Not trapping users’ data = GOOD

When users get what they want from you quickly and easily, they’re more likely to come back next time. (Shh. Don’t tell anyone else this vital secret.) Part of that is feeling that they aren’t “trapped”–that they can leave you behind if they want.

That’s why I was happy about Eric Schmidt’s quote from the Web 2.0 conference last year. He spoke against trapping users’ data so that they can’t get to it. Dan Farber covered the interview:

“We build a very good targeting engine and a lot of business success has come from that. We run the company around the users–so as long as we are respecting the rights of end users and make sure we don’t do anything against their interest, we are fine,” Schmidt said. He noted that history has shown that the downfall of companies can be doing things for their own self interest. “We would never trap user data,” he said.

Schmidt was asked if users could get all of their search history and export it to Yahoo. “We would like to do that, as long as it is authenticated….If users can switch it keeps us honest.”

(emphasis mine) I love that Eric said this. It echoes the “send your users away happy and they’ll come back” philosophy. It also gives guidance to teams at Google.

So I started making a list of the ways that Google lets you access your data:

– Gmail. This one’s easy. Google provides free POP access so that anyone can fetch their email out of Gmail.

– Search. If you sign in with your Google account to search, Google can offer not only personalized search but also let you retrieve your search history. Mihai Parparita did some digging a while ago, for example. The ability to securely access your search history as an RSS feed is documented in our help pages now. For example, the url works very well if you’re logged into your Google account. You’ll get an RSS feed back like this:

Search history feed

I believe you can add things like “&num=250” so that you don’t have to access 10 items at a time either. This feature is secured by password-protection (you have to be logged in), but it provides a nice way to access your own searches. Oh, and don’t forget to try out your personal search trends. If you’re logged in, the url is and you’ll get all sorts of neat data like your most frequent searches, clicks, and when you tend to search:

Search history trends

Okay, enough about search. Let’s look at some other products that let you get to your data easily.

– Google Docs and Spreadsheets let you export your stuff in more formats than I know: Word, Rich Text Format, CSV, HTML, XLS (Excel), and PDF. Even one I didn’t know: .ods? Ah, OpenOffice. Nice. 🙂

– Google Calendar. Also easy. From its launch, Google calendar has allowed iCal (.ics) and RSS export of calendar data.

– Google Talk uses the open XMPP protocol. The VoIP part of Google Talk is done with Jingle, another open protocol that Google helped with. I like that our IM chat is open to other clients, so you can talk from iChat and GAIM to Trillian Pro and Blackberries.

– Google Reader easily exports your list of feeds in OPML format, and can import OPML files as well.

– Blogger. Blogger can export data via FTP or SFTP and backup your blog.

– Google AdWords. I don’t use AdWords myself, but Google provides a free application called the AdWords Editor, and its features include a snapshot export feature: “Save a delimited file with your AdWords account information and show it to a colleague or keep it for reference.” So I’m assuming it’s not too hard to suck down your AdWords info. Yup, a couple minutes of searching found references to importing your Google ad campaigns into Microsoft and Yahoo.

– Google Groups. I was dreading checking on this one. Back in August, someone wrote to me and said “I run a Google Group with 7,500+ subscribers and I need to download the subscriber list, but I don’t see an option for that.” It turns out that we didn’t offer that as a feature back in August. We were able to help the fellow, but it didn’t sound like an often-requested feature, so I didn’t think the Groups team had gotten a chance to do this. But I checked and it looks like the Groups team got a chance to add this. Yay! For a group I owned, I clicked “Manage” and then “Browse membership list.” At the bottom right will be a button “Export member list” and clicking that will download a comma-separated value (CSV) file.

– Let’s see, where else can you store data at Google? Ah, a Custom Search Engine. There’s even a bookmarklet to let you add sites to your custom search engine as you surf the web. Can you get your entire list of sites exported from your Custom Search Engine? Yup. Go to your search engine’s control panel and click on the “Advanced” tab. You’ll get options to download your sites in XML or tab-separated value (TSV) file format.

– Lots of products like Google Analytics and the Google Webmaster Console also give options to export data in various formats.

Okay, so looking down this list, it looks like Google does pretty well in offering open access to your data, at least for all the important services that I checked. If you know of some way that Google doesn’t let you download your data, please feel free to mention it. I like that Eric said this, because it’s a really nice precedent to set.

66 Responses to Not trapping users’ data = GOOD (Leave a comment)

  1. Great tips Matt, thanks.

    That’s one goal I have too for my sites. Great usability and easy access to information. I would never want a visitor to have to hunt and search for something because they would, like you say, be less likely to come back.

    That’s why I keep coming back to Google to search for things: it’s easy 🙂

  2. This is definitely a good thing.

    It would be nice if I could export my chat history somehow, too (afaik that’s not exported with gmail, but I haven’t investigated). Better yet, I’d like to be able to /import/ the chat histories I have from other IM clients (gaim, adium) to google so I can search it all in the same place.

  3. One thing which is missing is an ability to download the Google Talk chat archives from my Gmail account.

  4. I really like hearing about Google’s openness, it’s very reassuring to know that I can take my data with me if I choose to leave.

    However, I can’t figure out how to take saved chats with me from Gmail if I migrate to another service, or even how to just download them.

  5. Yes, a very good summary there Matt!

    It would be interesting to see the same concept applied to the larger players in the social networking scene. 🙂

  6. Understandably, the personal user data will affect a user’s Organic SERPs with that user logs in –

    but does any of this collective information ever influence the General Organic Algos that the public sees.

  7. Hmm, looks like I need to actually start building up some *interesting* information to export! 🙂 My graphs of search activity are all predictably boring, my google docs are the same and let’s not even start on my Calendar! Ha ha ha 🙂

    Great tips Matt, thanks as always.


  8. Totally agree Mark, and I believe that the themes you highlight of Open Standards & Access are some of the main reasons I select Google products over the competition (Quality being the obvious other reason).

    However, I do not think that Google makes access to this data particularly easy, streamlined or consistent.

    So, for me it is not that Google doesn’t let me access my data…but my main feedback would be to make it easier and more consistent. For example, a single point of backup of raw files using FTP or HTTP for transfer.

  9. Amazing isn’t it? Thats why I love Google. Every company has its bad sides, but Google have proven in many aspects loyalty to customers.

  10. The personal search trends feature is quite shocking, with number of searches I do every day, its hard to believe I get any work done at all!

  11. Google is *reasonably* open. I’ll give you some rough edges, mostly around the use of AdWords.

    AdWords, with some (undocumented) twiddles, offers a unique advert ID; where referrer_info isn’t offered (e.g. bots, some browsers), it’d be nice for an advertiser to find out the site & page on which an advert was offered. This helps advertisers control distribution.

    AdWords offers some macros – “{keyword}” expands into the keyword. There are “missing macros”. The lack of these mostly acts to drive up the number of API operations, and hence cost of using AdWords. Some of this could be evaded by offering more data against that unique advert ID, but that would require extensive changes to analytics packages. Adding some maros would allow analytics and bidding systems to do more. Examples that would be useful include “match-type”, “position”, “cost”, “publisher” – you can probably infer the meanings… I’ve repeatedly asked account reps for more macros, but just get a truncated list of the macros that are known outside the Googleplex. There may be other macros implemented, but left undocumented. Finding the currently known ones involved locating a misnamed 2003 PDF…

    The Search API – was usefully REST, albeit with nearly uselessly low limit of searches, and is now targeted to be an AJAX component. AFAICS there is now no way to use the Search API to gather data on, for example, competitor messaging and position, except by getting a person to sit in front of screen (T’s and C’s appear to prohibit automation – I last reviewed this about a year ago, could be different now). A non-impression counting search tool offers some problems (could be presented on another interface) – but the current rules also seem to prohibit any automated uses for SEM, even as innocuous as seeing what positions you appear in and what your own snippet consists of (helpful to control that NOODP, metatags, etc).

    There is an AdWords Advertisment test interface – lets you see how your advert appears in context for a specific (geographical or IP) location. Doesn’t count the impressions. Clicking the advert does nothing that causes a payment, also doesn’t take you to the advertiser. Fabulous idea. T’s and C’s appear to prohibit automation. I can think of a number of ways of using this for large and small advertisers, but most of them are barred by the prohibition.

    The Webmaster Tools apparenty prevents a download exceeding 4MB. Hence lage sites with many inbound links to the home page can’t see all the links – and since the links are sorted, you can only see, say, sites in the first half of the alphabet…

    Although GMail offers a POP feed, it apparently rate limits. The rate is lower than or roughly equal to my rate of receiving spam. So my mail system never catches up… Downloads ten mails while eleven arrive… I’ve given up on trying to synch, since, when I’m on the road, emails definitely outstrip my zero download rate 🙂

    There’s a difference between data being available, and being *usefully* available. Some of it probably involves cost… Making Gmail POP available is a business cost and ducks the payback of advert exposure…

    But, yeah… you guys do a pretty good job. YUI’s nice, though… There’s others out there with a place in the sun, too.

    Cheers, JeremyC.

  12. Talking AdWords!

    Especially for large AdWords accounts, I wish to see an advanced tool assisting advertisers in analysing conversions based on parameters specified by the advertisers themselves. Because of lack of such tool advertisers are mostly forced at present to use a third party campaign execusion system.

  13. Matt.

    Is it me or does the fact that Google records my every search and when I search and what I search, well just a little worrying?

    Given that you know what I search for, would it be right to say personal targeted ads are on the horizon?

    P.S. Am I right in thinking you only sleep 3 hours every night?

  14. The only area I’ve become a little frustrated with when it comes to freedom of data and google is the personalized homepage.

    I’ve got somewhere around 100/150+ feeds across multiple tabs and I’ve yet to find an option to export these as an opml or similar.

    A 3rd party developer has come up with a widget you can add to a tab to export to opml, but it would be nice to have the option of exporting your feed addresses as an opml in the settings rather than relying on a 3rd party widget 🙂

  15. “If you know of some way that Google doesn’t let you download your data, please feel free to mention it.”

    I’m so glad you asked – actually, I just ran across this very thing. Or at least, if there’s a way to get it, I can’t find it. Google Base (for listings I want to appear in Froogle). I’ve a couple of clients for whom I upload extensive product listings, and you can log in to your Google Base account and see things like impressions, pageviews and clicks – BUT THERE’S NO WAY TO DOWNLOAD THAT DATA! I want to be able to show it to my clients, but with roughly 700 listings for one client and 1100 for the other, screen shots really aren’t practical. Plus I can’t sort it so that all the ones with activity come out on top.

  16. I agree that Google is rather open in these regards and allows you to export a lot. One thing to remember though is that as soon as Google products cross-integrate — e.g. a link from Gmail to add an event to Google Calendar — the costs for users of switching away are increased for any single product. As a practical example, let’s say I love Gmail and I hate Google Calendar, so I want to move to competitor Acme Calendar. Great, you guys offer exporting functionality for my events, so I’ll quickly move them from Acme. But you guys don’t allow me to set my preferred Gmail calendar integration software… so now I end up with a somewhat broken Gmail feature. This is not at all alarming on this scale, but it can be a problem for users down the road when Google heavily increases cross-integration (Google Checkout is being pushed in search result today, for example, cross-integrating another two theoretically “loosely coupled” services).

    I’m not implying that a company has or hasn’t got the right to do any of this, by the way, which would be yet another discussion. But in the end, any company won’t “trap data” for the sheer fun of it, but because they want to create a lock-in situation for their users to increase the costs of switching to competing products. So we need to look at the end result of whether or not the costs of switching are really “one click.” In a couple of years, Acme Calendar may accuse Google of doing the same that companies like Microsoft did with their past strategies of cross-integration (e.g. browser integrated into OS), unless Google goes all the way and supports “cross-integration options” for its services. Maybe it’s a question for the future: on the traditional OS, we have file-type associations. But how can we as users switch file-type association on the web OS Google and others are creating? Where can I toggle the “open X in Y”, or will Google always reserve the right to make that decision for me?

  17. Is it not making us all too “dependent” on google?

  18. The only thing I will call you out on is this, Matt. In Google Talk:

    Voice calls to other Google Talk users
    and a big red row of X’s.

    Honestly, I would not mind exclusively using Google’s client for Google Talk – but I can’t.

    There’s no Mac version. At all.

  19. I think it’s fantastic that we can export all of our data elsewhere, and I don’t think it’s bad that Google capture this data in the first place. Two things I think should be added:

    1. Be open and honest – make it plain exactly what data is being/will be kept, how long it will be kept for, etc.
    2. Allow people to modify/remove their own data at their will

    Both of these are based on the principle that we should own our own data.

  20. Matt,

    Not a single program from Google restricts the data download; everywhere we can download the data for free, because, I am using most of the products from Google.

    Thanks for the huge list 😉


  21. How about SketchUp? I can import Autocad files and other formats, but only save in its propietary format. I think that makes SketchUp a LOT less useful.

  22. Very cool Matt.

    I didn’t know I could access my personal search history in such detail. I checked out my history and there was nothing in it. It seems it didn’t start logging my history until after my first visit to it.

    Another export I found that you didn’t mention is Google Checkout Payouts. I can export that data via CSV.

    One thing I wanted to do when I used to use Google Base a lot was to export my items Impression/Clicks/Page View data but I wan’t able to. It seems this is still the case.

    Keep up the good work Google!

  23. Cool, Matt. When are you guys going to offer a delete button?

    I can delete my user accounts and all personal data almost everywhere else.

  24. One thing that Google is notably lacking is the ability to backup your data en masse – download all of your documents or all of your email.

  25. How can I access the data Google collects on my searches while I am *not* signed in? That is, the “secret” browsing history that Google turns over to law enforcement authorities on request?

  26. Andrew, I’d agree that anything we can do to make access easier (not just possible) is a good thing. We do pretty well, but we should keep looking for ways to be better.

    Kun, I try to get enough sleep. You can always clear your cookies; then the next time that you do a search, you’re in much better shape if privacy is a concern for you. I’d argue that AdWords being targeted to the phrase you type is the strongest type of targeting. I may be a TiVo fan, but if I’m typing about domains names, it’s better to see ads from registrars and hosting companies instead of ads for TiVos.

    JeremyC, I agree that YUI is pretty neat stuff.

    Pratheep, happy to help. 🙂

  27. Though this begs the question: can users delete their data? Without that ability, Google owns the data, not the end user. That, to me, is frightening. As a consumer, I am technically never allowed to ‘leave’ Google as my data would linger forever.

  28. Hi Matt,
    Coincidentally I’ve been struggling with a Google service’s lack of open data access very recently; that of Google Video. There’s no way to access the thumbnail or pass variables that change the resolution… there’s very little one can do with it other than paste in the object code to your site. Same with Youtube from what I’ve seen. The APIs on those kinda suck (or seem very limiting). It’s driving me nuts and means I’m going to have to develop some other way to manipulate that data… ho hum.

    I understand that there are all kinds of security, privacy and copyright concerns with a lot of what Google’s services do and think you guys do an amazing job in any case.

  29. I’m really impressed to hear that it’s Google’s priority to empower its users by giving them the freedom to use your services however we wish.

    People hate being locked down. Lock us down and we’ll tolerate, not enjoy, a product. Once we bite the bullet we move on and rid ourselves of the ball and chain, bridge burned forever.

    If you love us, let us go. If we love you we’ll return. 🙂

    p.s., the Miranda IM client is the best instant messenger of all time.

  30. Google Scholar + Trapped data = Bad Google User Experience

  31. I have one that is very frustrating –

    Google Finance Portfolio information.

    I would love to either be able to 1) export it somehow or even better 2) call my personal portfolio information through an API. If either of these options already exist, I apologize because I haven’t found them yet.

  32. I’ve been wishing for an API into Google Notebook for some time now… heck, even an RSS feed from it would be handy.

    The data isn’t trapped, though… you can export to Docs & Spreadsheets. That’s just not as handy as an API or feeds.


  33. I agree with Andrew that access to your data should be made easier.

    In the case of Gmail, POP access is nice and it is a necessary alternative for now, but it has several drawbacks and certainly does not pass the “mom” test.
    I would like to see a simple “Backup my Gmail to disk” option
    that would take care of downloading all emails (All Mail, Chats, Sent,
    Drafts, etc..) and all Contacts to an open standard format (readable by
    most email clients, cross-platform, etc..) and ideally that would
    preserve as much as possible of the metadata associated to the emails
    (labels, stars, etc…).

  34. How about not requiring me to use personalized search if I want to save my search history. There’s no logical reason I shouldn’t be able to do that.

  35. Brent, thanks for mentioning that.

    graywolf, Marissa has talked about that; the change was made to simplify the mental model of what people need to think about. Other than you, I haven’t seen that many people ask for search-history-but-not-personalized-search-when-I’m-logged-in-because-I-don’t-want-to-log-out-to-see-vanilla-search.

    corey, I believe that if you delete your Gmail account, it deletes that email, but I’m not an expert on that particular issue. Certainly you can do things like clear cookies between search sessions if you’re privacy-sensitive.

    Juan Around the World and Hagrin, I’ll try to ask someone about that. Michael Martinez, how does Google Scholar trap data? I’ve played with it, but haven’t noticed that.

    Tyler Martin, I’ll ask about that. It seems like offering those sort of options would be a good thing, but it may be that the teams have had higher-priority things that they’ve been tackling.

  36. Re: gmail, you missed a feature.

    Iffing I decided to stop using my gmail and start using Hotmail, not only is it easy to use pop to pull all of my existing mail out of my gmail account – I can also set gmail to forward the mail onward to hotmail, ensuring that anyone I forgot to tell about my new address (or who forgot to update their address book, or…) can still reach me.

    There really is nothing to stop me leaving – except the user interface and features that beat anything else that’s out there.

  37. Google does not have any real content, pretty amazing, brilliant! 😉

  38. Nice posting, good info 🙂
    What about exporting data from Google Code Project Hosting?

    There does not seem to be an export option for the wiki pages, issues tracked and the source itself. The repository checkout for the source does not count, as you do not get the repository database with the full history itself that way.

  39. Matt,

    I regularly read your blog.. and the contents are great – but, I have to say, YOUR BLOG IS THE FUGLIEST BLOG I HAVE EVER SEEN 😛

    You keep on posting the latest attractive graphs and aesthetically pleasing new features from Google – makes me think the fugliness of this blog is a deliberate ploy to highlight the sheer beauty of the google features you talk about..

    Seeing those screenshots is a bit like seeing an oasis after 10 days lost in the desert… (why is it that the phrase ‘artistic engineer’ is an oxymoron?)


  40. All you really need to do to modernize this blog is add a dos-prompt 😀

  41. dockarl: it’s not Matt’s design. It came from here:

    I tried the num=250 querystring and it didn’t work for me either when I was logged in or logged out. The max I could get was 100 either way (which is the same as before), on both the and DCs. Not sure if that’s a bug or it just hasn’t been rolled out across all DCs or maybe it’s in beta.

  42. Thanks for setting me straight Multi-Worded Adam, but it’s too late.. Matt’s blog has created a new word – FUNFUGLY.


  43. Great stuff, I learn so much here.

  44. Froogle shopping and wish lists. I believe mine was deleted by Google on accident. BTW, there is no Froogle support group for users and I can’t find the feedback form for Froogle.

  45. This is something I’ll be submitting as a feature request to the Google Apps folks, but not trapping of data is or particular concern to me. I’m a member of the executive board of a local (Toronto, Canada) tennis club, and I’d like to use Google Apps (or even just Google Documents & Spreadsheets) but because we have private member information, we need to ensure that the data is only stored on servers in Canada. Is Google working on something like that, so that people who are under different regulatory domains can ensure that their data doesn’t fall into the wrong hands (e.g., the US has much looser privacy laws than Canada does)?

  46. It seems to me that the only people who would ask for search-history-but-not-personalized-search-when-I’m-logged-in-because-I-don’t-want-to-log-out-to-see-vanilla-search would be the people involved in SEO.

    It wouldn’t make sense for a normal user not to want that.

    Although, given the amount of technical users who use Google vs other search engines, I still think it might be a useful option. Heck, It’d be interesting for me to see a side by side mashup of normal search vs personalized search.

    btw, no match for search-history-but-not-personalized-search-when-I’m-logged-in-because-I-don’ come on somebody… you know you want to register it.

  47. Thank you Google for getting it right and making it financially sound, now others must take heed. I recently read an article discouraging “walled gardens.” This is when publishers try to restrict users interaction with things outside the garden. An example is early AOL when users couldn’t email user outside of AOL. One of the worst examples is cell phone companies now “allowing” you to communicate with specific groups at a discount. Well thank you big telco for putting up artificial boundaries and then lowering them just for me. I’m sure I’ll thank you from a VOIP phone in a few years… How about Sony? BetaMax; Failed. Mini-Disk; Failed. MemoryStick; Failed.

    Anyone else what to be proprietary?

  48. Great insights, Matt. 2 thought streams…When I first saw the title of this post in my Google Personalized Home (which I use instead of Google Reader), I actually was curious as to another website debate re: trapping user’s data — that is, asking them to login or register in exchange for something/anything. That could be a comment, a star rating, a written review, having the right to read an entire review or article, etc. etc.

    ANYWAY, I digress. I’m glad I kept reading. One cross-service feature that I hope will one day be integrated for me is that of AdSense data with Analytics data. Having to build channels in Adsense to capture general revenue information for a l’il ol’ content site can get exhaustive. Today we export data from both services into Excel and combine them into one report but…alas, you’ve spoiled me and I want more. 🙂 My uber-geeky friends wag their tongues and say this is intentional ambiguity adhered to by Google product managers so people can’t game the system. That, or two product groups just can’t get along and the inner workings of the company have gotten too fat to make this integration happen quickly.

    My reaction? I’m not thinking that deviously. I’m only thinking like an end user of both services, nothing more. Which means I’m just trying to save time every month in VLOOKUP hell and hope that Google’s inter-service compatibility will continue to improve.

  49. Surely it would not be too hard for Google to build in a feature that allows you to have your personalised homepage without having personalised search activated.

  50. It would be nice to see this spirit extended in supporting OpenID. Support not just in the sense of just being an OpenID provider like some other sites have been doing. But accepting OpenIDs created at other sites as well.

  51. Matt wrote: “Juan Around the World and Hagrin, I’ll try to ask someone about that. Michael Martinez, how does Google Scholar trap data? I’ve played with it, but haven’t noticed that.”

    Citations from documents I cannot get to. Those “we found this citation but won’t show you the document” listings just drive me nuts. I’m not even talking about ACM.ORG listings.

    Here is an example (results may change so I’ll quote the listing below the link) for link spam

    On the front page of results, the last listing I see as I type this reads:
    [CITATION] Link spam alliances
    H García-Molina, Z Gyöngyi – 2005 – March 2, 2005. Technical report.

  52. Oops. Didn’t realize your blog software would hang on a partial HTML reference:

    CITATION] Link spam alliances
    H García-Molina, Z Gyöngyi – 2005 – March 2, 2005. Technical report. < http://blog. searchenginewatch …
    Cited by 2 – Related Articles – Web Search

    Anyway, I concluded by saying that I didn’t see the value in these types of listings to anyone.

  53. The two areas where my data seems trapped is the bookmarks and an opml export of the feeds on my google personalized home page.

  54. Re: Google Groups. The real data is in the messages of a group. Having the ability to export a membership list is nice. But only if I could export all the messages as mbox and take it to some other service, it wouldn’t be a lock-in.

  55. Martin, interesting–I’ll try to ask about that.

    Ryan, it’s definitely a pretty rare request (and much more likely to be from an SEO).

    dockarl, I go for the lo-fi text-heavy version of a blog because, well, for one thing I’m lazy. Stands up better to diggs too, I guess. I expect most people to read me from their feed reader, anyway.

  56. Hmm. While I strongly agree with the premise, and also believe Google is good at providing access to data, Google is not great. One example is Gmail. For example, I can’t export my GMail. I also cannot tell Gmail to add certain messages back into the POP3 queue to be downloaded again (this has caused me no end of grief!)

  57. I don’t think the side-by-side mashup or vanilla SERPs vs. personalized SERPs is going to be that big a deal, since there’s a very simple solution:

    Browser 1 (e.g. IE, just for illustrative purposes): shows personalized results. This would be the browser the SEO prefers to use by nature.

    Browser 2 (e.g. FF, also just for illustrative purposes): shows the vanilla results. This would be the browser the SEO doesn’t prefer to use, or downloaded for this specific purpose.

    Voila. Side by side results. I may be missing something, but it doesn’t seem all that complicated to me.)

  58. I’m delighted to read this. So when will Picasa web albums allow export of all images with XMP sidecars for the metadata in a single action, allowing migration to another vendor?

  59. Does anyone know if there is anyway to export Google Checkout orders (from sales that I have completed). I know that you can export the payouts (althought i have experienced much difficulty with this feature) but I want to export the orders so that I can use them for accounting purposes (Quickbooks format would be nice). It doesn’t seem that anything official has been released, but does anyone know of a hack that I could use to easily import this info?


  60. Matt,

    I love hearing that Google is willing to let users leave–decisions like that will ultimately pay off in the end by keeping you on your toes.

    I am disappointed to see that the “google start page” has no way to export either an OPML file or RSS/feed.
    ( in my case)

    Just an oversight?


  61. It would be nice to see this spirit extended in supporting OpenID. Support not just in the sense of just being an OpenID provider like some other sites have been doing.

  62. Google is taking over the internets.

  63. I guess Marshall over at ReadWriteWeb should have been subscribed to the Matt Cutts blog before he made his snippy comments recently!

    Bet it’s nice to feel ahead Matt!?

  64. Thanks Matt; saw this when digging about Google/Facebook portability. Tweeted about it, too. Nice to see Google leading the way (so when are you going to *accept* OpenIDs?)

  65. This post is very much about “What if user could really OWN all of his data?” mentioned at TheNextWeb’2008 (

    I think we can say “users’ data is not trapped” when you can switch from one provider to another (and back) without effort, virtually with a single click of the button.

    More specifically, it implies two conditions for a given service X:
    1) user having account at X can _easily_ export all his data and settings to any of X’s major competitors, web-based or otherwise, without loosing any metadata that is supported by competitor as well (for example, for GMail it is: [a] to download mail archives, stars and labels, filters and address book in a format Outlook / Thunderbird can import from; [b] to migrate all those data to other popular web-based email, defining industry standards for migrating user between webmail providers)

    2) user who once decided to leave X for its competitor (and who took his data from X) can come back to X, bringing his data updated/expanded by the competitor, and be sure all his X-specific settings and metadata will be preserved, and all data updates from competitor will be migrated to X seamlessly

    Considering things this way,
    = POP access to GMail doesn’t allow to migrate “state of inbox”: un/read state of each message, labels and stars attached etc
    = IMAP does well with un/read, but fails to be “migratable” when you use multiple labels per message: at alternative provider, you’ll always have such messages duplicated, one copy per label attached (which is a loss of vital metadata)
    = in GTalk, user’s data is stuck at GMail: you just can’t download chat history in, let say, MirandaIM-importable format
    = Google Groups: bit complicated. At least, a closed-group owner should be able to migrate the whole group, history of discussion, file an document archives, plus membership data to an entirely different platform (like Yahoo! Groups)–with minimal to none effort from a regular member of the group.

    And yes, being able to import email history to GMail from my MS Outlook .pst-files or chat history from MirandaIM to GTalk is also a very good thing, even if it about “further trapped by Google” 🙂

  66. People who want to download “Google/gmail chat history” can have a look at “How to do that”, without use of any software