Insights From Last Year’s Best Debates For 2026

Internet hosting my podcast, provides me a front-row seat to the unfiltered actuality of our business, the gritty, “in-the-trenches” actuality shared by consultants who handle thousands and thousands in spend.

Final yr, my company, together with Greg Finn, Christine “Shep” Zirnheld, Julie Friedman Bacchini, Andrew Lolk and Shawn Walker didn’t maintain again. They dismantled “finest practices,” known as out platform biases, and highlighted precisely the place the algorithms fail with out human arms.

Listed here are the ten most fascinating (and typically uncomfortable) issues my company shared final yr that you would be able to take ahead for 2026.

1. Google Is “Shaking The Sofa Cushions” (And You’re The Sofa)

We have to cease pretending Google’s incentives are completely aligned with ours. As Greg Finn and Christine Zirnheld from “Advertising and marketing O’Clock” identified, Google is, finally, a for-profit firm, and whereas it stays an essential promoting companion, its targets don’t all the time completely align with what’s finest for advertisers.

Finn put it completely: Have you ever ever observed that the “Suggestions” tab all the time suggests elevating your price range however by no means reducing it? That bias is actually constructed into the UI. With CPCs hitting record highs, “success” for the platform typically simply means “extra income” for the shareholder.

And we’ve seen this play out within the information. Optmyzr analyzed greater than 17,000 Google Advertisements accounts and located no consistent correlation between a high Optimization Score and strong performance. The truth is, lots of the best-performing accounts improved not by accepting Google’s suggestions, however by selectively rejecting them and focusing solely on fixes that really moved CPA, ROAS, or revenue.

So, the takeaway is easy: Cease treating suggestions as gospel. Deal with them as upsells, as a result of the info exhibits that blindly following them doesn’t reliably assist advertisers, however it does reliably assist Google.

2. Automation With out Guardrails Is Simply “Excessive-Velocity Waste”

The consensus from Shawn Walker from Symphonic Digital, Finn, and Julie Friedman Bacchini, President & Founding father of Neptune Moon & Managing Director of PPC Chat, was unanimous: AI can execute, however it can not strategize.

Walker famous that with out strict conversion high quality thresholds, Smart Bidding inevitably chases “low cost junk leads” as a result of they’re the simplest conversions to get. In the meantime, Julie warned about “algorithm drift,” the place a marketing campaign slowly expands into irrelevant search phrases as a result of it thinks it’s being useful.

Automation is critical for contemporary account administration, however that doesn’t imply “set it and neglect it.” Your job isn’t shifting bids anymore, it’s designing the correct layers of automation and guardrails so the algorithms work, not the opposite manner round.

I just lately examined how nicely AI may diagnose a drop in conversions, and it confidently recognized the marketing campaign’s restricted price range because the trigger. The reasoning was easy: Some key phrases have been nonetheless receiving impressions whereas others weren’t, so the price range have to be the bottleneck. However that didn’t make sense. Funds constraints often have an effect on campaigns broadly, decreasing visibility throughout the board fairly than selectively shutting off particular person key phrases.

When solely sure key phrases go darkish, the extra believable clarification is a bidding concern. And if bidding is automated, that always signifies the algorithm deems these key phrases as decrease high quality, leading to lowered bids and finally, disappearing impressions.

The larger level is that this: AI typically solutions with confidence earlier than it solutions with accuracy. It could actually completely assist you to refine advert copy or strengthen relevance, however it nonetheless struggles to know the nuanced and infrequently counterintuitive interdependencies inside a PPC account. In different phrases, it could help with execution, however it’s not but as dependable as a strategist.

3. The “Rule of 30” Is The New Regulation Of Gravity

Some of the sensible takeaways of the yr got here from Walker. We frequently debate how a lot information Good Bidding wants, however Shawn gave us the maths:

You want ~30 conversions per marketing campaign, per 30 days.

Not per account. Not throughout shared budgets. Per marketing campaign. Beneath that threshold, the machine is simply guessing. If you happen to’re questioning why your small campaigns are unstable, it’s not dangerous luck; it’s dangerous math. You’re ravenous the algorithm.

In Optmyzr’s 2024 research on the impact of bidding strategies on performance, we noticed the identical. 50+ conversions per thirty days are ultimate. 30+ is nice, and something much less isn’t nice. Nevertheless, I want to add one refinement to Shawn’s level. The actual threshold isn’t “30 conversions per marketing campaign,” however sufficient quantity on the conversion objective/actions Good Bidding is optimizing towards. Google’s techniques can use broader, account-level conversion patterns to cut back information shortage, and account-default targets and portfolio methods are designed to increase the training set past a single marketing campaign.

What really issues is having sufficient quantity for the motion you need Good Bidding to optimize towards. If a number of campaigns are all working towards the identical conversion occasion, they’ll successfully “pool” their learnings. In different phrases, campaigns don’t must hit 30+ conversions individually so long as the underlying conversion motion has sufficient combination quantity for the system to study and make dependable selections.

4. “Delicate Conversions” Are The Spine Of SMB Success

So, what do you do in case you can’t hit that magic variety of 30? It’s a must to feed the beast one thing else.

Friends closely advocated for shifting up the funnel. Walker detailed the need of “engaged customer” indicators, customized metrics like a consumer scrolling to a sure depth or spending time on web site, fired solely as soon as per distinctive consumer to stop inflation.

Whether or not it’s a PDF obtain, an add-to-cart, or a pricing web page go to, these “tender” indicators are not optionally available crutches; for smaller accounts, they’re the one strategy to generate sufficient information density for Good Bidding to operate.

In different phrases, micro conversions nonetheless matter. They offer Good Bidding a richer sequence of intent indicators to study from: Did the consumer examine merchandise? Did they view pricing? Did they return inside 24 hours? Did they have interaction with interactive instruments? In my expertise, these micro-signals are what stop smaller accounts from ravenous the algorithm and finally assist it acknowledge high-quality customers earlier within the journey.

5. SKAGs Are Lastly, Really Lifeless

If you’re nonetheless utilizing single key phrase advert teams (SKAGs) in 2025, you’re combating a battle that ended years in the past. Bacchini was blunt: SKAGs have “run their course.”

The granular management we used to prize is now a legal responsibility. It fragments your information, making it more durable for the AI to study. Andrew Lolk, Founder at SavvyRevenue, backed this up, warning that over-segmenting campaigns destroys shared learnings. The profitable construction for 2025 is radically easy: Consolidate till the info proves you must separate.

What does that imply? Properly, you need to break up campaigns when there are enterprise causes, like completely different bid targets, completely different promotions, and so on. Put merely, you separate campaigns solely when there’s a strategic purpose, reminiscent of assigning completely different ROAS targets to merchandise with completely different margins, or isolating seasonal stock, like ski jackets, from evergreen classes like swimwear, so every could be optimized by itself efficiency curve.

And whereas single-keyword advert teams are outdated, single-theme advert teams (STAGs) have turn into the trendy, simpler various. As an alternative of isolating every key phrase, STAGs cluster queries that share the identical intent and require the identical message, giving Google sufficient information to study with out sacrificing relevance.

A greater manner to consider it:

A STAG isn’t simply “all trainers phrases,” however it’s “all trainers for distance coaching phrases,” or “all waterproof trail-running footwear phrases.” Every theme represents a selected consumer intent that warrants a selected advert and touchdown web page combo

So, a extra reasonable STAG instance may appear to be:

Theme: Lengthy-distance trainers

• “finest long-distance trainers”
• “marathon coaching trainers”
• “long-distance trainers males”

All completely different key phrases, however they relate to the identical core motivation, the identical advantages to focus on, and the identical touchdown web page expertise.

STAGs protect the messaging management SKAGs as soon as supplied, however with out the info fragmentation that hinders Good Bidding from working at its finest. They offer you messaging precision whereas nonetheless feeding the algorithm sufficient quantity to study.

6. Cease Splitting Efficiency Max

Talking of consolidation, Lolk had some sturdy phrases for a way we handle Performance Max. A standard mistake is splitting PMax campaigns by asset group, model, or generic themes and not using a distinct ROAS goal.

His take? “Splitting = Ravenous.”

PMax campaigns don’t share information nicely. If you happen to break up them, you power every new marketing campaign to study from scratch, requiring double the quantity to stabilize. Except you have got a radically completely different ROAS goal for a selected class, preserve it collectively. And for the love of PPC, cease operating “feed-only” PMax, he says. Simply use Normal Procuring in case you want that management.

7. Search Is Making A Quiet Comeback

In a stunning twist, we repeatedly heard that ecommerce manufacturers have overemphasized PMax and Procuring, leaving cash on the desk in Search.

Lolk argued that Search is reclaiming its function because the high-intent workhorse as a result of it gives what PMax can not: diagnostic visibility and true messaging management. You may’t capitalize on a climate development or a selected seasonal second in case you’re ready for PMax to “study” about it. Search allows you to transfer quick, and it allows you to management the touchdown web page, a lever we’ve severely undervalued recently.

8. Your Aggressive Benefit Is Now “Put up-Click on”

With Google automating bids, focusing on, and even the artistic course of, what’s left for us? Bacchini says the reply lies after the clicking.

Differentiation is the brand new battleground. In case your provide is weak or your touchdown web page is generic, no quantity of bid tweaking will prevent. Shoppers typically dramatically underestimate their rivals and overestimate their very own worth propositions. As PPC execs, our price add is shifting from “technical setup” to “enterprise consultancy,” fixing the provide, the positioning, and the consumer expertise.

9. Generative AI Is Your New Junior Strategist

We moved previous the “AI will write my adverts” hype and received into actual use circumstances.

• Zirnheld defined that AI has turn into her go-to software for smoothing the communication hole between advanced PPC work and consumer understanding. She makes use of it to draft clearer explanations, refine messaging, and spark artistic ideas she will be able to develop additional. AI helps her speed up the early phases, permitting her to spend extra time on higher-value pondering.
• Walker described how AI has turn into a real technical power multiplier inside his workflow. He now makes use of it to put in writing Google Advertisements scripts, construct customized instruments, generate and debug code, and automate duties that beforehand required days of guide effort. AI successfully turns his concepts into working prototypes, permitting him to iterate sooner and push the boundaries of what one PPC supervisor can construct.
• Bacchini shared that AI has reworked how she researches rivals and analyzes positioning. As an alternative of manually combing by search outcomes and touchdown pages, she will be able to feed the whole lot into AI and immediately see patterns, themes, and gaps. It provides her a strategic overview in seconds, serving to her craft sharper messaging and perceive the place shoppers stand in a crowded panorama.

The consensus? AI received’t substitute you, however an skilled utilizing AI will completely substitute an skilled who refuses to the touch it.

In Silicon Valley, we used to lionize the thought of the 10x engineer, the sort of one that may out-code a whole staff, see round corners within the structure, and in some way ship issues at a tempo that felt virtually unfair. However recently, the tales I’m listening to in my very own community inform a distinct story: Lots of these “10x” engineers are beginning to fall behind the so-called mediocre ones who’re merely pushing the boundaries of what they’ll do with AI by their aspect.

And this not applies simply to engineering. In each function, those that study to companion with AI will outperform those that rely solely on expertise and hustle.

10. The “Search” We Knew Is Disappearing

Lastly, we touched on the existential shift. Shep talked about she now makes use of Perplexity.ai for analysis greater than Google. Greg Finn highlighted the instability of AI Overviews.

As I’ve been saying all yr, we’re witnessing a dramatic shift from key phrases to prompts. Search is not nearly matching a question to an advert; it’s about connecting customers who do advanced prompts with options, and possibly displaying an advert if that might be useful.

In an AI-driven ecosystem, the “prompt” turns into the brand new key phrase: a richer, extra contextual sign that displays not simply what customers sort, however what they’re making an attempt to perform. Advertisers who nonetheless assume by way of remoted key phrases will fall behind; those that assume in prompts, duties, and intent paths will thrive.

The Backside Line

The work of the trendy PPC marketer continues to shift from pulling levers to pondering critically concerning the levers being pulled on our behalf. Automation is not optionally available, however neither is oversight. The winners this yr have been the advertisers who understood the place algorithms shine, the place they stumble, and the place a human must step in with context that the machine merely doesn’t have.

And this evolution is much from over. As we head into 2026, I anticipate the debates on PPC City Corridor to get much more fascinating. We’ll seemingly spend much less time arguing about whether or not to undertake AI, and extra time unpacking how you can direct it, how you can measure it, and how you can stop it from homogenizing each account it touches. We’ll discover what occurs when prompts really turn into the brand new key phrases. And we’ll hear from practitioners who discover artistic, typically stunning methods to bend automation again towards profitability and technique, fairly than comfort.

If 2025 was the yr we discovered to inform the machine “No,” then 2026 stands out as the yr we learn to inform it “Do that and right here’s why.” The entrepreneurs who thrive might be those that don’t simply handle campaigns, however handle techniques, utilizing judgment, experimentation, and clear intent to information more and more highly effective instruments.

I’m wanting ahead to a different yr of unfiltered conversations on PPC City Corridor and to seeing what new onerous truths (and alternatives) we uncover collectively.

Extra Sources:

Featured Picture: Anton Vierietin/Shutterstock