Posted:
This month we have many improvements to celebrate. With 40 changes reported, that marks a new record for our monthly series on search quality. Most of the updates rolled out earlier this month, and a handful are actually rolling out today and tomorrow. We continue to improve many of our systems, including related searches, sitelinks, autocomplete, UI elements, indexing, synonyms, SafeSearch and more. Each individual change is subtle and important, and over time they add up to a radically improved search engine.

Here’s the list for February:
  • More coverage for related searches. [launch codename “Fuzhou”] This launch brings in a new data source to help generate the “Searches related to” section, increasing coverage significantly so the feature will appear for more queries. This section contains search queries that can help you refine what you’re searching for.
  • Tweak to categorizer for expanded sitelinks. [launch codename “Snippy”, project codename “Megasitelinks”] This improvement adjusts a signal we use to try and identify duplicate snippets. We were applying a categorizer that wasn’t performing well for our expanded sitelinks, so we’ve stopped applying the categorizer in those cases. The result is more relevant sitelinks.
  • Less duplication in expanded sitelinks. [launch codename “thanksgiving”, project codename “Megasitelinks”] We’ve adjusted signals to reduce duplication in the snippets for expanded sitelinks. Now we generate relevant snippets based more on the page content and less on the query.
  • More consistent thumbnail sizes on results page. We’ve adjusted the thumbnail size for most image content appearing on the results page, providing a more consistent experience across result types, and also across mobile and tablet. The new sizes apply to rich snippet results for recipes and applications, movie posters, shopping results, book results, news results and more.
  • More locally relevant predictions in YouTube. [project codename “Suggest”] We’ve improved the ranking for predictions in YouTube to provide more locally relevant queries. For example, for the query [lady gaga in ] performed on the US version of YouTube, we might predict [lady gaga in times square], but for the same search performed on the Indian version of YouTube, we might predict [lady gaga in India].
  • More accurate detection of official pages. [launch codename “WRE”] We’ve made an adjustment to how we detect official pages to make more accurate identifications. The result is that many pages that were previously misidentified as official will no longer be.
  • Refreshed per-URL country information. [Launch codename “longdew”, project codename “country-id data refresh”] We updated the country associations for URLs to use more recent data.
  • Expand the size of our images index in Universal Search. [launch codename “terra”, project codename “Images Universal”] We launched a change to expand the corpus of results for which we show images in Universal Search. This is especially helpful to give more relevant images on a larger set of searches.
  • Minor tuning of autocomplete policy algorithms. [project codename “Suggest”] We have a narrow set of policies for autocomplete for offensive and inappropriate terms. This improvement continues to refine the algorithms we use to implement these policies.
  • “Site:” query update [launch codename “Semicolon”, project codename “Dice”] This change improves the ranking for queries using the “site:” operator by increasing the diversity of results.
  • Improved detection for SafeSearch in Image Search. [launch codename "Michandro", project codename “SafeSearch”] This change improves our signals for detecting adult content in Image Search, aligning the signals more closely with the signals we use for our other search results.
  • Interval based history tracking for indexing. [project codename “Intervals”] This improvement changes the signals we use in document tracking algorithms. 
  • Improvements to foreign language synonyms. [launch codename “floating context synonyms”, project codename “Synonyms”] This change applies an improvement we previously launched for English to all other languages. The net impact is that you’ll more often find relevant pages that include synonyms for your query terms.
  • Disabling two old fresh query classifiers. [launch codename “Mango”, project codename “Freshness”] As search evolves and new signals and classifiers are applied to rank search results, sometimes old algorithms get outdated. This improvement disables two old classifiers related to query freshness.
  • More organized search results for Google Korea. [launch codename “smoothieking”, project codename “Sokoban4”] This significant improvement to search in Korea better organizes the search results into sections for news, blogs and homepages.
  • Fresher images. [launch codename “tumeric”] We’ve adjusted our signals for surfacing fresh images. Now we can more often surface fresh images when they appear on the web.
  • Update to the Google bar. [project codename “Kennedy”] We continue to iterate in our efforts to deliver a beautifully simple experience across Google products, and as part of that this month we made further adjustments to the Google bar. The biggest change is that we’ve replaced the drop-down Google menu in the November redesign with a consistent and expanded set of links running across the top of the page.
  • Adding three new languages to classifier related to error pages. [launch codename "PNI", project codename "Soft404"] We have signals designed to detect crypto 404 pages (also known as “soft 404s”), pages that return valid text to a browser but the text only contain error messages, such as “Page not found.” It’s rare that a user will be looking for such a page, so it’s important we be able to detect them. This change extends a particular classifier to Portuguese, Dutch and Italian.
  • Improvements to travel-related searches. [launch codename “nesehorn”] We’ve made improvements to triggering for a variety of flight-related search queries. These changes improve the user experience for our Flight Search feature with users getting more accurate flight results.
  • Data refresh for related searches signal. [launch codename “Chicago”, project codename “Related Search”] One of the many signals we look at to generate the “Searches related to” section is the queries users type in succession. If users very often search for [apple] right after [banana], that’s a sign the two might be related. This update refreshes the model we use to generate these refinements, leading to more relevant queries to try.
  • International launch of shopping rich snippets. [project codename “rich snippets”] Shopping rich snippets help you more quickly identify which sites are likely to have the most relevant product for your needs, highlighting product prices, availability, ratings and review counts. This month we expanded shopping rich snippets globally (they were previously only available in the US, Japan and Germany).
  • Improvements to Korean spelling. This launch improves spelling corrections when the user performs a Korean query in the wrong keyboard mode (also known as an "IME", or input method editor). Specifically, this change helps users who mistakenly enter Hangul queries in Latin mode or vice-versa.
  • Improvements to freshness. [launch codename “iotfreshweb”, project codename “Freshness”] We’ve applied new signals which help us surface fresh content in our results even more quickly than before.
  • Web History in 20 new countries. With Web History, you can browse and search over your search history and webpages you've visited. You will also get personalized search results that are more relevant to you, based on what you’ve searched for and which sites you’ve visited in the past. In order to deliver more relevant and personalized search results, we’ve launched Web History in Malaysia, Pakistan, Philippines, Morocco, Belarus, Kazakhstan, Estonia, Kuwait, Iraq, Sri Lanka, Tunisia, Nigeria, Lebanon, Luxembourg, Bosnia and Herzegowina, Azerbaijan, Jamaica, Trinidad and Tobago, Republic of Moldova, and Ghana. Web History is turned on only for people who have a Google Account and previously enabled Web History.
  • Improved snippets for video channels. Some search results are links to channels with many different videos, whether on mtv.com, Hulu or YouTube. We’ve had a feature for a while now that displays snippets for these results including direct links to the videos in the channel, and this improvement increases quality and expands coverage of these rich “decorated” snippets. We’ve also made some improvements to our backends used to generate the snippets.
  • Improvements to ranking for local search results. [launch codename “Venice”] This improvement improves the triggering of Local Universal results by relying more on the ranking of our main search results as a signal. 
  • Improvements to English spell correction. [launch codename “Kamehameha”] This change improves spelling correction quality in English, especially for rare queries, by making one of our scoring functions more accurate.
  • Improvements to coverage of News Universal. [launch codename “final destination”] We’ve fixed a bug that caused News Universal results not to appear in cases when our testing indicates they’d be very useful.
  • Consolidation of signals for spiking topics. [launch codename “news deserving score”, project codename “Freshness”] We use a number of signals to detect when a new topic is spiking in popularity. This change consolidates some of the signals so we can rely on signals we can compute in realtime, rather than signals that need to be processed offline. This eliminates redundancy in our systems and helps to ensure we can continue to detect spiking topics as quickly as possible.
  • Better triggering for Turkish weather search feature. [launch codename “hava”] We’ve tuned the signals we use to decide when to present Turkish users with the weather search feature. The result is that we’re able to provide our users with the weather forecast right on the results page with more frequency and accuracy.
  • Visual refresh to account settings page. We completed a visual refresh of the account settings page, making the page more consistent with the rest of our constantly evolving design.
  • Panda update. This launch refreshes data in the Panda system, making it more accurate and more sensitive to recent changes on the web.
  • Link evaluation. We often use characteristics of links to help us figure out the topic of a linked page. We have changed the way in which we evaluate links; in particular, we are turning off a method of link analysis that we used for several years. We often rearchitect or turn off parts of our scoring in order to keep our system maintainable, clean and understandable.
  • SafeSearch update. We have updated how we deal with adult content, making it more accurate and robust. Now, irrelevant adult content is less likely to show up for many queries.
  • Spam update. In the process of investigating some potential spam, we found and fixed some weaknesses in our spam protections.
  • Improved local results. We launched a new system to find results from a user’s city more reliably. Now we’re better able to detect when both queries and documents are local to the user.

And here are a few more changes we’ve already blogged about separately:


Posted:
(Cross-posted on the Official Google Blog)

Time to polish the champagne flutes and brush up on your movie trivia—it’s almost Oscar night again. Before you make any Oscar bets, get an edge by exploring Google Insights for Search. Out of the major entertainment awards shows (Tonys, Emmys, Grammys) the Oscars are the most popular in terms of search volume, and as we discovered last year, patterns in search behavior can help us predict which stars will go home with shiny gold statues. So without further delay, let’s open the (search) envelopes and see who the Oscar (may) go to this year.

Best Picture
Last year we found that for three years running, the films that won best picture had two things in common when it came to search data. First, the winning movies had all shown an upward trend in search volume for at least four consecutive weeks during the previous year. Second, within the U.S. the winning film had the highest percentage of its searches originating from the state of New York. Looking at search data for 2011, there were three films that satisfied these conditions—The King’s Speech, The Social Network and Black Swan. Our prediction was on the mark: The King’s Speech took home the Oscar in 2011.

This year, if we assume the two “winning conditions”—at least four consecutive weeks of increasing search volume plus highest regional interest from New York—will apply, then we can narrow down the nominees to a field of four: The Artist, Extremely Loud & Incredibly Close, Midnight in Paris and War Horse. But how to go from four to one?

Let’s again look back at last year’s finalists. When you compare search query volumes for The King’s Speech, The Social Network and Black Swan, the winning film, The King’s Speech, had the lowest search volume throughout the year leading up to the Oscars. It was the underdog that took home the statue.


We tried the same test on the Best Picture nominees from 2010. The nominated movies in 2010 that met the two conditions were The Hurt Locker and Inglourious Basterds. Once again, it was the the winning film, The Hurt Locker, that had lower search volume in 2009.

If the underdog trend holds this year, Extremely Loud & Incredibly Close could be our surprise winner. If we go strictly by search popularity, however, The Artist or Midnight in Paris have the best chances—among our group of four, they’re currently blowing the competition out of the water.


If we’re having a popularity contest, it’s only fair to look at all nine nominees for best picture. A 2-step comparison shows that the most popular films by search volume are The Help and Martin Scorsese's Hugo.

Best Actor
Of this year’s five nominees for Best Actor, Brad Pitt (Moneyball) is clearly the most popular—searches for Brad in the last 12 months far outpace any of the other leading men, as was the case in 2009 when he was nominated for The Curious Case of Benjamin Button. However, it could be Brad’s famous good looks that have us searching, which brings about a good point: the most searched-for nominee doesn’t guarantee a win. James Franco had the highest search volume in 2011 but Colin Firth won, and in 2010, George Clooney was the most-searched nominee but Jeff Bridges took home the Oscar.

The pattern emerging over the past few years is that the winner is generally in the middle of the pack in terms of searches and has relatively steady search volume throughout the year. First-time nominee Gary Oldman (Tinker Tailor Soldier Spy) fits that bill this year, but so does George Clooney (The Descendants). Maybe it will finally be George’s year to win Best Actor.


Best Actress
For the past three years, the eventual Best Actress winner has seen a spike of interest in the preceding December. Additionally, two of the three most recent winners have had the strongest regional interest within the U.S. from the cities of Los Angeles and New York City (2010 winner Sandra Bullock is the exception).

Among this year’s nominees, Rooney Mara is the clear breakout star, with a huge surge in search volume this past December for the young lead in The Girl with the Dragon Tattoo. However, it’s Meryl Streep who has the highest regional interest in NYC and while Rooney is popular in LA, she’s even more popular in San Francisco. So it could be her name that is announced when the envelope is opened—or not.


Of course, we don’t have a Magic 8-Ball or access to the names in those top-secret envelopes, so our predictions are just that—but it’s always enjoyable to look at how what people are interested in online plays out in the real world. As you prepare for your Oscar viewing parties this year, put a visit to Insights for Search on your checklist before the red carpet walk begins (fun fact: searches for [red carpet] peak at Oscar time every year). Between dry cleaning your tuxedo and making hors d'oeuvres, tune in to a pre-Oscar hangout on the +Good Morning America page, where the live discussion will be the fashion dos and don'ts of the big night. You can also stay up to date on all Oscar news on +Oscars, the official Google+ Page of the Academy Awards.

Posted:
I’m a big music fan, so I spend a lot of time searching for music bands and artists. When I hear a new song or a friend tells me about an artist I should check out, I search to watch their music videos, find out recent news about them, and to learn more about them in general. I can also see right on the results page which sites have songs or samples of songs that I can listen to.

But to me, music is best heard live, so it’s always exciting to find out that a band I love is coming to town. For fellow concert-goers, now when you search for bands or artists, you may see upcoming concert tour dates right on the results page if they’re playing in your area. If they aren’t touring near you, the new results for concerts won’t appear, but if the band happens to be coming to your town within the next few months, you can see the concert dates listed under the band’s official website. You can then click on the band’s official site to learn more or click on other web pages to learn more about the event or to buy tickets.

For example, I was searching for the [black keys] to listen to some of their music online. I wasn’t explicitly looking for their tour schedule, but now I know I should keep my calendar clear on Friday May 4, since they will be coming back to the Bay Area, and I remember how great they were at Outside Lands last year:


To find upcoming concert tour dates, we aggregate relevant data for events from multiple websites and show it under the band’s official website with links to the event sites where you can find out more about the event or purchase tickets.

If you have a website listing upcoming events and would like them to appear in search, you should add rich snippets markup to your web pages. After following the instructions to mark up your events, use the rich snippets testing tool to test your markup and see how it would appear on Google.

This is currently available to those searching on google.com in English, and we’ll continue to expand the feature to more countries in the future.

Posted:
(Cross-posted on the Official Google Blog)

Today we’re making it easier for you to hone in on that perfect image or explore your topic visually with an update to related search links. Related search links have been around for awhile—they’re the row of blue links running across the top of your image search results—but today we’re making them more visual to help you find exactly what you’re looking for or just have fun exploring.

For example, when planning a trip to Greece, I may not know what places are worth a visit, so I search for [greece] on Image Search. Now, with more visual search links, I can hover over the links on the top of the results, like [santorini greece], and see a panel pop up with images of Santorini. Without having to type more words into the search box or clicking through, I can quickly glance at the pictures of Santorini. If I decide to click through, I find new links for further refined or related searches, such as [oia santorini greece] or [santorini greece sunset]. Now I’m sold, I want to see more Santorini images.

You’ll start to see these links whenever you search for images as we roll this change out globally over the next few weeks.

Posted:
(Cross-posted on the ITA Software By Google blog)

Planning trips is fun and exciting, and ideas for where to go on your next adventure can strike you when you're in the most unexpected location. Last week I met a friend for coffee, and he told me about his recent trip to Maui. His story and photos inspired me and I immediately made a note to myself to check flights to Hawaii at home.

Since we launched Flight Search back in September, we’ve been hard at work improving the features, increasing coverage and making it easy to find flight results directly from google.com on your desktop. Starting today, we're also making it easier to find flights departing from the US on your mobile device.

For example, when you search for [flights from Chicago to Daytona Beach] you’ll see a table that shows available flights, including duration and prices. You can adjust dates on the page, or click any flight to further research and book your trip.


The Flight Search feature on mobile browsers offers all the benefits of Flight Search on desktop:
  • Find flights quickly with results that load instantly and a list that’s easy to scan. 
  • Discover places to go on a map - see ticket prices for various destinations by surfing the map. You can filter by price, airline, or flight duration. 
  • Find the best time to go - Click the calendar icon to see what dates will get you low prices. 

The feature is available on Android and iOS devices. To learn more about the Flight Search feature, see our tips at http://www.google.com/insidesearch/flights.html.

Posted:
(Cross-posted on the Official Google Blog)

It’s Valentine’s Day, and all you need is love.

Well, maybe you need a few more things.

This year, lovebirds in the U.S. are pulling out all the stops and are expected to hit a 10-year spending high on romantic goods. Whether you’re looking for a box of chocolates or buying a diamond ring, we’ve taken notes on how Google can turn any last-minute Cupid into a polished Romeo. Think of us as Cyrano de Bergerac, whispering words of wisdom in your ear for dishes, dates and romantic inspiration.  

Sweets for a sweetie
They say the way to a man’s heart is through his stomach. Searches for [valentine’s day dinner], [valentine’s day recipes], [romantic dinner] and [romantic recipes] ramp up at the start at February, hitting their peak on the 14th.


If you’re whipping up a homemade treat, you can use Google Recipe view to search for thousands of heart-shaped cookie recipes, and tailor the ingredients (and the calorie count) for the best fit.

If you’re not a maestro in the kitchen yet, we can help. Peruse some trendy dishes and learn how to cook from YouTube’s Next Chefs, who have created a highlight reel of their best aphrodisiacs. We’re not playing favorites, but we do love the Sweetest Vegan’s red velvet beet cupcakes, which will come in handy for many couples—searches for [vegan valentines] have more than tripled since 2006.

Don’t know a truffle from a trifle? Then let a professional handle the meal. If you haven’t booked a table yet, you’re not alone: searches for [valentines day reservations] typically peak on February 9 and remain high through the holiday. For ideas, check out Zagat’s recommendations for romantic hotspots in your neighborhood. A word to the wise, though: If you’re planning on popping the question on Valentine’s Day, avoid dining out, as 69 percent of those surveyed in Zagat’s recent Valentine’s Day Survey feel that restaurant proposals are “cheesy.”  

Perfect planners and last-minute cupids
Women have a head start on the menfolk when it comes to Valentine’s Day gifts, searching earlier (and more often) than their male counterparts—about 160 percent more since January.


The staple romantic gifts haven’t lost their appeal. Searches for [jewelry gifts] have grown over 10 percent, searches for [flower delivery] have increased nearly 20 percent, and searches for [couples massage] have jumped nearly 50 percent over last Valentine’s Day.

To impress, some folks are thinking outside the (heart-shaped) box. For the daring, create your own [valentines scavenger hunt]—searches are up more than 20 percent from last year. Or, add a personal touch—searches for [personalized valentines day gifts] are up over 20 percent compared to last year and searches for [homemade valentines gift] are up over 60 percent since last year. The last-minute lovers don’t have to despair, though. Visit Google Shopping to find gift ideas and filter results to see which items are in stock nearby.  

Celebrating solo
Celebrating Valentine’s Day solo doesn’t mark you as a Miss (or Mr.) Lonelyhearts. Instead, treat yourself to a night on the town. View interactive results for nearby movie showtimes on your mobile phone, and know that if you indulge your inner cynic by skipping the rom com and catching a horror film, you won’t be alone: in the past 30 days, searches for [horror movie] are 230 percent higher than searches for [romantic movie].  

Vive la romance
Everyone loves a fairytale ending, so let’s wrap up with two final ways to make your heart grow two sizes too big today. For a close-up look at romance at its finest, check out the “Awww: Romantic Proposals” YouTube Slam. Vote for your favorite mushy, creative, artistic, or—in one case—magic proposal.

Finally, today’s homepage doodle gives a nod to love, both young and old. Though “Cold, Cold Heart” plays in the background, we bet yours will warm just a bit.

Happy Valentine’s Day, everyone!

Posted:
Every day, people search on Google for health information. Many of these searches relate to symptoms they or their loved ones may be experiencing. You might be trying to understand why you’ve had a headache every morning for a week or why your child has a tummy ache all of a sudden. Our data shows that a search for symptoms is often followed by a search for a related condition.

To make the process easier, now when you search for a symptom or set of symptoms, you'll often see a list of possibly related health conditions that you can use to refine your search. The list is generated by our algorithms that analyze data from pages across the web and surface the health conditions that appear to be related to your search.

For example, if you search for [abdominal pain on my right side], you’ll be able to quickly see some potentially related conditions and learn more about them by clicking on the links in the list.


The list of health conditions you see is aggregated from what’s written on the web about the symptoms you searched. The list is not authored by doctors and of course is not advice from medical experts.

We hope this feature makes it easier and faster to research symptoms and related health conditions on Google. We’re humbled by the number of people who turn to Google with such important questions, and we are working especially hard to make our search results even more useful for health searches.


Posted by Roni Zeiger, MD, Chief Health Strategist

Posted:
Here’s the latest installment of our monthly series on “search quality highlights,” with 17 new quality improvements to read about for January. In addition to this month’s big announcement of Search plus Your World, you’ll find short summaries of other changes to our high-quality sites algorithm, spelling systems, snippets, search preferences, speed, freshness and much more. It’s all part of our ongoing effort to be transparent about how search works and the ways Google is constantly evolving to answer your questions.

Here’s the list for January:

  • Fresher results. [launch codename “nftc”] We made several adjustments to the freshness algorithm that we released in November. These are minor updates to make sure we continue to give you the freshest, most relevant results.
  • Faster autocomplete. [launch codename “Snappy Suggest”, project codename “Suggest”] We made improvements to our autocomplete system to deliver your predicted queries much faster.
  • Autocomplete spelling corrections. [launch codename “Trivial”, project codename “Suggest”] This is an improvement to the spelling corrections used in autocomplete, making those corrections more consistent with the spelling corrections used in search. This launch targets corrections where the spelling change is very small.
  • Better spelling full-page replacement. [launch codenames “Oooni”, “sgap”, project codename “Full-Page Replacement”] When we’re confident in a spelling correction we automatically show results for the corrected query and let you know we’re “Showing results for [cheetah]” (rather than, say, “cheettah”). We made a couple of changes to improve the accuracy of this feature.
  • Better spelling corrections for rare queries. This change improves one of the models that we use to make spelling corrections. The result is more accurate spell corrections for a number of rare queries.
  • Improve detection of recurrent event pages. [launch codename “neseda”] We made several improvements to how we determine the date of a document. As a result, you’ll see fresher, more timely results, particularly for pages discussing recurring events.
  • High-quality sites algorithm improvements. [launch codenames “PPtl” and “Stitch”, project codename “Panda”] In 2011, we launched the Panda algorithm change, targeted at finding more high-quality sites. We improved how Panda interacts with our indexing and ranking systems, making it more integrated into our pipelines. We also released a minor update to refresh the data for Panda.
  • Cross-language refinements. [launch codename Xiangfan] Previously, we only generated related searches based on the display language. With this change, we also attempt to auto-detect the language of the original query to generate related search queries. Now, a user typing a query in French might see French query refinements, even if her language is set to English.
  • English on Google Saudi Arabia. Users in Saudi Arabia can now more easily choose an English interface to search on google.com.sa.
  • Improved scrolling for Image Search. Previously when you scrolled in Image Search, only the image results would move while the top and side menus were pinned in place. We changed the scrolling behavior to make it consistent with our main search results and the other search modes, where scrolling moves the entire page.
  • Improved image search quality. [launch codename “endearo”, project codename “Image Search”] This is a small improvement to our image search ranking algorithm. In particular, this change helps images with high-quality landing pages rank higher in our image search results.
  • More relevant related searches. Sometimes at the bottom of the screen you’ll see a section called “Searches related to” with other queries you may want to try. With this change, we’ve updated the model for generating related searches, resulting in more useful query refinements.
  • Blending of news results. [launch codename “final-destination”, project codename “Universal Search”] We improved our algorithm that decides which queries should show news results, making it more responsive to realtime trends. We also made an adjustment to how we blend news results in Universal Search. Both of these changes help news articles appear in your search results when they are relevant.
  • Automatically disable Google Instant based on computer speed. [project codename “Psychic Search”] Google Instant has long had the ability to automatically turn itself off if you’re on a slow internet connection. Now Instant can also turn itself off if your computer is slow. If Instant gets automatically disabled, we continue to check your computer speed and will re-enable Instant if your performance improves. We’ve also tweaked search preferences so you can always have Instant on or off, or have it change automatically.

And here’s a recap of some other January improvements we’ve already blogged about:


Stay tuned for more in the coming months. We have some exciting ideas for how we can continue innovating on transparency in search, and we hope you enjoy what’s on the way.