What I learned managing $20M+ in Google Ads at Google

The TL;DR

• I advised 55 clients per quarter across $20M+ in annual ad spend, primarily ecommerce plus B2B SaaS, financial services and local services
• The vast majority of accounts I worked with were profitable on ad spend; most agencies cannot say that
• The most useful diagnostic for whether your agency is doing real work: ask for the Google Ads change log from the last 90 days
• I still work with Googlers from the outside, which means clients get access most agencies cannot match
• I report in profit-based language (MER, contribution margin, nCAC:LTV) — the language CMOs/VPs actually use to make budget decisions

---

Most Google Ads advice online is recycled by people who learned from blogs.

That’s a strong claim to open with but it’s true. Many of the operators writing about paid search today learned from the operator class above them, who learned from blogs, who learned from courses, who learned from earlier blogs. The information tends to loop without much new input. Misconceptions get cemented because everyone repeats the same things.

The view from inside Google is different. You see patterns across hundreds of accounts that no single agency operator gets to see. You watch the same mistakes appear across verticals that have nothing in common. You see what actually correlates with profitability versus what just looks good in monthly reports.

This is the post I would have wanted someone to send me if I were evaluating a Google Ads partner. It’s what I saw from inside the platform, what stuck and how I run my own engagements now.

The scale of what I saw

I advised 55 clients per quarter across roughly $20M+ in annual ad spend.

My portfolio was predominantly ecommerce — brands ranging from emerging DTC to mid-market multi-channel retailers. Beyond that, the mix included B2B SaaS, financial services and local service businesses (HVAC installation and repair, window and door companies, similar verticals).

That cross-section matters. Most agency operators specialize in one or two verticals and develop pattern recognition that doesn’t transfer well outside them. When you see the same problems show up in HVAC accounts that show up in DTC apparel accounts that show up in B2B SaaS funnels, you start to recognize what’s actually universal about Google Ads performance versus what’s situational.

The lessons in this post come from that cross-section.

The outcome that matters most

The vast majority of clients I engaged were profitable on ad spend. Most Googlers and agencies cannot say that with a straight face.

That sounds like a sales line so let me define what I mean.

Profitable on ad spend means the campaigns generated revenue that exceeded media costs plus the agency or in-house cost to run them, with enough margin to justify the working capital tied up in waiting for cash to come back from the platform. It does not mean ROAS over 1x. ROAS over 1x means you broke even on top-line revenue, which for most businesses is a meaningful loss after COGS, fulfillment and overhead.

The reason most accounts are not profitable on this definition is structural. They’re run for activity (changes shipped, optimizations completed, reports delivered) rather than outcome (more contribution margin in the business). Activity is easy to measure and easy to bill against. Outcome requires aligning what the agency does with how the business actually makes money. Most engagements never get there.

The accounts I worked with that were profitable shared three patterns.

First, the brand owner could articulate their unit economics. They knew their gross margin, their contribution margin after variable costs and what they needed CAC to be to make growth viable. If you cannot answer those questions about your own business, no agency can run Google Ads profitably for you.

Second, the conversion data flowing into Google was accurate and reflected actual business outcomes. Not just “form filled” or “added to cart.” Real revenue events, real qualified leads, in real time.

Third, the people running the account understood the account was a system. Not a list of campaigns to optimize. Not a set of keywords to bid on. A system where structure, match types, bidding strategy, audience signals and creative all interact. Changes get made with awareness of how the whole thing fits together. Think about the roles of dentist and primary care physician — they work with different parts of the same system (your body) for one shared goal (long-term health).

When those three pieces are in place, profitable Google Ads is not hard. When they’re not, no amount of optimization fixes the underlying problem.

Most agencies collect the retainer and never touch the account

This is the dirty secret of the agency model.

A meaningful portion of Google Ads engagements consist of a monthly report, a status call and almost no actual changes to the account. The work that justifies the retainer is not happening. Or it’s happening at a fraction of what the retainer is paying for.

I cannot emphasize how many conversations I had with marketing coordinators or directors at agencies where I asked “how is the product feed optimization coming along?” and was met with “that’s still on the list for next month.”

To be clear, this is not bad-faith fraud. It’s a byproduct of agencies optimizing for client retention rather than client outcomes. Active changes carry risk. A change that doesn’t work generates a tough conversation. A change that does work eventually surfaces a ceiling on how much more can be improved. Both make agencies vulnerable to churn.

Quiet account management — small tweaks reported as bigger work, monthly reports that highlight any positive movement regardless of cause, status calls that emphasize “what we’re seeing” without pointing at what to change — keeps the relationship friendly and predictable. It just doesn’t move your business forward.

There is a legitimate version of low-touch account management. Accounts that are mature, well-structured, hitting their targets and generating predictable conversion volume can run with minimal intervention. That state is called maintenance mode and it exists. The problem is most agencies treat early- and mid-stage accounts as if they’re already there. They’re not. Accounts that haven’t yet hit their efficiency floor still have meaningful gains available if someone is willing to stake their reputation and make changes.

Here’s the math. An account spending $30K a month, running at an effective CPA of $80, has roughly 375 conversions a month flowing through bidding strategy. That account is probably 20-30% inefficient on CPC if Quality Score and bidding isn’t actively managed. It’s probably leaking another 15-20% of clicks to broad-match queries that should be negative keyworded. It’s probably missing 10-15% of incremental conversions due to learning-phase resets from frequent ad rotation. Compound those and you’re looking at $5K-$8K of monthly waste, every month, that real account work could close over time.

The diagnostic is simple. Ask your agency to send you the change log from your Google Ads account for the last 90 days. Google Ads tracks every change made by every user with timestamps. Real account work leaves a trail. If the change log shows fewer than 5 to 10 substantive changes per month on an account spending more than $10K, the work isn’t happening at the level the retainer suggests.

This is one of the highest-leverage diagnostics a business owner can run on their existing agency relationship.

Re-engaging “failed” accounts almost always surfaces the same problems

I re-engaged dozens of clients who had concluded Google Ads was not adding contribution margin to their business. In nearly every case the conclusion was wrong. The measurement was wrong.

The pattern was consistent enough that it became a checklist.

Conversion data quality. The account was tracking the wrong events, the right events with the wrong values or events firing twice leading to inflated total counts. Smart bidding optimizes against the data it gets. When the data is broken, the optimization is broken.

Misattributed conversions. Last-click attribution was being used to evaluate Google Ads against paid social and organic. Google Ads frequently sits in the middle of a customer journey that ends with a branded search or direct visit. Last-click strips Google Ads of credit for the visits that started the journey. Data-driven attribution exposed the real picture in most of these accounts. Third-party marketing mix modeling (MMM) software was even more accurate.

Wrong match type strategy. Accounts running exact-match-only on competitive ecom queries were either spending nothing or losing every auction. The Google Ads platform changed underneath them. The agency or in-house team did not update.

Blended ROAS hiding incremental loss. This is the most insidious one. Brand search runs at 5-10x+ ROAS in most accounts. Non-brand runs at 1-3x. Blended ROAS averages those together and looks healthy. The non-brand campaigns might be losing money outright while the brand campaigns are mostly just intercepting visits the business would have gotten anyway. Until you separate the two, you can’t tell whether your non-brand investment is producing real growth or just absorbing margin.

Bidding strategy mismatch. Manual CPC on an account that should be on Target CPA. Target ROAS on an account with conversion volume too low to give the algorithm signal. Maximize Conversions on a business with unit economics constraints the bidding strategy can’t see.

After fixing the measurement and the structural issues, the same accounts that had been “unprofitable” turned profitable. Often quickly. The work wasn’t magic. It was diagnostic precision applied with discipline.

If you have an account you’ve written off, there’s roughly a 70% chance you’re looking at one of these five problems. None of them require a new agency to identify. They do require someone who knows what to look for.

You should work with Google reps

Hear me out for a second. I have no skin in the game now.

At Google I had access to forecasting tools, budget planning resources, competitive intelligence dashboards and the teams who built new products (Performance Max, Demand Gen, AI Max).

Most companies running Google Ads above a certain spend threshold have access to most of this too. The reps assigned to accounts have similar tools available to them. The competitive intelligence Google sees on your account is the same Google sees on every account in your category. Most companies underutilize all of it because they don’t know what to ask.

I do know what to ask. And I have continued working relationships with Googlers from outside the company. Those relationships matter.

A few specific examples of how this plays out for clients.

Forecasting data shifts a budget conversation. Most CFOs ask “how much should we spend on Google Ads next quarter?” The most complete answer involves running a forecast against current account performance, projecting where saturation hits and identifying where additional spend stops producing proportional return. Google has tools internally that produce this projection (lost conversions due to bid or budget constraints). Most operators on the outside never request it. I do.

Competitive intel reframes a category strategy. When a client thinks they’re competing on a small group of keywords, the auction insights and category data inside Google often show the actual competitive set is much larger or different than they assumed. Understanding the real auction landscape changes which campaigns matter and where defensive spend is justified.

Rep relationships unlock beta features. Google releases features to a subset of advertisers before public rollout. Performance Max channel control, YouTube pause ads (like on Netflix), native AI video/photo generation — all of these showed up earlier for accounts I was supporting. A responsive Googler means the rep believes the account is worth pushing forward. That belief comes from the operator running the account, not necessarily from the spend level.

This kind of access is the difference between an account that hits diminishing returns at a certain scale and an account that keeps finding new ground to grow into.

Most agencies cannot do this. They burned their rep relationships through low-quality conversations or never built them in the first place. The relationships work in both directions — the operators reps trust get the time and the access.

Report in business language, not marketing language

Most marketing org leaders report on ROAS, CTR, CPC and impressions.

VPs, Directors and CMOs make decisions on Marketing Efficiency Ratio (MER), nCAC:LTV ratios, AOI per marketing dollar, contribution margin and payback period.

The gap between these vocabularies is where most bad budget decisions get made.

A marketing director can show their CMO a quarterly report full of ROAS improvement and look like they’re winning. The CMO looks at the same data and sees that overall marketing spend is up 40% while contribution margin is flat. The director is reporting on what’s easy to measure inside the marketing function. The CMO is making decisions on what determines whether the business is growing profitably.

I tie outcomes to profit-based measurement while optimizing based on channelization.

Here’s what that looks like in practice.

A DTC ecommerce brand running roughly $80K a month in Google Ads at a 3.2x ROAS. Their marketing director was reporting this as a win. Their CFO was raising flags because total marketing spend across all channels had grown 60% year-over-year while gross margin compression had narrowed contribution margin to almost nothing.

Reframing the conversation from ROAS to MER changed what the data revealed. Their MER was sitting at 2.1x against a break-even MER of 2.6x given their gross margin. The business was technically unprofitable on a fully-loaded marketing basis even though each individual channel looked acceptable on ROAS.

The fix wasn’t about cutting Google Ads. It was about identifying which channels were actually incremental and which were absorbing demand that would have closed anyway. Google Ads non-brand turned out to be doing real work. Google Ads brand was largely catching organic traffic. Meta prospecting was producing the appearance of conversions through last-click attribution while contributing minimal incremental revenue.

After restructuring around incremental performance, the brand hit MER 3.4x within two quarters. Same revenue. Lower total marketing spend. Healthier contribution margin.

That’s what reporting in business language looks like. It’s not a different toolset. It’s a different orientation toward what the data is for.

Build frameworks, test them & refine regularly

Every client engagement gets a custom strategy and toolset. The starting point is a set of proprietary frameworks and tools I built and tested across the hundreds of accounts I worked with at Google.

These aren’t templates. They’re pattern-recognition shortcuts that came from seeing the same problems and solutions appear repeatedly across verticals.

For ecommerce:

Landing pages that convert. The structural elements matter more than the design. Hero positioning, social proof placement, friction reduction at the conversion step and trust signals at the right moment in the page. Most ecommerce sites I worked with had landing pages that worked against their own paid traffic. The pages were built for browsing rather than for converting paid visitors with high intent.

Checkout upsell flows. Most brands miss this entirely. The post-add-to-cart, pre-checkout moment is where AOV gets built. Subscribe-and-save offers, bundle upgrades, accessory recommendations — placed correctly, these add 15-25% to AOV without affecting conversion rate. Placed wrong, they tank conversion.

Post-purchase reengagement. The 60-90 day window after the first purchase is where customer lifetime value gets built or lost. Email and SMS sequences designed around real consumption cycles and category behavior, not generic “we miss you” flows.

For lead generation:

High-converting lead capture pages. Form length is not the variable. The variable is matching the offer to where the visitor is in their decision process. Top-of-funnel offers on bottom-of-funnel ads waste budget. Bottom-of-funnel offers on top-of-funnel ads waste leads.

MQL definitions that predict revenue. Most lead gen accounts track leads as if all leads are equal. They aren’t. The MQL definition matters more than the lead volume number. I work with sales teams to define what a real MQL looks like for their business and feed that signal back to Google for bidding optimization.

Angle and offer testing driven by sales call recordings. Marketing teams test angles based on what they think will work. Sales teams hear what customers actually say on calls. Using sales call recordings to surface the language and objections that come up in real conversations produces ad copy that converts at higher rates because it speaks to actual concerns rather than imagined ones.

Across all of these, the thread is the same. Frameworks should start from how the buyer actually behaves. Not from what’s convenient to build inside an ad platform.

What this means if we work together

The frameworks aren’t theoretical. They run through every Google Ads engagement I take.

Worth saying directly: ultimately some of the parts of the role that required pushing products prematurely led me to leave Google. I now work for the advertiser, not for the platform. That orientation changes which recommendations get made and which products get pushed. It’s the only sustainable position when you’re advising on someone else’s money.

When we work together, the lessons in this post become operational reality. The diagnostic posture toward existing accounts. The reporting in business language. The Google rep relationships managed on your behalf. The frameworks adapted to your specific business.

That’s the offer.

---

Real expertise comes from real volume across real verticals. The frameworks here were earned by watching hundreds of accounts succeed and fail under conditions I had a front-row seat to.

If you want me to evaluate your Google Ads account through this lens, I’m happy to do it. No pitch. No follow-up sequence. Just an honest read on what I see and what I’d change.

I'm Aayaam — a former Google strategist who loves to share ideas and learn in public.