Blog// paid search

Twenty Years of Google Ads: What Changed, What’s Gone, and Why Automation Doesn’t Remove Business Responsibility

Matt Van Wagner reviews 20 years of Google Ads, showing how automation reshapes the game while businesses remain responsible for strategy and budget.

14-10-2026 · 8 min read
Twenty Years of Google Ads: What Changed, What’s Gone, and Why Automation Doesn’t Remove Business Responsibility

Matt Van Wagner has worked in paid search since the early 2000s. Over that time the industry has moved from enthusiasts mastering a new sales channel to an environment where algorithms and large platforms make the decisions. According to Search Engine Land, his retrospective is not nostalgia but a set of guideposts for anyone spending budget on contextual advertising today. For a small-business owner in Russia and the CIS, it is useful to understand which familiar schemes have stopped working, where Google’s interests diverge from the advertiser’s, and which decisions remain in his zone of responsibility.

From ‘Gold Rush’ to Consolidation: How the Market Changed

In the early 2000s traditional channels — trade publications, exhibitions, direct mail — were losing effectiveness. Search offered a fundamentally new way to find a buyer: at the moment of intent. According to Search Engine Land, the early conferences had a collaborative atmosphere: the market was growing fast enough for everyone to get a share. Agencies often underpriced their services, failing to recognize the value they delivered to clients.

The emergence of Search Engine Land and the SMX conferences changed the communication format. SMX Advanced skipped basic topics and went straight to advanced examples. This attracted speakers with deep experience, and informal coffee chats became as valuable as the official sessions. Over time the market matured, the same names appeared on agendas, and breaking in became harder for newcomers. Van Wagner, who participated in speaker selection, introduced anonymous rating of theses — without names, companies, or gender — so ideas would win on merit rather than reputation.

For business this means: the era of ‘just launch ads and sales will come’ is over. Competition for attention inside the auction has grown, and access to quality expertise now requires a targeted search rather than attending any event.

Gone Tactics: Why Complex Structures Were Built Before and Why Automation Removed Them

Earlier, professionalism was judged by single-keyword phrases in ad groups, duplication of match types, and layered negative-word structures. The goal was to rigidly allocate traffic among campaigns and control every bid. According to Search Engine Land, automation has made these constructions largely unnecessary: algorithms themselves find connections between queries and ads inside broad groups.

Van Wagner notes the loss of direct access to Google engineers. In the past, events like Google Dance and figures such as Matt Cutts gave marketers a chance to discuss changes with the people building the product. Now the platform is closed; feedback goes through official channels and representatives like Ginny Marvin. In Van Wagner’s view, Marvin’s experience on the advertiser side helps her understand market problems, but a peer-to-peer structural dialogue no longer exists.

The main pitfall: automation does not cancel the need for business knowledge. If you know for sure that queries containing ‘free’ or ‘vacancy’ do not lead to sales, it is irrational to make the AI relearn at your expense. Systems should be able to accept expert constraints as input, not only derive them through trial and error.

Conflict of Interest: Why Google’s Recommendations Are Not Always in the Advertiser’s Best Interest

Advertisers want profitable sales and measurable results. Google wants you to spend your budget. According to Search Engine Land, Van Wagner emphasizes: the platform’s commercial incentives and business goals do not automatically align. This is not about dishonesty among Google employees but about the structure of the business.

A classic example is the default broad match. A newcomer launches a campaign without changing the setting and pays to train the system on things the platform already knows. A broader risk: advertisers increasingly spend their own money teaching algorithms what they already know about their business — seasonality, product margins, off-target geography.

Van Wagner does not call for abandoning automation — it is moving in one direction. But he insists: the platform should provide tools for implementing existing expertise before training begins, not after the budget has been burned.

Lessons from History: Florida, Google Instant, and the Risk of Depending on One Channel

In 2003 the Florida update, released for the holiday season, wiped out the rankings of companies that had built their business on yesterday’s SEO techniques. Some of the tactics were aggressive, but the lesson is universal: organic rankings can vanish overnight. According to Search Engine Land, this incident cemented Van Wagner’s place in paid search — PPC offered an alternative traffic source independent of algorithmic whims.

Another episode was Google Instant, when results began to appear as the user typed the query. Google gained massive free media exposure while Microsoft spent millions advertising Bing. The takeaway: a platform’s ability to turn a product change into a news event is a powerful competitive asset.

Microsoft followed a mimicry path: it made its advertising system familiar to Google users. This lowered the entry barrier for cross-platform campaigns but erased differentiation. For advertisers it is convenient — one skill set, one reporting — yet it increases dependence on a single set of rules.

What This Means for Small Business in Russia and the CIS: Concrete Steps

The situation in our market has nuances: Yandex.Direct remains the main player, Google Ads works through intermediaries or in a self-service mode with limitations, but the platforms’ logic is similar. Automation of bidding strategies, match-type expansion, interface recommendations — all of this has arrived here as well.

  1. Audit default settings. Check each campaign for broad match, automatic extensions, conversion goals pulled from analytics without verification. Turn off what you do not control.
  2. Lock in negative keywords at the account and campaign level before launching automated strategies. The list of irrelevant queries (‘free’, ‘do it yourself’, ‘job’, ‘internship’, irrelevant regions) is your asset, not junk for AI training.
  3. Keep a change-and-error log. Van Wagner gives an example: an accidentally added negative keyword ‘credit’ in an expensive financial campaign burned about $7,000 in four hours. Acknowledging the mistake and correcting it quickly would have saved more than hunting for external culprits.
  4. Split responsibility: platform handles scale and speed, you handle economics. Do not accept the recommendation ‘raise budget by 20%’ without a unit-economics calculation. Auto-strategies optimize a goal (click, conversion, ROAS) but do not know your margin, LTV, or operational constraints.
  5. Maintain a second traffic source. Dependence on a single search engine is a risk akin to the 2003 SEO dependence. Set up Yandex.Direct and, if needed, social-network or marketplace ads so that shutting off one channel does not halt sales.
  6. Seek independent expertise. Platform conferences and blogs are useful, but filter the information: who is paying for the talk, whose product is being promoted. According to Search Engine Land, an independent editorial voice is the main thing the market loses without outlets like Search Engine Land. In the Russian-language segment look for authors who critique platform updates rather than merely repeating press releases.

Main Takeaway: Skepticism Toward Algorithms Is Part of the Job

Over two decades the tools have changed radically, but the principle has not. The platform is a business with its own KPIs. Automation removes routine, but it shifts to the advertiser the responsibility for high-quality input data: goals, negative keywords, conversion values, product margins. No ‘smart campaign’ can replace decisions: what margin you are willing to give up per order, which customers you do not want, where your delivery geography ends.

Van Wagner puts it simply: do not trust default recommendations. Verify each setting against your economics. In Russia and the CIS this is more relevant than anywhere else: the cost of a mistake in rubles, given current exchange rates and inflation, is higher, and platform support is often limited to chatbots. Controlling what the algorithm learns is the only way to avoid paying for someone else’s training with your budget.

Frequently Asked Questions

What has changed in Google Ads over the past 20 years?

Over the past 20 years, Google Ads has evolved significantly from manual management to sophisticated automated systems. New ad formats have emerged, targeting and analytics capabilities have expanded, and artificial intelligence has become a key element in optimization.

Why does automation not eliminate a business's responsibility?

Automation in Google Ads streamlines routine tasks, but strategic planning, goal setting, and budget control remain the business's responsibility. It is essential to set appropriate parameters for automated systems and analyze their results.

What main mistakes do companies make when working with Google Ads?

Common errors include lacking a clear strategy, neglecting data analytics, and blindly trusting automation without proper oversight. Additionally, poorly optimized landing pages and ignoring ad testing are widespread.

Based on: Search Engine Land — Matt Van Wagner shares what’s changed, what disappeared and what PPC still gets wrong

Analysis & review by
Alexander K.

Web development, digital advertising and marketing expert since 2012. Founder of AIFORSALE.

How we can help
Online storesMarketingAI agentsContact us
← All articles