From the category archives:

Search Engine Optimization

Clarifying a couple points

by Matt Cutts on March 13, 2010

[Just as a reminder: everything below is my personal opinion. I haven't sent it to anyone else at Google for a review, etc.]

Valleywag used a recent podcast I did as material for two points in Six Delusions of Google’s Arrogant Leaders. The two assertions that used my comments as material were “Google’s wealth means Google ‘gets it’” and “Google must sacrifice user privacy to grow.”

Valleywag has either misinterpreted what I said, or I didn’t express myself clearly, because I don’t believe either of those claims. I’ll try to explain the intent of what I said, in case I wasn’t clear during the podcast. I’ll address the latter claim first (“Google must sacrifice user privacy to grow”), because I certainly don’t believe that “Google must sacrifice user privacy to grow.” I think Google benefits the most when users understand what Google is doing and why; I also think that user trust in Google (and by extension our privacy policies) is paramount to our success.

A good example is our Google Ad Preferences page. As one blog concluded a couple days ago: “Google’s Ad Preference Manager, with its persistent opt-out plug-in, offers precisely the kind of robust opt-out that privacy advocates have always demanded.” And it’s not that we’re shy about talking about privacy; Googlers Alma Whitten and Nicole Wong recently talked privacy for an Ars Technica article that came out earlier this week. It’s a long article, but an example useful fact is that if X is the number of people who visit the Ad Preferences page and opt out, 10X people don’t opt out and 4X people actually edit their categories to improve the targeting relevance of the ads they see. Let me say that again: four times as many people change their settings to make their ads *more* relevant than opt out of interest-based targeting. I think the Ad Preferences page is a good example where users get more transparency and control regarding their privacy.

Another example where Google helps your privacy (rather than sacrificing it) is the Google Dashboard. This is a single site that gives you an overview of what information Google has from various services, and allows you to edit and to manage settings. This is another example where Google is trying to give more information to users, not less. I could point out lots of examples where we try to debunk privacy misconceptions. Where we actively fight for our users’ privacy. Or where we talk about privacy and engage in debates about user privacy. And of course there’s Google’s full privacy center (with videos!) at http://www.google.com/privacy.html . Suffice it to say, I don’t believe that Google “must sacrifice user privacy to grow.”

Okay, what about the other claim that Valleywag used me for: “Google’s wealth means Google ‘gets it’”? Ryan Tate wrote “It’s a truly bizarre moment, in which Cutts defends some horrendous management decisions based on Wall Street trades.” I don’t agree with that either, so let me try to clarify. Eric Schmidt joined the company in 2001. The first time I got to meet Eric was at the weekly TGIF meeting where he was introduced to the wider company. He answered questions for an hour, and I thought his answers were spot on. He was one of the original authors of lex, a well-known Unix utility that I had used in the past, so I knew that he was also a solid engineer and technologist. Schmidt also had experience at large companies (Sun and Novell).

All in all, I was very happy and impressed that Eric was joining Google. When I went home that day, my wife asked what had happened at work. And I replied with something like “I think the value of our stock options just went up a lot.” What I meant by that was that I thought Google had recruited the perfect person to lead the company from start-up to the next level. I still believe that. Eric has been a truly great CEO–and I’m not just saying that because for the last several years he has worked for $1 a year. :) Maybe I didn’t tell the anecdote well or clearly, but my intent was to explain that I think Eric Schmidt has been a great CEO right from the beginning of this decade, not to defend any decisions “based on Wall Street trades.”

If you want to listen to the full podcast, it’s available, but I hope this post helps to clarify.



{ Comments on this entry are closed }

Chrome support for Greasemonkey

by Matt Cutts on February 9, 2010

Back in December, I happened to click on a Greasemonkey script in Chrome and was shocked that it just worked. At the time, I wrote a note within Google that said

Whoa. I just clicked on a Greasemonkey script in the latest dev version of Chrome (4.0.266.0 on Linux). Chrome offered to install the GM script, so I said okay. The script ran perfectly in Chrome with no changes at all! I don’t know how many Greasemonkey scripts will run in Chrome unchanged, but at least some will.

Last week brought that news as an official announcement. My guess is that scripts that don’t use specific Greasemonkey APIs should be fine.

(Side-note: I found a good post from November that claims that ~60% of Greasemonkey scripts don’t use any sort of special API calls at all. The top API calls appear to be GM_getValue and GM_setValue (16.5% of Greasemonkey scripts), plus GM_xmlhttpRequest (15.5% of Greasemonkey scripts). It’s unclear which of these functions might be worth supporting. Some could have security implications (GM_xmlhttpRequest). Others like the get/setValue functions could be done by using other ways to store data.)

So this is cool. There’s a good chance that your favorite Greasemonkey script might just work in Chrome. Personally, I recommend the dev channel version of Chrome. It gets all the cool features early, and it’s been very stable/fast for me.



{ Comments on this entry are closed }

Keep an eye on changing pages

by admin on January 25, 2010

Google just launched a nice feature on Google Reader: the ability to keep an eye on pages for changes. This works even if the page doesn’t have its own RSS feed. This sort of thing is very handy. You could use it to spot new things on a privacy policy page or watch for changes in the executives page at another search engine.

Check out the blog post, but it’s easy to use: just add any url to Google Reader.



{ Comments on this entry are closed }

Important blog post on Google blog

by Matt Cutts on January 12, 2010

This is an important blog post. Go read it from the source.



{ Comments on this entry are closed }

A short note about real-time search

by Matt Cutts on January 8, 2010

This is just a quick blog post to share some info about Google’s real-time search based on the earthquake that shook the Bay Area this morning. I didn’t feel the earthquake myself, but the real-time folks looked at how we did. Here’s what they found:

Overall, realtime search triggered in under two minutes from the earthquake happening and within a minute of the first tweets appearing. The rough timeline (in Pacific time) is

~10:10 – An earthquake happened. (The USGS says the earthquake happened at 10:09:35 a.m.)
10:11 – The USGS government web site started to track the earthquake, with a “?” magnitude.
10:12 – Google’s realtime onebox triggers.
10:13 – USGS web site marked the magnitude as 4.1.
10:20 – USGS site updates their feed.
~10:25 – Google’s earthquake onebox gets updated earthquake info.

All in all, not too shabby, but still a perfect opportunity to identify ways to do better. For example, Stephen Shankland said it took about six minutes for him to see realtime results trigger, adding “which struck me as fairly impressive for such a mammoth operation as Google search.” We looked into it, and it looks like a data center timeout meant that a few people didn’t see the realtime results for a few extra minutes. Nothing to be ashamed of, but something we’ll look at improving in the future. Most searchers did see the realtime results trigger within a minute or two on Google.

Greg Sterling showed a screenshot (taken by Danny Sullivan) over on Search Engine Land that showed an earthquake onebox with stale info. It looks like that happened because the feed we were using from the USGS took ~10 minutes to update. Going forward, I’m sure we’ll check whether we could do anything differently on this, but if a feed has older info for a few minutes, there’s little we can do about that.

Overall though, I think Google acquitted itself quite well on this earthquake. Realtime search triggered quickly and accurately for the vast majority of people. And Google’s realtime search was able to pull in not only tweets, but relevant blogs and news articles. The search quality team will keep working to make sure that the triggering speed/thresholds, relevance, freshness and comprehensiveness of real-time search matches the high expectations that people already bring to Google’s web search results.

One last thing: I love that many Google employees’ first instinct when they feel an earthquake is to start searching on Google to see how well our real-time search works. :)



{ Comments on this entry are closed }