Internet Explorer 8 beta 2: Privacy vs Monetizability

image Once upon a time, when I was a young turk, I would assiduously download every last doodad that my employer created as soon as it shipped – or often long before, happily reaching for the pile of floppy disks as I rebuilt my computer for the umpteenth time following the latest toxic combination of untested software.

Age (and a need to still be able to work on my computer) has slowed me down. So I passed over IE8 beta 1, preferring to read about others’ experiences of the new “standards mode” that is the default rendering mode for the new browser.

But last week, only hours after its public availability, I downloaded and installed IE8 beta 2. Why? Because it contains a raft of new features for protecting user privacy. I’ve blogged previously about the eternal tension between user privacy on the web, and the measurement and tracking that is so essential to many websites’ business models. Put simply, if users’ behavior could not be measured online, a lot of online businesses would go out of business.

 

What’s new?

So how does IE8 contribute to the debate? Well, there are a number of minor features to protect users, and one major one. The minor ones include a nice feature in the address bar to highlight the actual domain of the site you’re looking at:

image

This makes it much easier to spot phishing attacks, since many phishing sites try to confuse users by including familiar looking domains as subdomains of the real site, e.g.:

http://www.barclays.co.uk.evildomain.com/…

Another nice feature, related to phishing, is the “Smartsite Filter”. This allows the user to check the current website against a known list of bad sites.It’s essentially a UI into the automatic phishing filter that was built into IE7  – but it allows users to report sites as well as check them, adding a Cloudmark-like element of user contribution to the process of spotting evil sites.

This feature is rolled up under a new Safety menu, which also contains options to view the privacy policy info for a site (which shows all the cookies that were served and/or blocked, per IE7), and the security report for a site (any problems with the site’s SSL certificate etc). Neither of these features is new, but it’s nice to see them called out in their own menu.

The other small enhancement worth noting is that the “browsing history deletion” feature has become smarter – you can elect to delete the cookies etc. for all sites except those in your favorites list. This is a step forward, but it still mystifies me that IE has no easy way for browsing the cookies (and their content) on your computer, and selectively deleting them (as Firefox has had since v2, it pains me to say).

 

InPrivate Browsing & Blocking

The big new security/privacy feature in IE8 is called InPrivate Browsing (others have dubbed it “porn mode“, but I am above such lewdness). InPrivate Browsing allows the user to browse without storing any cookies or browsing history, or locally cached files. It’s good for when you’re borrowing someone else’s computer, or if you share a computer and don’t want the other people who use the computer to know what you’ve been up to (now you are starting to understand where the “porn mode” nickname comes from).

The naming of the InPrivate functionality is somewhat confusing. Once you turn on InPrivate Browsing (either from the Safety menu or using Ctrl+Shift+P), something called InPrivate Blocking is also activated. InPrivate Blocking prevents your browser from sending requests for third-party content that it thinks are principally for the purpose of tracking your behavior. The big difference here is that this isn’t just blocking third-party cookies – it’s third-party content. That’s tracking pixels, third-party JS calls, and yes, ads.

InPrivate Blocking will block third-party requests if one of the two following conditions have been met:

  • The request URL has been made in a third-party context on more than 10 other domains
  • You have specifically added the request URL through an InPrivate Blocking Subscription

To understand the first condition, take a look at the screenshot below, which is the dialog that comes up if you select InPrivate Blocking from the Safety menu when InPrivate Browsing is active:

image

You’ll notice that there are some third-party request URLs that come up, well, a lot. googleadsyndication.com is the domain that Google AdSense ads are served from; and you will doubtless know what comes from google-analytics.com. In the dialog above, the four URLs across these two sites have each been requested at least 20 times in a third-party context, and I’ve only been using IE8 for a few days. With the default settings (“Automatically block”), these URLs are blocked when I am in InPrivate mode.

The other way of adding a URL to the blocked list is to subscribe to an InPrivate Blocking list. This is an RSS or Atom feed of URLs that IE8 should block in InPrivate mode. I have created a subscription list which blocks third-party requests to analytics.live.com – the domain for adCenter Analytics’s tracking JS and pixel. You can try it out by clicking here.

The power of the feed-based approach to InPrivate Blocking is that privacy advocacy sites can post a single link to a feed XML file which users subscribe to; if that file changes, the users’ blocking lists change. So you can expect to find “click here to block ALL tracking pixels and ads” links on such sites in the not-too-distant future. You can take a look at your InPrivate Subscriptions through the Manage Add-ons option in the Tools menu:

image

 

“Aargh! This sucks!”/”Great!” [Delete as applicable]

Whether news of this functionality sends a shiver down your spine or warms the cockles of your heart depends on whether your business depends on online advertising or web analytics. Popular third-party analytics systems like Google Analytics, or third-party ad servers like Atlas Enterprise will lose data on users who enable InPrivate Browsing; and even a less popular service that might not normally be blocked automatically could end up on common “Opt-out” feeds and have its tracking blocked, especially if had a poor reputation for privacy.

I must admit that when I first read of this functionality, I was – ahem – a little apprehensive, for the reasons above. And in truth, only time will tell what proportion of users are engaging InPrivate browsing (although, given the nature of the functionality, we’ll not be gathering this data). But my gut feel is that, whilst this capability is a welcome addition to the privacy and security arsenal of Internet Explorer, actual take-up of the feature will be low. It n
eeds to be invoked explicitly, of course, and the blocking of persistent cookies means that some desirable features of websites (such as being able to remember you from visit to visit) will be disabled. So I imagine it will be used sparingly by the vast majority of users.

Even so, this feature could easily add another 1 – 2% to the existing disparity between different measurement systems (such as an in-house web analytics system and a third-party ad server). Though there are techniques that vendors could use to work around the automatic blocking – the best example being the use of CNAME DNS entries to make the third-party tracking URLs look like first-party URLs – these techniques will add complexity to the implementation of such systems; so it might be easier for us all to live with a little less certainty.

 

If you’d like to read more about the new features in IE8, there’s a ton of stuff over at the IE blog. And, with my Microsoft hat firmly on my head, I should say that the IE team has done an outstanding job with this beta, which is performing really well for me, and rendering most sites flawlessly, with just a few slight layout differences cropping up here and there. Well done, guys.

8 thoughts on “Internet Explorer 8 beta 2: Privacy vs Monetizability

  1. This is a great post Ian. I thoroughly enjoyed reading it and learned a lot about how IE8 works (and used your link to download and install it a few mins back -hopefully your affiliate chq is on its way :).
    If its ok I wanted to add two thoughts.
    1] We have got to realize that the primary thing we are solving for is customer experience and secondarily for collecting data. Website Visitors should have a choice about privacy and I firmly believe that it is that confidence in the web (and web monitoring) that will perhaps one day result in cookies not being a four letter word.
    2] Decision Makers in corporations (usually fanned by external “experts”/ “consultants”/ bloggers(!) ) underestimate the value in web analytics data, even with full privacy settings on. Even with 100% of your website visitors browsing in privacy mode.
    Sure Unique Visitors would be imprecise. Yes you won’t get much from segmenting New vs Returning Visitors, and such. But your visits number is fine. Your referrers are fine. Your keywords are fine. Initial sessions from campaigns are fine (as is conversions at least for those first sessions). Your page views are fine. And …. well I could keep going.
    We should be on a quest to ensure that our data has as little inaccuracy as possible, then use the wealth of insights that are already there rather than bitching about the small % that’s missing.
    It is important to pause and think of faith based initiatives like ads in magazines or on the telly and how they are measured. On its worst day the web is infinitely more measurable.
    Thanks again for a lovely post.
    Avinash.

  2. Avinash,
    A pleasure, as ever, to find you in my comments box. Absolutely agree about the need to provide choice, which is why I think the IE team have pitched it right with this release – the functionality is there, but not enabled by default.
    However, with InPrivate blocking switched on, it’s important to understand that ALL tracking pixel and JS calls to commonly encountered tracking services will be blocked. This includes GA, Webtrends, Atlas, Doublelick, and to a lesser extent Omniture (because most of Omniture’s tracking goes to ‘first-party’ CNAME destinations) – anything that comes up more than 10 times as a third-party call in the user’s browsing history. So you’re not just losing cookies – you’re losing page impression data. Users with InPrivate enabled will disappear in this situation. So in a (theoretical) 100% InPrivate situation, you would see no data at all.
    But I agree that the main response to web analytics & ad server vendors to this development is to encourage customers to focus on drawing valuable conclusions from the data they do have, not fretting about the data they don’t. And another 1-2% is livable with, but I do pity the vendor customer service folk who will have to explain this ‘discrepancy’ again and again to clients.
    Cheers,
    Ian

  3. Ahh….
    Thanks for the clarification. Time to dust up the old web server log files and write a chq to ClickTracks or Urchin to get the log file based analytics solution. 🙂
    Atleast that will give us some of the data I had mentioned in the first comment!
    -Avinash.
    PS: I think our mutual friend John Marshall said it best: “Our solution is less inaccurate than the other guys”. In there is the perfect framing we should give online marketers.

  4. Some of this functionality also appears in Google’s Chrome browser, including domain highlighting and a version of private browsing. The third-party cookie control isn’t as refined in Chrome, though.
    I’d be interested in knowing how fast IE8 is compared to other browsers. The improved speed in Chrome alone makes it worth using. Pages appear to render much more quickly than they do with other browsers, including (oddly) Firefox.

  5. Sean,
    Nice to hear from you (loved your book, by the way). There’s a good TechCrunch article about Chrome performance vs other browsers, here:
    http://www.techcrunch.com/2008/09/04/so-is-chrome-the-fastest-or-what/
    Only caveat I’d point out is that because Chrome launches a new process for every browser tab (which IE does to a slightly lesser extent), its memory footprint is higher than other browsers; so on memory constrained systems this could end up impeding performance.
    Cheers,
    Ian

  6. According to the Chrome comic, memory recycling is more efficient by launching those new processes. Supposedly when a tab is closed, all of the process’ memory is recovered, versus leaving behind garbage in RAM. So they contend that the footprint is actually smaller than other browsers over a long period of use. Plus the stability advantage of isolating the tabs from each other.
    For Firefox fans intrigued by some of Chrome’s functionality, Lifehacker does a pretty good job detailing how you can get Firefox to look and act more like Chrome:
    http://lifehacker.com/5044518/enable-chromes-best-features-in-firefox
    http://lifehacker.com/5046737/chromifox-makes-firefox-look-like-google-chrome
    It will be interesting to see how quickly Chrome is adopted by the public at large. For sure our little “geek” community is going crazy over it. But remember that it’s taken Firefox 5 years to achieve its 10-15% browser share…
    Wow, doesn’t all this browser war stuff take you back? I know I have that Mosaic floppy around here somewhere! 😉
    Cheers!
    Aaron

  7. Hi Ian
    Thanks for reading the book! I’m pleased you enjoyed it!
    Thanks for that techcrunch link – my gut feeling is that Chrome renders faster than other browsers I’ve used even for pages with no Javascript element. The Javascript is clearly faster, but I’m surprised that the plain rendering seems to work better too.
    As Aaron says, it’s funny that there’s a browser war in 2008. It’s interesting too what’s driving it. Last time around everyone thought there was a market in developing and even perhaps selling browsers. This time around, all the browsers are free and it’s about what other web experiences they might enable (out of which, presumably Microsoft, Google and the end users represented by the open source community, will benefit).

  8. Nice post Ian!
    It will be interesting to see how ad networks (especially behavioral networks) will overcome this issue of 3rd party cookie blocking, if the incognito / inprivate browsing becomes more prevalent. Most of these networks use the 3rd party cookie to get a bigger audience reach. However, those behavioral targeting systems that does the tracking at the ISP level will not have this issue, I suppose. Any thoughts?
    Best,
    Siva

Leave a Comment