{"id":4970,"date":"2026-08-09T01:42:06","date_gmt":"2026-08-08T22:42:06","guid":{"rendered":"https:\/\/ivitskiy.com\/blog\/definitive-traffic-seminar-2023\/"},"modified":"2026-08-09T01:42:06","modified_gmt":"2026-08-08T22:42:06","slug":"definitive-traffic-seminar-2023","status":"publish","type":"post","link":"https:\/\/ivitskiy.com\/blog\/en\/definitive-traffic-seminar-2023\/","title":{"rendered":"How Google’s AI Works Internally: Igor Ivitskiy at Definitive Traffic Seminar 2023"},"content":{"rendered":"\n
In short:<\/strong> Google Ads AI learns from the conversion signals an advertiser sends back, so a cheap conversion can be harmful when it represents the wrong person or action. At Definitive Traffic Seminar, Igor Ivitskiy explained why signal speed, statistical volume and business value have to be judged together. His warning came from a campaign that produced 22,000 apparent conversions at $8 each in 10 days, yet the sales team found that none of those users bought anything. The lesson was to optimize for evidence of real business value, not for an attractive platform number.<\/div>\n\n\n\n

Definitive Traffic Seminar<\/a> took place in Chicago, US, on May 23\u201325, 2023, as an in-person conference. The session was presented by Igor Ivitskiy<\/a>.<\/p>\n\n\n\n

\u201cHow Google\u2019s AI Works Internally\u201d<\/strong> connected three questions that sit behind campaign performance: which conversion signal teaches the system, how AI interprets audience context, and how apparently successful optimization can amplify bots or unqualified leads. It placed those questions beyond routine keyword research for Google Ads<\/a> and inside the feedback loop that guides automated campaigns.<\/p>\n\n\n\n

$180,000<\/div>
spent in 10 days while 22,000 apparent conversions at $8 produced no buyers<\/div><\/div>\n
Source: Definitive Traffic Seminar 2023 talk, Igor Ivitskiy.<\/div>\n\n\n\n

Google AI follows the feedback it receives<\/h2>\n\n\n\n
\"Search
Search query passing through layered ranking stages<\/figcaption><\/figure>\n\n\n\n

Igor\u2019s core analogy came from his earlier research on non-Newtonian fluids. Both the experimental medium and an advertising market are dynamic, unpredictable and time-dependent. A delayed or sparse measurement can describe a state that has already changed. In an automated campaign, the conversion signal plays the role of that measurement.<\/p>\n\n\n\n

That signal has three competing qualities: it should arrive close enough to the ad interaction, occur often enough for the system to distinguish a pattern from noise, and represent something that matters to the business. The hard part is that the fastest and most frequent event is often weaker than a qualified lead or a payment.<\/p>\n\n\n\n

Automation therefore does not remove judgment. It moves judgment upstream, into the choice and interpretation of data. The same principle applies when campaigns use wider reach, including broad-match keywords<\/a>: more freedom for the system makes the quality of the feedback loop more important.<\/p>\n\n\n\n

\n
1<\/span>Conversion signal<\/strong>
The system needs feedback that is timely, repeatable and connected to business value.<\/div>\n
2<\/span>Audience context<\/strong>
Clear semantic signals help AI identify people whose interests or recent intent fit the offer.<\/div>\n
3<\/span>Feedback integrity<\/strong>
Bots and accidental form fills can teach the system to find more of the wrong outcome.<\/div>\n<\/div>\n
Source: three-part framework from the Definitive Traffic Seminar 2023 talk.<\/div>\n\n\n\n

The metric is not the objective<\/strong><\/p>

I treat every conversion action as a teaching signal. If the action is fast and frequent but has no commercial meaning, better optimization can move the campaign further away from the business result.<\/p><\/section>\n\n\n\n

This way of thinking is useful beyond one campaign type: it gives marketers a framework for questioning what an automated account is actually learning. Our school returns to that strategic layer in its training, while the talk recording preserves the detailed examples and mechanics.<\/p>\n\n\n\n\n\n\n\n

Key insights<\/h2>\n\n\n\n