NO XMAS PRESENTS FOR Google Now and Siri: Chirpy scamps get a C+
Analyst's test shows Cupertino dumping Google's results, both must try HARDER
A series of tests on the effectiveness of iOS and Android's voice-controlled search apps shows that Google Now is catching up on Apple's Siri software, but still has some way to go in beating its Cupertinian competitor.
Piper Jaffray analyst Gene Munster asked the two personal-assistant programs 800 questions apiece – half spoken outdoors and half indoors – about nearby things, shopping and other commerce, directions to places, and general information, and also tried to run the operating system by voice.
In both environments both Google Now and Siri understood the question correctly 94 per cent of the time; Google answered with 81 per cent accuracy, compared to Siri's 83 per cent. Overall, Munster gave both utilities a C+.
Unsurprisingly Google Now was strongest on navigation questions, but also did well on pulling up local services and information. Siri scored very well for allowing fuller voice control of the operating system, with Google Now not understanding how to stop and pause music, for example.
"We believe that our test suggests that Google Now and Siri are comparable in terms of understanding the spoken queries and returning the correct result," wrote Munster in a research note to investors.
This is the third time Munster has carried out the testing and he noted that whereas Apple's results have improved by two percentage points in the last year, Google Now has come on 9 per cent, with the Chocolate Factory's engine seeing its biggest improvement in its ability to take voice commands in noisier environments.
One major change over the last series of tests is Apple dramatically reducing Siri's reliance on search results from its rival Google. Last year Siri used Google's data 27 per cent of the time, but in the latest testing round this had dropped to just 3 per cent. To compensate Siri now makes greater use of Bing, Wolfram Alpha, and Wikipedia. ®
Sponsored: Hyper-scale data management