Tag Archives: Google Instant

Search while you type isn't a new concept, it has been around for years but, it doesn't really work for users. Predictive text on the other hand does work for users, it's really simple and extremely fast. Google's new Instant search results combine predictive analysis with instant results and a new scroll to search feature that automatically suggests predicted queries to users all in real time. These predictions are based on years of data and billions of previous searches but, Google's results are the same. What makes Instant search radically different, is speed and most of all the feedback it gives users.

For example, Instant search provides users with feedback about misspellings and provides suggested spellings for queries before users even search. This is no accident, Google knows users intend to spell queries correctly and has been working behind the scenes on spelling improvements for some time. Improvements like these have been integrated in and rolled out with Google Instant. Instant feedback about misspellings is great for users but, may come as a shock to webmasters focused on tricking users with misspellings as their SEO strategy. Google Instant is no threat to ethical, white hat SEO efforts or unique quality content that users value. Bottom line, users "aren’t going to fundamentally change what they’re looking for."

In addition to improved spelling, Google has also recently improved how they handle proper nouns. Better handling of proper nouns helps Google extract more information, especially about named entities. Proper nouns and named entities often share a common trait, they're capitalized. Google already has entity related patents and recently acquired MetaWeb a company that specializes in this field. Mapping multiple named entities to one "thing", increases data captured about each entity as well as the whole. Improved understanding of named entities, improves data about potentially vital pages and increase the quality of results even in the absence of relevant keywords. All of these help Google Instant entice users to explore more of the space around their query. Combining better spelling with better quality and better targeted results, decreases the percentage of "unique queries". These unique queries are difficult to monetize and often result in a poor experience.

The day before Google launched Instant search, CEO Eric Schmidt said, "Never underestimate the importance of fast!" When it comes to speed, Google Instant makes other sites, including Yahoo and Bing seem much slower by comparison. Google Instant is so fast in fact, it increases the perceived latency of other sites. This factor could help increase Google's market share especially, if dedicated users start leaving other engines for Google. Increased focused on site performance is no doubt more critical than ever before.

A few other notes, Instant has Google's improved triggering for realtime queries and that could tie directly into their "Social Layer" scheduled for release in Q4 2010. You'll find Google Squared technology in Instant results for queries like [inventor of airplane]. You may notice, the scroll to search feature actually pushes results down the page and the footer search box is no more. It's possible that Google Instant's GUI emphasizes images, video and highly positioned AdWords ads (colored background) more, because they flash in and out of view in certain cases.

What does the near future hold for SEO, PPC and analytics? In coming months expect to see, more data missing from analytics and fragmented with Google Webmaster Tools and/or other sources. It's quite possible Google Instant thwarts automated queries on some levels and for that reason, ranking reporting software may be even more inaccurate. AdWords impression data will be less accurate for testing and virtually worthless in terms of historical comparison.

Google Instant Event Live blog, please excuse typos and spelling errors.

According to Marissa Mayer 11 hours can be saved for every second spent searching thanks to Google Instant. Google Instant revolutionizes how users interact with Google. This new search feature has 3 main components:

  • Instant results - Results appear instantly to the user.
  • Predictions - Google is able to predict what users search for as they search.
  • Scroll to search - This feature offers users options for their query. Google Instant is a better, faster, cleaner experience but like riding a bike... you never forget how to ride.

google instant efficiency

Google Serps

Google Incremental Search

All of these new features render the "search" button almost useless. Google is still seamless users just have additional options. Google Instant provides feedback that users can use find what they are looking for online. Feedback from Google Instant is also used via mobile but this feature won't be available until Fall 2010. Google "Trusted Tester" video illustrates testers interaction with Google Instant which will roll out later today.

Up next is Ben Gomes to talk about technical challenges with Google Instant. According to Ben, Google Instant had to be efficient. To do this Google focused on user interface design, search as an application and efficiency to scale. In order to test all of this Google tested more than 100 testers in their usability lab. These "Trusted Testers" were monitored using eye tracking technology.

Google Autocomplete

In the past Google has relied on (X)HTML but Google Instant is AJAX based. To do this Google first sends autocomplete search requests, autocomplete predicitions and then receives results page. In addition, Google detects when the system breaks.

If Google returns search results with every letter and there are 20 characters per query on average, how can Google do this and not melt down their servers.

When it comes to scale and efficiency Google can maintain low cost per query by anticipating the query. Google does this because efficiency matters to users and users matter to Google. For that reason Google has already rolled out 500 updates this year. Today is the biggest change to hit Google in terms of user interaction in years. Google Instant makes search available to users.

According to Marissa Mayer, Interaction, Comprehensiveness and Understanding (users, web, squared) are the future of search. Google Instant is first step into the future of search. According to Marissa, users will save 350 million hours a year.

The event is closing with a video by Google's creative lab inspired by Bob Dylan.... Results as you type, no clicks, search on!

Does this change PPC? No change in how ads are served or ranked.

Are 20% of queries still unique? Not sure but if you make search easier and faster more folks search

@Wired asks how Google Instant ties into search history? Marissa Mayer, fields and says the inputs remain unchanged.

How much personal data is required for speed improvements? How much does this tax data centers?
Marissa Mayer says personal data requirements are the same. Google is using prediction not personal data. Johanna answers the data center question and confirms the cost has increased but engineering has been able to maintain.

Sergey Brin fields "human computer interaction" question in terms of speed of advancement in this area. According to Sergey techno advancements and openmindenesss are resulting in major advances that will change the way you interact with these kinds of devices.

What impact will Google Instant have on SEO? According to Ben, there will be little impact in terms of SEO.

When will this be available in China? Marissa says Google's goal is to roll out on Google.com.hk

Do you believe users want faster search? YES By speeding up search Google expects more traffic and more queries. Marissa says as users adapt to Google Instant the won't go back.

Question from author of "Inside Larry and Sergy's Brain" asks about privacy and what impact Instant has. According to Google Co-Founder Sergey Brin Google spends lots of time focused on privacy and Google Instant is no different.

BBC asks how Google Instant will change behavior? According to Ben neighboring queries become more important.

During testing how many users didn't want to use Google Instant? How does Google Instant relate to caffeine. According to Mayer, few didn't like. Ben follows up by saying the two have an indirect conncetion.

How does this impact SEM (SEO &PPC)? Little...

Next question, "Sergey did you ever think you would reach this point". Based on Moore's Law of Computation is has been amazing and a tribute to people who work on computing everywhere.

How is Google getting folks to contribute content? Sergy, great question. That has been one focus for a long time, for example AdSense. For more user gen content Google has worked on projects like Blogger, Google Docs. In addition, there are lots of places where users gen content.

Are there plans to make faster? Ben says we are going to enjoy the speed but you "ain't seen nothing yet".

Ad Age asks about brand impact and rankings in terms of page 2? Ben says user intent doesn't change no matter what predictions are going on behind the scenes. Johanna points out the Google Instant gets users to results faster. Google totaly focuses on users and that is good for ads.

Will Google server less ads? No

Will this change result in more searches and more adds clicked? Ben says focus on user and all else will follow.

When will Google release Google Instant in Japan and China because characters can cause problems? Marissa Mayer, says expect all these in addition to mobile in a few months.

Is this part of a bigger plan? Mayer, yes... In fact we look at queries from years ago. Ben says we are always working to make things better.

ONE FINAL QUESTION, oh who could it be? Danny Sullivan asks about what happens when results are not correct? According to Google, users learn how to pull things up from the top except for Danny's Dogpile question. Google asks Danny, why he is using Google to search for [search engines] :)

Google Instant Meetup later today at 111 Minna in SFO