Don’t Care About Market Share

Published 19 years, 2 months past

In a fashion vaguely reminiscent of the process by which weeds keep growing back no matter how you try to rid yourself of them, the question of browser market share has once again been rearing its foul, misshapen head.  Dan kicked off a round of it over at Simplebits, but it’s recently been popping up other places as well.  I heard discussions about market share at SES Chicago, perhaps unsurprisingly, but I’ve also been seeing the question on various mailing lists and other forums.

The only thing more frustrating than the persistent recurrence of this unnecessary question is the inappropriate gravity it seems to acquire in so many minds.

Look, I’ll make this very simple for everyone.  If you’re trying to figure out what browsers to support (or not) in terms of layout consistency on a given site, then the answer is very easy.  Whatever the site’s access logs tell you.  End.  Of.  Story!

For example, the stats for the past few days’ worth of visitors to Complex Spiral Consulting tell me the following:

User AgentPortion of hits
Firefox 43%
IE6 30.8%
Mozilla 8.8%
Safari 8.6%
Opera 2.4%

(For those who are curious, IE5.5 makes up 0.8% of hits.  Various flavors of IE5.x below IE5.5 total roughly 1.2%, but note that Windows and Mac users are lumped together there.)

Those statistics tell me quite a bit about the people who visit the CSC site, and I can use that information to decide what to do about browser support.  You know what those numbers tell you about which browsers to support (or not) in your designs for sites on which you work?  Absolutely squat.  Anyone who uses those access statistics to make decisions for their own work is a fool, and a misinformed fool at that.

In every design, we have to ask what browsers need to have a consistent experience, which ones can be given a reduced experience, and which ones get no design at all.  The user logs from another site are useless in trying to make this decision.  The “global statistics” from firms like WebSideStory are just as useless in this case.  They may be entirely accurate, but they are also entirely irrelevant when it comes to making design support decisions.  The only stats that matter are the ones that come from the site you’re designing.

In a like manner, I don’t care if you think visitors to your site or some other favorite site of yours are an accurate reflection of the overall Internet population or not: that opinion is similarly irrelevant.  It’s rather like me claiming that the people who come to our annual holiday party are an accurate reflection of partygoers in general.  Maybe they are and maybe they aren’t, but either way I don’t think you should plan your all-night rave to accomodate the kinds of people who drop by our house to have homemade bread and soup and chat about babies, politics, science-fiction movies, and the weather.  And vice versa.

(Do remember that your site’s stats may reflect its current behavior instead of your potential audience.  If your site is already broken past the point of usefulness in Safari, then you’re going to see very low Safari numbers.  Make sure that you’re comparing apples to apples, and only compare the numbers in your access logs for browsers that can already use the site.)

As for the related question of “at what percentage level do I decide a browser isn’t worth bothering about”—well, that’s really up to you, isn’t it?  I certainly can’t tell you when it’s worthwhile to stop worrying about IE5.0, or Netscape 4.7, or Mosaic 1.2.  I know what I think is appropriate for the sites I work on—and the process of finding the answer is different for every site.  It has to be, because every site is different.

Now, if you want to share your user demographics with anyone who wants a peek, hey, have fun with that.  If data exhibitionism is your thing, who am I to judge?  Just don’t pretend that the bits of data you’re exposing to the world are representative of everyone else’s, because I guarantee you that they are not.  As for anyone who happens to glance at your data: I hope they realize the same thing.


Comments (41)

  1. Pingback ::

    davidhay » Blog Archive » Stat Track

    […] ch by Dave Shea, yet nonetheless I will present my stats, if only to chip in about browser share. And may I just say […]

  2. “Do remember that your site’s stats may reflect its current behavior instead of your potential audience.”

    This is a very important point. Imagine how many site owners there are that don’t realise (or they do, but keep quiet about it) that one reason they aren’t seeing a lot of visitors with other browsers than Internet Explorer is that their site blocks other browsers. Not by giving them unstyled HTML, but by completely stopping them from entering the site. Then they can use their statistics to justify their IE-centric development.

  3. Mike D: I agree that users of NN4 aren’t getting a “full experience” when they receive an unstyled page, but in some cases that is a benefit. For example, my school district runs NN4 on ancient macs. Providing any rich content (including images in a large format) would bring the browser (and computer) to its knees. Also, most modern web sites are designed for a resolution of 800×600, while our older computers have a maximum of 640×480. A current CSS layout would cause an aggravating amount of scrolling. I would be interested in hearing from actual NN4 users and learning how much a lack of style inconveniences them.

  4. Well said.

    I said this on Dan’s site and I’ll say it here as well: If you decide to present a completely styleless raw HTML site to a certain browser (e.g Netscape 4), you are effectively dropping support for that browser. You can say to yourself that these users just aren’t going to get the “full experience” but in reality, you are putting forth a site which looks broken, feels broken, and from a human interaction standpoint, *is* broken.

    Quite simply, every user that comes to your site expects at least a certain level of design. But presenting them with *no* design whatsoever, you leave an instantly bad impression.

    I’m not saying this as if it’s a bad thing because it’s not. It’s a very good thing. As long as we are judicious and conservative with what browsers we drop support for, we can help speed up browser evolution, lower lead-times in creating and testing pages, and do all the things we’d like to do on the web. For me personally, my general rule of thumb is 1%, but this can be influenced severely by what the trends say. If a great new browser comes out but its market share is still small (e.g. Safari), I’ll support it no matter what the numbers say. If a crappy old browser is clearly on its way out and has no potential upswing in usage (e.g. Netscape 4, IE 5), I might dump support a little before it hits the 1% line.

  5. It is a well known fact that statistics is a dangerous thing, and need to be evaluated in it’s context no matter what it concerns, right? :)

    I totally agree you can’t use statistics from another site, maybe even with a completely different aim and reader group. Though, when a new concept is to be launched, a new site, some statistics can be useful to get a general feeling. But no one site’s statistics can be interpreted as the statistics.

  6. The reason I thought it was worth bringing up, is to encourage people to look at their server logs. They may be surprised at what they see. I’m also generally curious as to what other sites are showing for IE5 numbers. I probably should’ve titled the post “When Can I Hide From IE5?”, because as you said, it’s your own site’s statistics that should matter the most. Agreed, there.

  7. Thanks for the post: it re-affirms what I’ve been pushing on a web project.

    I’m not sure what percentage of the public at large still uses netscape 4.79, but I’ve been keeping a log with my employer’s site and there hasn’t been a single Netscape browser below 7.0 in over 8 months. I can now push for a CSS-positioned layout (up to IE 5.0, but hey, I’ll take it) and dump the tables without fear of angering clients or the sales staff.

    We’re using the same web stat strategy for javascript: 3-4% have it disabled, so we’ll use it, but only with a transparent alternative that does not alert the user that he’s missing something or give him a page with no design. The idea of bloated, flashy DHTML menus is out; simple styled lists are in.

  8. I apologize for the double comments, but I left out the IE 5 stats that Dan referred to in his comment. The site shows 15-20% IE5, 3-5% IE 5.5, and 75-90% IE6 depending on the month. Other browsers are usually 1% or less. Most of our customers are corporate with an approved browser that everyone uses, so if we get a spike in traffic from one company, it can throw off the stats a few points. Is it odd for IE 5.0 to be more popular than IE5.5?

  9. Another thing to notice is that web-statictics is inaccurate in the first place.

    I did an extensive amount of manual log analysis in May, and found that a huge amount of visitors (for which your browser stats are calculated) are not real people – or the same person. The analysis was done and compared with 4 individual sites and 3 statistic tools (1 in-house + 2 external products).

    On a daily basis anything between 52% and 86% of the recorded visitors needs to be treated as non-persons. And since you do not know what browser the non-person entries is using, your browser stats is practically invalid.

    BTW: we could just add (+/- xx) to indicate the unknown factors.

    User Agent Portion of hits
    Firefox 43% (+/- 32)
    IE6 30.8% (+/- 22)
    Mozilla 8.8% (+/- 7)
    Safari 8.6% (+/- 7)
    Opera 2.4% (+/- 2)

  10. Since I’m just starting out with my own site soon, having just learned (x)html and some css, and I’m going to design the site directly to standards and push firefox.

    My mother runs firefox on her 150Mhz pentium desktop and it runs great, displays css remarkably, and so on.

    Everyone needs to get in the habit of upgrading browser software, and become active consumer in this technology, because it is the most important communication tool in history, and people need to become responsible to our collective future.

    It might sound harsh, but if my mom can run a standards compliant browser on a 150mhz system with 48MB of ram, then anyone can… so no mercy.

    Hopefully other new web designers take the same approach because it would go a long way towards making the internet environment stronger.

  11. Progressive enhancement is the key word; and checking the logs for a particular site. Some of my better clients understand this, and stopped worrying about pixel precise layouts. As long as the content remains accessible and attractive in older browsers, I can get them to play along.

    The one browser that is a difficult case is IE Mac. It doesn’t go away, even on OS X. (on commercial sites, I keep seeing half the Mac audience using it — might be specific to Japan. Safari is still having some character encoding problems when sending from data. And OS 9 is still used by many people).

  12. Is it odd for IE 5.0 to be more popular than IE5.5?

    No, it is just an indication of the lazyness of users in upgrading.

    IE5.0 was shipped with Windows 2000. IE6 was shipped with Windows XP. IE5.5 was not shipped as default with any major OS (I think it may have been in Windows ME)

  13. The only time that you cannot look at your own site’s logs is when you aren’t designing for any particular site. This is the case when developing web applications. How do the phpBB folks decide which browsers to support? I wouldn’t be surprised if they take a look at stats like WebSideStory’s. And in a product that’s meant to be used on a large number of sites of different varieties, I think that makes a lot of sense.

  14. Trackback ::

    Lost Below the 49th

    Browser Share
    Ok, I am with Eric on this one. I do not give one iota about designing to a specific browser. I know what I use, and my pages look good in that browser. I also know that the browser I

  15. Adam touched on an un-mentioned fact to this point: what if you cannot get access to your client’s web logs?

    I’m serious.

    Sometimes it takes an act of Congress for someone in marketing to talk with someone on the web team to talk with someone in the IT department who actually likes the aforementioned people and will provide that information.

    In that situation, I make a judgement call on what browsers to support and to what degree of perfection: [skip browser list]. If it’s not perfectly backward-compatible, I don’t lose sleep anymore.

    And I don’t lose sleep over dialup users either. Or users with black and white monitors. Or users on rotary telephones. Or Lynx users who still us Pico as an editor and Elm as an email client.

  16. Those silly statistics. Even on my own web space I odn’t really trust them. They declare that 40% of my visitors use “Mozilla 5” which I’ve never heard of. I was quite surprised to see how often Google vists me too. That skews the stats.

    It’s quite remarkable to me how many “upscale” (by upscale I mean non-corporate) individuals are using non-IE browsers. And after we hear how much market share IE has. I “Don’t Care About Market Share!”

    Here’s a question: How long of a time period is reasonable to analyze when deciding browser support?

  17. Like-minded sites will have pretty much the same audience, that’s why browser statistics for the creative people which comment here are so simillar and so different from those of other sites, such as alexa.com

    So, Eric, I can look at your stats and have a good idea of what I can expect on my own site (if our sites had simillar content)

  18. I don’t think browser statistics should be used much at all when deciding which browsers to aim for, and which to allow for a site to degrade gracefully. It’s simple:

    1. Build for standards, and test in the most standards compliant browsers available.
    2. Fix for broken browsers where required so that full accessibility is maintained, but it doesn’t have to look perfect.

    For my own site, my current host does not even provide access to statistics or logs, so I really don’t have a choice. But even when I do switch to a better host that does, I won’t be putting to much weight on them for decision making.

  19. Your stats will be skewed if your page is less accessible or less attractive in some browsers. Stats of site in flash show that 99.9% have flash installed. IE-only site has only IE visitors, and so on…

    Stats show who browses your site, not who would like to.

    For that matter you’d have to open your page to all browsers first or at least measure browser popularity on entry-page(s) only.

    Another thing is effort you need to put to support some browsers.Firefox-targetted page will usually work flawlessly in Opera, but design for IE5 (which may have similar share) requires almost a redesign. Supporting “top5” is not always the best solution.

    All in all, I think you should not believe your server logs.

  20. Hey Eric,

    You missed the most obvious reason that people don’t just base their designs on their access logs: they may be working on a completely new site. Which makes everyone else’s access logs a good place to go for advice, since theirs are blank.

  21. How do the phpBB folks decide which browsers to support? I wouldn’t be surprised if they take a look at stats like WebSideStory’s.

    Well, we the phpBB folk still only use such stats as a guide. Afterall, phpBB is installed on sites ranging from those whom the users are non-technical and more than likely still 90%+ IE to those such as MozillaZine where the 90%+ are very well Mozilla/Firefox users. Other phpBB sites probably get everything inbetween as well…

    A little look back to when phpBB2.0.0 was in development (2001), we already had the goal of core markup output by the application to be XHTML compliant. So why does subSilver, the style phpBB ships with, use an HTML4 DTD? Because it was making use of Mozilla’s (0.9.x I think) DOCTYPE switching to get it to work in sloppy mode, as strict mode had a very annoying line-height bug in table cells (IIRC). The web as it was then was quite different, as Eric himself was only just making live his ComplexSpiral demo.

    So less your server stats say whatever user-agent has a market share below statistical noise, then the truth is that you should cater for all that have anything above 0%. Probably why web developers should chose the most standards compliant browser of the day (if someday in the future it’s the successor of IE, then so be it), use that as the gold standard to develop to, then work downwards in how your (or client’s) site degrades.

  22. Great post, Eric. I know we’re talking browsers here, but your post reinforces rather well the idea that each site’s audience is completely unique. (they make up their own set called “example.com’s audience”) And every decision made on a particular web site that involves the people who use it should be made with only that population in mind. Not an ideal population, not an average population, not the population that they’re talking about on your favorite blog, but the population of the people who are actually taking time to come to your site.

  23. I think what Eric is saying is, “know your audience”, which can be applied anywhere, regardless of medium.

    I’ll tell you one thing: I love security flaws (well, not really).

    Our sub-Standards browser support in our corporate audience is 0.08% simply because of those beautiful IT brothers out there enforcing security. It may be just IE 6.0, but it sure beats NN 4. :)

  24. Because it was making use of Mozilla’s (0.9.x I think) DOCTYPE switching to get it to work in sloppy mode, as strict mode had a very annoying line-height bug in table cells (IIRC).

    It sounds a lot like you were relying on table layout in a non-standard way:

    Does Netscape 6 Break Your Table Layouts?

    The article doesn’t go into why it breaks, but the comments do.

    Another thing to bear in mind is that httpd logs merely record the transfers out of your server, not the downloads to the clients. The can be a major difference between the two depending on external factors like caching. I wouldn’t recommend paying attention to server logs for much beyond performance tuning.

  25. Eric, I agree with 98.62% of what you say. You rightly criticise the use of “global stats” from firms like WebSideStory and The Counter because they are not representative of the site you’re trying to design.

    I mistrust these stats for another reason: I have no reason to suppose they (and almost all other stats that are offered) are reliable or statistically valid. They generally fail to explain their sample size and period and they don’t provide any assurance that they identify browsers (and non-browser UAs such as spiders) accurately. There are huge pitfalls in doing this, often resulting in the understating of non-IE browsers. Also, in quoting percentages to one or even two decimal places, they fall into what I call the Abyss of Spurious Accuracy and seek to drag developers in there as well.

    The css-d Wiki page http://css-discuss.incutio.com/?page=BrowserStats contains my own modest contribution to the debate, primarily for those to whom the subject is new.

  26. I realise your entry is more concerned with the fact that ignoring a browser is a personal choice. But taking a maybe slightly different slant, surely there will come a time when, as an industry as a whole, we should push to code to the better browsers? Maybe not now, but in the future, we shouldn’t be coding hacks for people on bad browsers, but pointing them to a better one – which would then allow everyone to push the boundaries again. It’s all very well being backwards compatible, as long as it doesn’t hinder progress. So in a way, the browser share issues are important?

  27. If I may just throw another variable into the mix here.

    there are so many young people leaving collage, Uni, High School, etc. I would love to know that they understand the whole browser debate? What I mean is, do they even know that a browser existed before IE 5.0?

    As a lot of the web designers of the future are just leaving higher education right now (in Australia) to enter the real world, I would like to think they have the knowledge they need to really understand where the web has come from, and not just where it is going..

    I suppose what I am asking is, are the web designers of the future going to be as backward compatible as the web designers of this era? Do they even know that older browsers don’t support CSS, etc? Unless of course they are the self educating type.

    I am personally of the opinion that a solid “web curriculum” is needed within the education systems world wide if these new comers are going to meet the standards set by “geeks” (sorry) like us / you guys ;). To this day there is still not one “course” that would satisfy even my expectations (as an employer) of a “Qualified Web Designer” and this is the year 2004.

    By the way, this is all coming from someone who has only 3years experience of running his own web design business, but the point is if you don’t care about what you a serving then why bother serving it at all? (unless time is a constraint when trying to rebuild your own site ;) .)

    Sorry I think I may have strayed off topic…

    Just my humble opinion, cheers.

  28. Our site gets 7-10 thousand unique visitors a day with this breakdown for November:

    IE6: 60%
    Firefox: 16%
    IE5: 9%
    Mozilla: 5%
    IE5.5: 3%
    Opera <1%
    NN4 <1%

    Our site is not quite as geeky as Meyerweb or Complex Spiral, but is comprised almost exclusively of web developers. Although the stats are completely different than Eric’s, it does support his opinion unequivacably.

  29. I generally aggre with all of what Eric posted, but I just wanted to throw in two more reasons why general statistics may be widely off-scale.
    1. As Eric mentioned some websites don’t allow acces to anything other than IE, as this doesn’t go in reverse (I don’t know of any site that does not allow IE) the general statistics of, for instance, WebsiteStory are biased, as a portion of the hits for which they seem to gather statistics are quite likely based on above mentioned sites.
    2. I, and every IT specialist in my surrounding uses FireFox and with FireFox comes this nifty tool called AddBlock. The result, as far as WebsiteStory is concerned, I’m a ghost, I don’t count because I don’t want them to see me, they’ll just require more time for a page to finish loading…

    However, it’s verry nice to have a post of someone whom people listen to to back up my story when I try to convince someone that browser-statistics don’t matter.

  30. To underline your statements Eric, I noticed this article at CNET today. Which, and I quote, said amongst other things:

    “It seems that people are switching from Microsoft’s Internet Explorer to Mozilla’s new Firefox browser,” Niels Brinkman, co-founder of OneStat.com, said in a statement in November.

    Microsoft has disputed these numbers, claiming that they do not represent corporate users.

  31. My response got too long to be a comment, so I posted it here.

    Martin Reurings:
    There are definitely problems with the figures in OneStat’s November press release. They just don’t make sense when compared to their previous releases. More here

  32. I agree completely. Know your audience. If you DO care about which browser to design for, only look at your own site statistics. If you don’t care, no statistics and factual infos would be necessary.

  33. Since being “enlightened” earlier this year, I will not make a website unless it looks the same in all browsers since IE 5.5 etc, that way I don’t rule out any browser.

  34. The problem I find with statistics like that is that visitors of any site fit certain criteria. On my site, Firefox is a majority, but it tends to attract bloggers, Palm user, and… myself (true FF user). Care about your market share, not the big published figures.

  35. Trackback ::

    just to let you know

    Browsers and Marketshare
    Good piece on the recently surged discussions on browser share and IE6’s potential end of No.1 rule. Despite me being fairly upbeat about Firefoxes success, Eric hits the nail on the head again as usual: We design for a community…

  36. Firefox 43% (+/- 32) me too :)

  37. Pingback ::

    Max Design - standards based web design, development and training » Blog Archive » Some links for light reading (21/12/04)

    […] t reading (21/12/04)

    December 21st, 2004
    Web design world cool-down Don’t Care About Market Share HTML tags ALT attribute (ALT tag, […]

  38. Pingback ::

    Eric’s Archived Thoughts: Don’t Care About Market Share at go.

    […] gly Websites Spore Gameplay Video – Google Video » […]

  39. I agree with the “know thy audience” approach, adn with giving ptiority to your own siter statistics.

    That said, there is one thing to be aware of – what is known as the “self-fulfilling prophecy”. porneL touches on it above and Luke Barton fairly creates it.

    If you design syour site one way, so that it works primarily, or only, for visitors using one type of technology be it browser type/version, connection method, reading method etc, then the logs for you site will be massively skewed to the audience you designed it for. Sure there will be te occasional other visitor show up, but they won;t come back and over time your stats will show exactly what you created them to show – not becasue there aren’t users out there with something different that would like to visit/use your site, but becasue, for one reason or another, they are unable to becasue you designed it that way.

    I have a small book that was given me as a humourous, and yet serious, gift by one of my mathematics professors when I was an occasional TA for him. It’s called “How to Lie with Statistics”.

  40. My apologies for the typos in the above – sore hands and lazy proof reading – mea culpa.

  41. Trackback ::

    RAMsey

    A Visit to Browser Stats…

    Matt Mullenweg writes:
    I”m at An Event Apart in Chicago and Eric Meyer just said that browser statistics were “worse than useless.” More specifically, the only browser share numbers that matter are the one for sites you run, not what the web at …

Add Your Thoughts

Meyerweb dot com reserves the right to edit or remove any comment, especially when abusive or irrelevant to the topic at hand.

HTML allowed: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <em> <i> <q cite=""> <s> <strong> <pre class=""> <kbd>


if you’re satisfied with it.

Comment Preview