Vivian Voss

Google Search: The Time Tax

performance web

Performance-Fresser ■ Episode 19

I have been supporting Kagi since before it launched. Not because I enjoy paying for things that used to be free. Because my time is not negotiable and my search results are not advertising inventory. Ten dollars for hours saved in research every month is not an expense. It is a rather obvious decision.

Thirty per cent of a knowledge worker's day goes to searching for information. One full month per year lost to bad results. Not obscure things. Things that should have been the first result. The tool that was built to save time now consumes it. One does appreciate an industry that turned "I'm Feeling Lucky" into a 30-per-cent-of-your-day problem.

The Decay

Google's user trust score dropped from 7.05 to 5.73 in a single year. Its global market share fell from 91.47 per cent to 89.57 per cent: the largest annual drop in a decade. People are not leaving because something better arrived. They are leaving because Google got worse.

Leipzig University studied 7,392 queries across Google, Bing, and DuckDuckGo. The finding is worth reading twice: higher rank correlates with lower text quality. The majority of top results are affiliate-monetised content optimised for the algorithm, not for the reader. Algorithm updates have, and one does enjoy this quote, "only a temporary positive effect." The spam adapts. The algorithm updates. The spam adapts again. The search engine updates again. The spam wins. Comprehensively.

Google Search: The Numbers Trust Score 7.05 5.73 -18.7% in one year Market Share 91.5% 89.6% largest drop in a decade Zero-Click 58.5% of searches end at Google Leipzig University (7,392 queries) Higher rank correlates with lower text quality. Algorithm updates have "only a temporary positive effect." Sources: Market My Market, SociallyIn, SparkToro/Datos 2024, Bevendorff et al. 2024 Forrester, Slite Enterprise Search Survey 2025

This is not a fringe complaint from privacy enthusiasts. This is the mainstream user experience of the world's dominant search engine, measured by academics, confirmed by the market, and visible to anyone who has searched for a recipe in the past twelve months and been offered seventeen life stories before the ingredient list.

The AI Overviews Tax

Half of all US queries now trigger AI Overviews. The organic click-through rate dropped 61 per cent. Zero-click rate with AI Overviews: 83 per cent. For every hundred searches, eighty-three give you an AI summary and nothing else. The open web receives seventeen visitors. The rest stay inside Google, reading an answer that may or may not be correct, with no way to verify it because the source links are hidden beneath a fold that nobody unfolds.

AI Overviews: The Click Collapse 50% of US queries trigger AI Overviews -61% organic CTR with AI Overviews 83% zero-click rate end at Google 37% finance answers inaccurate Google removed health queries from AI Overviews entirely in January 2026 after recommending gluing cheese to pizza and eating rocks for minerals. 58.5% of all Google searches end without a click to the open web (SparkToro 2024). The search engine no longer connects you to the web. It replaces it.

37 per cent of finance-related AI Overviews contain inaccuracies. Google removed health queries entirely in January 2026 after the glue-on-pizza and rocks-for-minerals incident made international news. The feature that was meant to save you a click now gives you wrong answers faster. Rather efficient, that.

58.5 per cent of all Google searches end without a click to the open web. For every 1,000 queries, only 360 clicks reach an independent website. The rest stays inside Google or goes nowhere. The search engine no longer connects you to the web. It replaces it.

The Alternative

When a search engine is free, you are not the customer. You are the inventory. Every result is a negotiation between what you need and what someone paid to show you. The first page of results is not a ranking of relevance. It is an auction.

When you pay, that negotiation disappears. The product becomes the answer.

Kagi charges ten dollars per month. No ads. No tracking. No AI Overviews inserted between you and your answer. Two independent, self-built crawlers (Teclis and TinyGem), curated sources, and 50,000 paying subscribers whose incentive is search quality, not ad revenue. It is the best result quality I have seen from any search engine, and I have tried most of them.

Brave Search runs its own independent index, built from scratch, the only one outside Big Tech in the western world. Fifty million queries per day. No surveillance. No ads. Both Kagi and Brave are funded by users, not advertisers. Both are independent of Google's index. Both work.

Who Pays, Who Benefits Free (Google, Bing) Customer: advertisers Product: your attention Incentive: ad clicks, not answers Paid (Kagi, Brave) Customer: you Product: search quality Incentive: answers, not ads The price of a good search engine is a coffee and a slice of cake per month. The price of a free one is your afternoon.

The price of a good search engine is a coffee and a slice of cake per month. The price of a free one is your afternoon.

The Pattern

Google Search was the fastest tool on the internet. Then it became an ad platform with a search feature. Then it became an AI summary engine that sometimes, when the mood takes it, gets the answer right. Each iteration made the company richer and the user slower.

The "I'm Feeling Lucky" button still exists. The feeling, rather less so.

30% of a knowledge worker's day spent searching. Trust score down from 7.05 to 5.73 in one year. 83% zero-click rate with AI Overviews. 37% of finance answers inaccurate. 58.5% of searches end without a click to the open web. The search engine that once finished your sentence now wastes your afternoon.