better.

It can reduce repetition. It can improve pacing. It can manage frequency across more surfaces. It can match creative to context. It can make ad loads smarter. It can help determine whether a pause screen, a mid-roll break, a live sports moment, or a second-screen exposure should be treated differently.

That’s the useful version.

The dumb version treats every impression as raw material for yield extraction and lets the machine optimize around what’s easiest to sell.

That’s how viewers get the same ad six times, buyers get a report full of completed impressions, and everyone pretends the system is working because the campaign technically delivered.

CTV doesn’t need more automation just for automation’s sake.

It needs better decisioning around the actual conditions of the ad experience. What’s the viewer doing? What’s the content moment? What’s the interruption cost? What’s the creative asking from the viewer? Is the ad built for full attention, partial attention, sound-on, sound-off, passive awareness, or action?

Automation can help answer those questions at scale, but only if the business cares enough to ask them.

Automation Makes Accountability Harder to Locate

Automation doesn’t just change how advertising runs. It changes where accountability goes.

When a human buyer makes a bad call, at least everyone knows who to blame. Maybe that person had bad data, bad incentives, or a bad day, but the decision has a location.

Automated systems blur that.

Was it the platform’s model? The agency setup? The advertiser’s conversion signal? The creative feed? The measurement window? The campaign objective? The attribution logic? The budget constraints? The optimization event? The landing page? The brand safety settings? The inventory source? The AI-generated asset nobody reviewed closely enough before it reached the market?

The answer is usually yes.

That ambiguity is convenient for the people selling the system.

When things go well, automation gets credited as intelligence. When things go sideways, everyone starts talking about inputs, learnings, edge cases, model behavior, and the need for additional calibration.

Funny how the machine is a genius during the case study and a misunderstood intern during the postmortem.

That’s why advertisers need clearer ownership. Not just of the campaign. Of the system around the campaign.

Who approved the objective? Who validated the data? Who reviewed the creative outputs? Who set the exclusions? Who checked the placements? Who decided the measurement standard? Who has the authority to shut it down when the automation starts optimizing toward nonsense?

The Real Question Is Control

The automation story sounds like margin expansion, product improvement, and labor efficiency all rolled into one. For platforms, that’s powerful. More automated tools can make the ad system easier to use, more scalable, more profitable, and harder for advertisers to leave.

That’s a great business.

For advertisers, the question is whether the automation creates more control or quietly asks them to give control away.

There’s a difference between using automation to improve your operating leverage and using automation to outsource your judgment.

The first makes the business stronger. The second makes the business dependent on someone else’s black box.

You don’t have to reject automation to worry about that. In fact, the smarter you are about automation, the more you should care about control. What are the inputs? What are the guardrails? What can be audited? What can be overridden? What is the system allowed to change? What does the platform report, and what does an independent measurement layer confirm?

Those aren’t anti-automation questions.

They’re adult supervision.

The platforms want advertisers to believe the machine gets smarter as it gets more control. Sometimes it does. Sometimes it mostly gets better at absorbing budget, expanding definitions, and explaining itself after the fact.

Skip Says

Automation should make advertising easier to operate. It shouldn’t make it harder to understand.

The machine can help with execution, testing, pacing, versioning, targeting, and optimization. That’s useful. But automation doesn’t replace strategy, creative judgment, clean inputs, measurement discipline, or the need to ask whether the campaign created actual value.

The best marketers won’t reject automation.

They’ll stop treating it like adult supervision.

Because faster advertising isn’t automatically smarter advertising. Sometimes it’s just the same bad idea with better throughput.

Ask Me Anything

Whether you’re fed up, fired up, or just want the truth behind the trends, send me your questions using this form. Anonymity guaranteed. Bullshit not included.