Google Email Leak Explained: What’s coming next to Google Ads?

Episode Transcript

[0:00:10] Mike: Hello and welcome to another episode of Growing Ecommerce. I’m your host, Mike, of Smarter Ecommerce, also known as smec. Unfortunately, Chris is on the road today, so he’s not able to join us. We’ve got a chilly morning here. There’s a chill in the air, and if I look out my home office window, there’s a wall of white fog. The poet John Keats once called this the “season of mists and mellow fruitfulness,” but when I look out the window, I’m reminded of the opaque advertising industry. You never know what kind of day it’s going to be. Will the sun heat up and burn off the fog, giving us one of those brilliant, crisp, clear days? Or will it continue in this direction and start to get rainy and dark?

I’ll drop the absurd metaphor and move on to talk about Google’s PowerPack in some detail today. I’m going to discuss a previously confidential email that you’ve likely never read before, but I think it’s essential for understanding the current strategy of Google Ads and the direction Google is headed as a platform. We can also compare how that might diverge from a competitor like Meta. I’ll introduce the cast of characters, explain what the PowerPack is and which campaign types are involved, and give you Google’s view on how it all works together—as well as a contending view on how things might not be as harmonious as they seem. Then we can talk about those conflicts and some possible resolutions. I’ll try to keep it moving along briskly.

P-Maxification: A Quick Refresher
Just over a year ago, I was on stage talking about something I called “P-Maxification”. For new listeners and as a refresher, my P-Maxification thesis was that Google Ads was being rebuilt board by board, like the ship of Theseus, in an increasingly automated fashion until the whole platform would eventually have nothing in common with the original pieces. The key tool Google used in this end-to-end automation strategy was Performance Max.

At the time, we saw different campaign types migrating into Performance Max. Within Google Ads as a platform, it was absorbing all other ad-serving mechanisms. PMax is a cross-network campaign type that can serve anywhere across Google’s surfaces. There was also a higher-level dimension to this where every other platform was racing to build their own PMax. So P-Maxification was occurring not only within Google Ads but also in the industry at large.

I still think there’s a lot of validity to this thesis, but I’m going to share an email thread I discovered between two senior executives in Google’s sales organization. While Google is labeled “big tech” and builds itself as a technology company, I would argue it is ultimately an advertising, sales, and marketing organization.

The Leaked Email: “That Was Our Previous Strategy”
These two senior sales executives had a very interesting exchange. The first email landed at 8:36 AM on May 22nd, 2024—exactly one day after Google Marketing Live. Executive A expressed concern that Google “doubled down unambiguously that all our AI goodness is PMax”. He felt the general consensus among the sales team was that they didn’t believe Demand Gen was going to be a “thing” and that anything that is not PMax is structurally disadvantaged.

This reflects my P-Maxification thesis. Even within Google’s own sales team, the feeling was: if you’re not PMax, you’re nothing. Why do you need Demand Gen when PMax serves everything else?

The reply came from Executive B at the end of the following business day. He agreed they were pushing PMax hard but explicitly stated, “That was our previous strategy”. This was a lightbulb moment for me. While I was talking about P-Maxification in September 2024, internally at Google, they were already calling it their “old” strategy by May.

The question then becomes: if that was the old strategy, what is the new one? Executive A had mentioned Demand Gen, and Executive B talked about something then called “Search Max” or “PMax for Search”. This technology went into beta and is now entering general availability under the name AI Max for Search.

AI Max for Search is essentially the continuation of Google migrating campaign types toward PMax. They focus-grouped the idea of migrating keyword search campaigns to PMax, but the reception was presumably bad, so they brought PMax features over to search instead. If it doesn’t have “AI” in it, shareholders aren’t interested. With ads coming to AI Overviews and AI Mode, they landed on AI Max for Search. Going all-in on PMax is the old strategy; the new approach is broader, and that’s the PowerPack.

The Khrushchev Thaw of PMax
I call this the “Khrushchev thaw” of PMax. There has been a “Cold War” relationship between advertisers and Google Ads—an unhealthy relationship for both parties. Then things started to change; Google took a different tone and started opening up. These emails are like leaked diplomatic cables or encrypted comms from the Kremlin proving this isn’t imagined.

We’ve seen structural changes. Standard shopping campaigns were clearly disadvantaged in the past as PMax had default priority, but that restriction was lifted a year ago. More recently, Google sales decks even encourage hybrid approaches, running shopping and PMax campaigns in parallel. Google likely realized they were getting stuck at a certain point with PMax alone. Executive B also mentioned that nobody was particularly excited about PMax in advertiser conversations; at best, they were willing to go along with it despite real frustration.

The Cast of Characters
Last year, Google talked about the “PowerPair”—the combination of Performance Max and standard search campaigns with broad match and smart bidding. This year, they rolled out the “PowerPack,” a holy trinity of Performance Max, Demand Gen, and AI Max for Search.

Performance Max — Finding Conversions
PMax’s role is to find conversions across all Google channels, including Display, Gmail, Discover, YouTube, Shopping, and Search. Originally pitched as a full-funnel campaign type, Google now frames it differently. They talk about “marginal return on ad spend,” meaning Google finds the next cheapest conversion it can. This is similar to how Meta advertising works. Google has been clear that PMax is bottom-funnel only, operating in the purchasing phase.

PMax has a relentless performance focus, but there is also a lot of low-quality “spam” inventory mixed in. Placement reports often show YouTube, Display, and search partner placements that advertisers aren’t thrilled about. My biggest criticism of PMax remains the traffic quality; it works on average, but it could work better if we could trim the fat.

Demand Gen — Discovery and Consideration
Demand Gen is a mid-funnel campaign type and the successor to Discover campaigns. Discover ads could never really scale, and Google was eager to sell more YouTube ads and monetize YouTube Shorts. Modern Demand Gen campaigns are roughly two-thirds YouTube, one-quarter Discover, and a small amount of other cross-network placements.

Its role is discovery, consideration, and feeding the funnel. I like that it’s highly customizable; you can do channel targeting for things like YouTube Shorts-only or Gmail-only campaigns. However, I feel its reach quality is decreasing over time as the Display Network is being included more often.

AI Max for Search — Expansion
AI Max is the newest and least mature technology. While PMax’s goal is finding conversions and Demand Gen’s is nurturing, AI Max is expansionary. Its job is to take your search campaigns and find traffic or incremental conversions you might be missing.

It offers high transparency and customization, but it can be aggressively expansionary. AI is not strategic; it doesn’t know what it doesn’t know. In a mature search build, AI Max might look in strange places because you are already covering the likely ones. I’ve heard examples of AI Max bidding on competitor terms where you might have had a strategic reason not to be there. It also tends to scale heavily into the search partner network or brand traffic. You need to put guardrails in place.

Tying it All Together — The Funnel View
Imagine a classic marketing funnel:

• Top of the funnel: Google’s display and video inventory. YouTube is increasingly important as Google believes the open web is dying.
• Mid-funnel: Demand Gen.
• Bottom of the funnel: PMax.
• Expansion layer: The radiating circle at the bottom is AI Max for Search, reaching beyond existing boundaries.

This presents a coherent view, but the reality is much more conflicted. There are huge amounts of redundancy and overlap in ad inventory and technological features. I am skeptical that there isn’t cannibalization or self-competition occurring.

Feature Duplication
An example of triplication is “Dynamic Search” technology. Dynamic Search Ads (DSA) have been around for years, automatically generating headlines and selecting landing pages. Inside PMax, this is called “Final URL Expansion.” Inside AI Max, it is called “Text URL Optimization”. We have the same technology three times under different names. Dynamic Search Ads will likely go away because they aren’t part of the “protected” PowerPack. Running all three creates a risk of them competing with each other rather than finding incremental gains.

Inventory Overlap: PMax vs. Demand Gen
Both PMax and Demand Gen serve YouTube, Display, Discover, Gmail, and Maps. They are supposed to be separated by their funnel position (bottom vs. mid), but you can make Demand Gen conversion-oriented and add a product feed to it. If you put a tROAS on Demand Gen, you are dragging it into the lower funnel where it overlaps significantly with PMax.

My philosophy is to let Demand Gen be Demand Gen—keep it mid-funnel as a feeder for everything else. Instead of giving them the same goal, find ways for them to support each other. For example, use Demand Gen for “sleepy” long-tail products that aren’t getting traction in PMax. Use Demand Gen’s channel targeting to fill gaps identified in PMax channel reports.

Wrapping Up
P-Maxification remains a real trend, but Google has moved on to a broader strategy of end-to-end automation with more tools. Unlike Meta, which aims for a “black box” that runs itself without intermediaries, Google is adding back some control and transparency.

The clean funnel view Google presents isn’t the whole story. Everything is trying to expand as much as possible, and the interactions are not as simple as advertised. It is harder than ever to maintain a hygienic, well-organized account.

Think critically about what Google tells you. Test these technologies, but don’t feel rushed to be an early adopter. AI Max currently feels half-baked to me. Take things one step at a time and carefully consider the roles and interactions of every campaign in your account. If you want to learn more, reach out to us at smarter-ecommerce.com.

Thanks for listening, and we’ll see you next time.