Search is amazing compared to 10 years ago at AltaVista
We all use web search engines a lot...about 3 Billion times a day around the world. We hear pundits say that search sucks, or hasn't advanced much beyond 10 blue links. Eleven years ago I was director of engineering at AltaVista …
Context & Ripple Effects
Don Dodge, who ran engineering at AltaVista eleven years before writing this, is answering a specific critique that has been building since the mid-2000s: Jakob Nielsen's 2006 argument that search engines act as leeches on the web, and the recurring pundit line that search 'sucks' or has gone nowhere beyond ten blue links. His counterweight is scale and craft — roughly three billion queries a day worldwide — plus his insider's view of how far ranking, indexing, and result quality have moved since the AltaVista era.
The debate had a predictive thread running through it: Search Engine Land was already asking in 2007 what search would look like by 2010 and declaring local search finally delivering on its promise that September, so Dodge's 2011 retrospective lands as a scorecard moment — did search actually advance as fast as its defenders claimed, and where does the next decade of improvement have to come from?
First-order effects
- Dodge directly rebuts the '10 blue links' critique, giving Google-era search a credibility argument grounded in his own AltaVista tenure rather than vendor marketing.
- AltaVista becomes the baseline in the comparison — a former category leader now serving mainly as the yardstick against which current search quality is measured.
Second-order effects
- Pundits pushing the stagnation narrative are forced to argue specifics — freshness, spam, local results — rather than the blanket claim, shifting the debate toward measurable quality dimensions like the local search gains Search Engine Land highlighted back in 2007.
Third-order effects
- With query volume at billions per day and incremental quality gains getting harder to see, the structural question becomes whether future differentiation comes from relevance tuning alone or from new interaction models — the tension that keeps resurfacing every time someone declares search finished.
The trend: Web search is transitioning from a maturing relevance-refinement game into a contest over new interaction models, as each generation of critics declares ten blue links a dead end and each defender points to cumulative quality gains since the AltaVista era.