Why test with Internet Explorer 7?

Jonathan Klein asked a good question in the comments on this recent post:

Is there a reason why you test with IE 7? I understand that it may expose performance issues a little bit more because it has fewer connections per domain and executes JS extremely slowly, but on almost all sites IE 8 is going to be more widely used (and IE 7 market share continues to drop over time).

Here’s how I responded:

I used to test with IE 8, then made the switch to IE 7 after some consideration of the fact that it’s still the third most popular browser version in the world, and a 15% market share is still pretty significant. Put it this way: If you owned a bricks and mortar shop, would you be happy knowing that 3 out of 20 prospective customers had a hard time opening your door? It made sense to me to identify IE7 as a reasonable baseline for performance.

Since I wrote this, this issue has been ricocheting around my brain, and it’s been bouncing off some other things I’ve read. Now I want to expand on my response.

I frequently see two major — and potentially dysfunctional — disconnects between the tech community and non-technical folks. And before I go any further, I’ll confess that I’m as guilty of these disconnects as anyone.*

The disconnect between tech and marketing/sales

First, think about what 15% means. From one perspective, it seems statistically insignificant, not worth bothering over. But from a marketing and sales perspective, 15% is enormous. Those are 15 people out of every 100 people who visit the average ecommerce site — 15 wallet-carrying human beings who, by virtue of being at the site, have demonstrated that they probably want to spend money.

For every dev team that works hard to build and maintain an ecommerce site, there’s a marketing and sales team working just as hard — and spending a lot of money — to bring new prospective customers to the site. They create online promotions, banner ad campaigns, flyers, billboards, print and TV ads. And 15% of that effort and money gets wasted if the site rebuffs IE 7 users.

The disconnect between tech and end users

A recent example from the browser wars:

Last week, Mashable trumpeted the demise of IE with the headline With Less than 50% Market Share, IE Is Now Losing the Browser Wars.

Reality check: At 49.87%, IE still has more market share than Firefox, Chrome and Safari combined. Whatever my personal feelings about IE might be, I don’t call that losing.

This article caused some jubilation within the tech community — the usual schadenfreude that bubbles up any time there’s negative news about a major player. Predictably, commenters on the Mashable post argued the merits of their favourite browsers. One person stated, “The people who hated IE and switch to Firefox, are going to be the people who hate Firefox and switch to Chrome if Mozilla does not get quicker with their release, updates and innovations.”

I’ll come back to that comment in a minute.

Yesterday, I read an excellent blog post by Nicholas Zakas, entitled What’s a web browser? In it, Nicholas describes a conversation he had with his mother, a very well-educated woman who, nonetheless, did not know what a browser is. He explains why he’s sharing this story with us:

As web developers and engineers, we live in a fantasy world where every user understands exactly what he or she is doing on the Internet and has a deep appreciation for web standards. The real world doesn’t reflect the fantasy at all. Users don’t understand the Internet; they don’t understand web browsers. If they don’t understand web browsers, that means they also don’t understand if they’re using a good browser or a bad browser, nor do they understand how to upgrade it.

For many, the Internet is just an icon on their desktop with a particular logo and color just like the toaster is a black or silver box that sits on the kitchen counter. What brand of toaster do you have? Who cares? You know how to make your morning bagel with it, right? For many of the billion Internet users, web browsers are simply a box that displays stuff. Whether or not that stuff has rounded corners and drop shadows is irrelevant. Whether or not you want them to change that box is irrelevant.

On Twitter, one commenter responded to Nicholas’s blog post with this: “If you work as a web dev, your parents should know what a web browser is. Be the change you want to see in the world, yo.”

Here’s where the Mashable comment and the Twitter comment get user experience profoundly wrong:

  • The average internet user does not shop around for the best browser, and only a rare few update with any regularity. Most people accept the default browser that comes with their operating system. And the average person replaces their computer every four or five years.
  • No matter how much you or I might prefer otherwise, most people don’t know what a web browser is, don’t care, and never will. And that’s okay.

Getting back to the original question, this is why I test on Internet Explorer 7, and will continue to do so until its market share actually does become insignificant. This does not mean I won’t or don’t test in other browsers. This is a multi-browser world that we live in and our testing needs to reflect that.

*The irony in all of this is that I have to fight my own tech snobbery as hard as anyone else. I haven’t used IE 7 for 3+ years for anything other than testing. I’m a Chrome user who upgrades to the dev builds daily. I’ve converted my parents to Chrome and scoff when my wife opens IE 8 to browse on her computer. I need to remind myself daily that I am not the norm.

Related posts:

9 thoughts on “Why test with Internet Explorer 7?

  1. Pingback: Twitted by ti_pierre

  2. Nice post, I think you’ve converted me. There is a strong argument to say that if you optimize for IE 7 and make that browser perform well then more modern browsers will follow suit. I think one exception would be domain sharding – when you are trying to decide how many domains to stripe your static content across it usually makes sense to pick a number that isn’t ideal for IE 7 but works well for your user base as a whole. Unless of course you can do it dynamically based on user agent (which I believe is a Strangeloop feature).

    Thanks again for the further explanation.

    -Jonathan

  3. I have to agree with the author. It is important to test all browsers in order to reach the highest number of potential customers or leads.

  4. Pingback: Early findings: Internet Explorer 9 will not make your site faster by default — Web Performance Today

  5. Pingback: End-of-year performance report: Top retail sites are slower, not faster, than the rest of the pack — Web Performance Today

  6. Pingback: Is it time to give up Internet Explorer 7 as our default test scenario? — Web Performance Today

  7. There’s really no need to test in any browser other than IE7. Pretty much all other browsers are standards compliant so if it looks good in IE7 it’s going to look good across the board. The only browser I’ve ever had trouble with is IE. Microsoft thought they were big enough to mold the internet into their image, ignoring web standards in favor of their way of doing things. Even now IE9 is only partly compliant which makes it the most standards compliant browser Microsoft has ever released. I think it’s worthy to celebrate every percent of lost users they suffer. When or if IE is ever gone from the world, it will make web development exponentially easier. But until then, and despite my urge to ignore the ancient browsers people insist on using, I guess testing in IE7 is a sound practice.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>