Google Ads & Meta Ads, understanding the online advertising landscape

More than 70% of businesses don't achieve their marketing objectives. Our mission is to help professionals find leaks, to inform them about the AI landscape, and build profitable strategies for the long run. Get a free consultation with an expert.

Request Your Free Audit See the audit checklist →
Last reviewed: July 2026 Covers: Google Ads & Meta Ads (Facebook, Instagram, Threads) Written for: business owners, not marketers

The short version

1Meta's ad business is now growing roughly twice as fast as Google's (about 24% vs. 12% projected for 2026). Analysts expect Meta to overtake Google in total digital ad revenue for the first time this year. Google is still bigger and still growing; the gap is just closing.
2Google is changing how Smart Bidding works starting August 17, 2026: budget-limited Target CPA and Target ROAS campaigns get tuned to hit your target precisely instead of sometimes beating it for free. If your costs have looked unusually efficient lately, budget for them to drift up.
3Both platforms now run most of your account on automated bidding by default. That's powerful, and dangerous the moment nobody is watching what it's optimizing for.
4The single biggest source of wasted spend we find in audits is rarely a "bad campaign." It's broken, incomplete, or duplicated conversion tracking feeding bad data to the algorithm.
5A great ROAS or a high "optimization score" doesn't automatically mean a healthy account. Sometimes it just means you're being charged for customers you already had.
6Meta's new AI ad-creation tools are genuinely powerful, but they've already made public, documented mistakes: miscategorized products, mistargeted audiences, broken images. Treat AI-generated creative as a first draft that needs a human check, not a finished ad.

Perspective

First, we observed that the core problem is that the platforms occupy every seat at the table. Google and Meta run the auction, advise you on how to bid in it, execute the bidding for you, and then measure their own results. No other market works like this: imagine a stock exchange that also managed your portfolio and wrote its own performance reports. Big advertisers can afford independent testing to check the machine's claims; a small business just has to trust the dashboard.

The "AI democratizes advertising" pitch inverts reality. When execution skill is automated away and everyone runs the same algorithm, the small player loses the one advantage they could actually earn: being smarter than the competition. What's left to compete on is budget, and auction prices drift upward while the platform pockets the efficiency gains. It's no accident that the loudest promises of simplification come from companies whose revenue depends almost entirely on ad spend.

Meanwhile, transparency only ever moves in one direction. Year after year, advertisers see less of their own data and control less of where their money goes, always in the name of convenience. And on the organic side, the old bargain — you make content, search sends you visitors — is being dismantled, as AI answers keep users on the results page and the sites that fed those answers get scraps.

Regulators have started agreeing with the diagnosis, but remedies arrive years late and softened, while the platforms rebuild the landscape faster than courts can rule on the old one.

The takeaway for small businesses and the people who serve them: the leverage that remains lies in what platforms can't or won't provide — independent measurement, strategy, distinctive creative, and owned audiences that don't depend on renting visibility.

Market share & dynamics

Both platforms are changing faster than most advertisers realize. Here's what's actually shifted, and why it changes how you should read your own results.

$60.4B
Google Search ad revenue, Q1 2026 (up 19% year-over-year)
Source: Alphabet SEC filing, Apr 2026
$55.0B
Meta ad revenue, Q1 2026 (up 33% year-over-year)
Source: Meta SEC 10-Q, Apr 2026
68%
Of US Google searches ended with no click at all, Jan–Apr 2026
Source: SparkToro / Search Engine Land, Jun 2026
~2×
Meta's projected 2026 ad-revenue growth rate vs. Google's (≈24% vs ≈12%), enough that analysts expect Meta to overtake Google in total ad revenue for the first time this year
Source: eMarketer, via Marketing Dive, Apr 2026
Google Ads

Search is still dominant, but the click is changing shape

  • Still growing, but Amazon is growing faster. Google's own Search ad revenue rose about 19% year-over-year in Q1 2026. Even so, eMarketer now names Amazon, not Google, as the single biggest driver of new growth in US search advertising.
  • Most searches never produce a click. 68% of US Google searches ended with zero clicks in early 2026, as AI-generated summaries answer more questions directly on the results page. That doesn't mean search ads stopped working. It means the searches that do click increasingly matter more.
  • A real antitrust loss for a different product than the one you use. In April 2025, a federal court ruled Google illegally monopolized parts of the publisher-side ad-tech market (the tools websites use to sell ad space). That's a genuine loss for Google, but a separate business from Google Ads search campaigns. Remedies and an appeal are still playing out.
Meta Ads

Feed and Reels dominance, and new surfaces opening up

  • Accelerating, not slowing down. Full-year 2025 ad revenue hit $196B (+22%), and Q1 2026 alone brought in $55B (+33% year-over-year), growth analysts expect to outpace Google's this year.
  • A new required choice for EU audiences. After the EU ruled Meta's "pay-or-consent" model didn't give users a real choice, Meta rolled out a new EU ad-personalization option to every EEA user starting January 2026: share full data for personalized ads, or share less data for a more limited experience. If you advertise to EU audiences, your targeting precision there now depends partly on which option people pick.
  • Delivery runs through an AI shortlist first. Meta's "Andromeda" system filters tens of millions of possible ads down to a shortlist of about a thousand before its ranking models make the final call. A single underperforming ad rarely gets a fair chance to recover on its own. The system moves on fast.

Bidding strategies

Both platforms want to bid for you automatically. That's usually the right default, but "automatic" is not the same as "unsupervised."

What you wantGoogle AdsMeta Ads
Spend efficiently toward a cost-per-result targetTarget CPACost Cap
Spend efficiently toward a return-on-ad-spend targetTarget ROASROAS Goal
Get the most results possible from a fixed budgetMaximize ConversionsHighest Volume
A hard ceiling on what you'll ever bid per auctionManual CPC (outside Smart Bidding)Bid Cap

Google is recalibrating how precisely Target CPA/Target ROAS track that number starting August 17, 2026 (see below).

Google Ads

Smart Bidding, governed

Changing August 17, 2026

Google is recalibrating budget-constrained Target CPA and Target ROAS campaigns to track your set target more precisely, instead of sometimes beating it for free, as they can today. Google's own example: a campaign targeting a $10 cost-per-result that's currently averaging $5 will drift back up toward the full $10. If your Smart Bidding results have looked unusually efficient lately, budget for that efficiency to normalize.

  • Target CPA / Target ROAS: set a cost-per-result or return target and Smart Bidding chases it across every auction.
  • Maximize Conversions / Conversion Value: spends the full budget for as many (or as valuable) results as possible, with no explicit target.
  • All of them need enough conversion volume to learn from; thin data is the most common reason automated bidding underperforms for smaller accounts.
Meta Ads

Advantage+ and automated delivery

  • Highest Volume: spends the full budget for as many results as possible, no target set.
  • Cost Cap / ROAS Goal: chases a cost-per-result or return target, similar in spirit to Google's Target CPA/ROAS.
  • Bid Cap: a hard ceiling on what Meta will ever bid in an individual auction, for the most manual control.
  • Every option feeds Meta's Andromeda retrieval system, which shortlists your ad against tens of millions of others before ranking even begins, so the same tracking-quality and creative-relevance factors that affect Google's algorithm affect this one too.

The bidding mistakes that quietly cost the most

1

Mixing brand and acquisition spend

Bidding on your own brand name buys traffic that was already coming to you: the customer typed your name, knew you, and was about to click anyway. Worse, blending brand and acquisition in one report inflates your average ROAS and makes weak acquisition campaigns look healthy.

The fix: split brand and acquisition into separate campaigns and read two ROAS numbers, never one blended average. If you already rank first organically for your own name, test pausing brand bidding; it rarely costs revenue, and it recovers margin immediately.
2

Letting automated campaigns run without exclusions

Performance Max and Advantage+ spend your budget wherever the algorithm finds the easiest conversions, and the easiest conversion is usually a customer who was going to buy anyway (past purchasers, cart abandoners, people already searching your brand). Without guardrails, these campaigns can take credit for sales you didn't need to pay for.

The fix: exclude branded search terms and already-warm audiences where the platform allows it, and track what share of "new" revenue is really coming from people who already knew you.
3

Optimizing to ROAS instead of real margin

ROAS measures revenue per dollar spent. It says nothing about what you keep. A campaign selling a 60%-margin product and one selling a 15%-margin product can show the identical ROAS of 5, while one is profitable and the other loses money on every sale once landed cost and logistics are factored in.

The fix: bid to a cost-of-sale (spend as a % of revenue) target set per product category, based on actual margin, not a single blended ROAS goal across everything you sell.
4

Assuming the platform allocates budget better than you would

Left alone, automated campaigns concentrate spend: sometimes 30–50% of a Shopping or catalog budget lands on one or two best-selling items, starving everything else of the exposure it would need to prove itself. The algorithm is optimizing against its own signals, not your margin mix.

The fix: segment campaigns by margin tier, force dedicated budget toward higher-margin products the algorithm is neglecting, and check the real spend distribution regularly instead of trusting the "optimization score."

Conversion tracking

This is the root issue. Every bidding strategy, every automated campaign, every report above is only as trustworthy as the tracking underneath it.

Why this is the highest-leverage fix in either platform

Smart Bidding and Advantage+ both optimize on the conversion data you feed them. If that data double-counts orders, misses events, drops basket value, or attributes sales to the wrong click, the algorithm doesn't fail loudly. It optimizes confidently in the wrong direction. You end up cutting a genuinely profitable campaign because it "looks weak," and pouring budget into one that's quietly double-counting its own sales.

Google Ads

What to check

  • Enhanced Conversions: supplements standard tracking with hashed first-party data (like email or phone) so conversions survive cookie loss and ad blockers. Google is actively pushing adoption: since mid-June 2026, it's begun rejecting old-style offline conversion uploads that never adopted the newer method.
  • Consent Mode: adjusts what Google can measure based on a visitor's actual cookie consent choice. Required for compliant measurement in the EU/EEA and UK; if it's missing or misconfigured, expect gaps in reported conversions that have nothing to do with real performance.
  • GA4: the standard source for importing website conversions into Google Ads; worth confirming your key events are actually flowing through, not just installed.
Meta Ads

What to check

  • Conversions API (CAPI): sends conversion events from your server, not just the visitor's browser, recovering data that iOS privacy settings and ad blockers keep from the Pixel alone.
  • Automatic deduplication: Meta's Conversions API Gateway now auto-generates a shared event ID between Pixel and CAPI, so the same purchase shouldn't get counted twice. Worth confirming in your own account rather than assuming; duplicate conversions are one of the most common ways automated bidding gets fed a bad signal.
  • AI Pixel "enrichment": a newer feature that auto-infers page and product details to improve data quality. It defaults on with advance notice, not silently; worth knowing it's there and deciding deliberately whether to keep it on.

Advanced tools & automation

Both platforms have pushed hard into AI-driven, asset-based campaign types. They're genuinely powerful; the mistake is treating them as "set and forget."

Google Ads

Performance Max, Demand Gen & what's new

  • Performance Max: runs across Search, Display, YouTube, Gmail and Maps from one campaign. Channel-by-channel performance reporting, once a limited beta, is now standard; use it to see where your budget actually goes instead of trusting one blended result.
  • Demand Gen is absorbing Display. Google announced in May 2026 that Display campaign management is folding into Demand Gen, with a full transition targeted for 2027. You'll still be able to run Display-only delivery if you want it, but plan for the interface to change under you.
  • AI Mode ads are still a US-only test as of Google's most recent recap; worth watching, not yet worth building a strategy around.
Meta Ads

Advantage+ & generative creative

  • Advantage+ Shopping & Creative: Meta's automated campaign and creative tools, including generative AI for backgrounds, image expansion and ad copy variations.
  • The AI creative tools need a human check. In July 2026, eMarketer documented real, named cases of Meta's generative ad tools getting it wrong: a product miscategorized into the wrong listing, an ad for a women-only group generated centering men, a product image rendered with two sets of handlebars. Treat AI-generated creative as a fast first draft, not a finished ad, and check whether accounts were silently re-enrolled into AI features after opting out.

Curious what your own account would show?

We'll read it the same way: from the P&L up. You'll see exactly where budget is working, where it's leaking, and what to fix first.

Request Your Free Audit

Further reading