Browser

Should Web developers keep up with browser statistics?

Tony Patton explains why developers may want to keep up with Web browser statistics and describes where to find this information. Discuss how browser market share impacts your development work.

TheCounter.com data for August 2008 show that Internet Explorer 6.x and 7.x make up 77% of the Web browser market, with 17% of the market going to Firefox. So do you use this knowledge when designing a Web application? I do to some extent.

Although I don't track Web browser statistics very closely, I do keep up with the latest browser version and what is the most popular (or used), and I design with standards in mind. I also test with the most used browsers (which is why I need to know what is going on) to ensure the Web application works as planned.

Here are three reasons why I can see that developers may want to keep up with Web browser statistics:

  • To monitor which Web browsers are most used on their site.
  • To create a test plan.
  • To design with the most popular browsers' standards in mind.

Keeping up with Web browser statistics

If you are interested in monitoring Web browser usage, there are a variety of resources available. These resources provide details (numbers, percentages, etc.) on the visitors and their browser to a particular site or group of sites. The following list includes a sampling of these sites.

  • TheCounter.com offers a subscription service where users can track browser usage for their sites. It counts/tracks only the page on which it is installed. You may also access global statistics for all sites.
  • W3Counter offers a free and a paid solution for monitoring a Web site's audience. It also offers global statistics.
  • W3Schools has a variety of wonderful free tutorials for Web developers. In addition, it provides browser statistics and other details about its visitors. (Note: W3Schools places Firefox with a 42.6% share. The larger share for the open source browser should not be a surprise since the site provides training on Web standards as well as some proprietary technologies.)
  • Market Share is a commercial product for tracking individual site usage statistics. This product also provides a global report.

It is worth noting a few possible anomalies with these statistics. For example, overestimation may occur with page refreshes, loading resources like css or JavaScript files, and feed readers may be counted. (Wikipedia has a good explanation of these anomalies.)

Using Web browser statistics

Web applications should be designed using Web standards with CSS for presentation and JavaScript for behavior, but knowing browser statistics reveals what standards are recognized by the user community. But you shouldn't lean on browser usage statistics to deliver non-standard functionality that isn't supported by other browsers.

Browser statistics are crucial when developing a test plan; that is, an application should be tested with the most popular browsers. Browser statistics also include details about platform, screen resolution, country, and more, so these additional details can be used to construct test plans.

Developing for intranet and Internet applications

When it comes to Web applications, there are two main categories: intranet and Internet. The intranet applications live within a vacuum not accessible to the outside world. The audience for an intranet application is captive and usually consistent, as the browser choice is often dictated by management. This presents a more stable environment for developing applications because there are no worries about a user that may use the application with Netscape 2.0.

Internet applications are open to the world. While they may only target a certain user group and particular browsers, anybody with a connection and browser may access the application. This can introduce more design considerations as the uncertainty of browser can affect what is delivered, or a group of browsers may be targeted (as opposed to one with intranets).

Internet applications are more prevalent than intranet applications, as more businesses open their applications to business partners and the world. However, delivering non-standard functionality on an intranet can be troublesome when a new browser version is delivered that changes what it may or may not support. After all, you do not want to revisit application design with each browser version. You can use the data to decide what standards to utilize, but only intranet applications should be fully aware of the browser used.

How does browser market share impact your work?

Do you think it's important for Web developers to keep up with browser statistics? Do browser trends play into your Web development projects? Do you target one or more browsers? Should browser statistics play any role in application design? Share your thoughts and experience with the Web development community.

Tony Patton began his professional career as an application developer earning Java, VB, Lotus, and XML certifications to bolster his knowledge.

---------------------------------------------------------------------------------------

Get weekly development tips in your inbox Keep your developer skills sharp by signing up for TechRepublic's free Web Developer newsletter, delivered each Tuesday. Automatically subscribe today!

About

Tony Patton has worn many hats over his 15+ years in the IT industry while witnessing many technologies come and go. He currently focuses on .NET and Web Development while trying to grasp the many facets of supporting such technologies in a productio...

33 comments
Justin James
Justin James

It's important to look at them every 6 months or so, just to make sure that you aren't surprised by something. It isn't like you are going to suddenly re-write two thousand HTML files just because Safari crossed some threshold of market share. Put simply, the same two HTML engines have been pushing back and forth on the same disputed 10% market share margin for a few years now. Firefox has its adherents, IE has its camp, and between the two are the people who don't really care and use whatever their default is, whether it be IE or the copy of Firefox that their local computer guru installed. As a developer, I try to write code that adheres as closely as possible to the HTML spec, and renders as closely as possible to my intent in Firefox and IE. If the two show it slightly differently, of well, I can live with that. If it shows them really different, that means I wrote some really bad code, since they both render standard code pretty similarly in most cases. I'm certainly not going to try torely upon browser sniffing + browser=specific hacks to make my code look right! J.Ja

djmort
djmort

I know that it is more important to make sure your page links are accurate and correct. Your link to SSL certs is "broken."

Ceespace
Ceespace

Once ina while I check thecounter.com the browsers / OS and screen resolutions stats. This can be handy - I downloaded Firefox and started using that to check my work with after it started up the chart, prior to that I used Opera and Safari to check with, based on what my client had

mattohare
mattohare

People are browsers. Not the software. The stats I get tell me a lot more than which piece of software they use to see the site. We can see some information about where they are located, any search strings they may have used on a search engine, and what pages they may have viewed. Not all of this will be strictly accurate, but it can do a lot to determine your market. If I can, I check out the parts of my site that the search strings may represent. I may find ways to add content to the pages. If I get a lot of viewers from French-speaking areas, I may add a French version to the site. Really, if all you see is what piece of software people use to visit your site, you're missing most of what's there.

jreeve
jreeve

Most certainly. Especially if you are developing a web-based application. Knowing that a small percentage of your audience uses IE 6.x vs IE 7.x will save you a lot of headache when trying to get things working in IE.

Duggeek
Duggeek

The simple answer to parent is, "yes". The addendum to that answer is, "...but not the ones presented here." The reason for that answer is largely due to the fact that nobody--but nobody--can know how many people actually have IE or Firefox (or Opera, or Safari, etc) installed. This article falls short on the most important part of the question; What's really being measured here? What these web statistics measure are the different browsers being used. The first fallacy is to presume that these numbers represent the population in general, which they really don't... they represent a sampling of users that are respectively visiting sites enabled with these web-trackers. W3C has a certain sampling, while W3Cschools has another sampling. Since theirs is a voluntary participation, this is generally seen as a 'balanced' set of statistics. (no bias from fiscal pressure or limitation to a customer base) Numbers from TheCounter.com are skewed by their own subscriber base. Not only that, the results given are vague and do not present a qualified analysis of their "global" numbers. (such as their quantitative sampling count) These numbers create more questions for me than what they would purport to answer. For instance, why are the 'monthly' statistic pages present a figure of >100 days? Shouldn't a monthly report show the figures for just that month? Perhaps it would be more salient if the sites in my charge used that service. Market Share is likewise misleading with 'global' numbers... the truth behind it: If people are visiting a site that isn't one of their subscribers, then they don't know about that visitor. My general point is that; browser-type statistics are only as good as their source. I firmly stand behind my initial 'simple' answer, but to truly understand that answer we should understand more about the question. If the question is, "Should web developers keep up with browser statistics for the respective sites that they manage?" ...then the answer is a resounding "yes". Designing towards the standard is always best, but then there's best practices. Until 100% of browsers are 100% standards-compliant, there will always be a need to watch the ebb-and-flow of the so-called Browser Wars. Winner or none, the important thing is to present your site in the best manner possible to your visitors.

jose.a.nunez
jose.a.nunez

Well, if your target audience is an intranet with all users in IE, then yes, do not bother on fixing FF Compatibility. If your target audience is the Internet... then yes, go ahead and ensure your site is compatible with both IE and FF.

johja
johja

The critical issue is that what you develop works for your audience. Most of my audience uses the following: IE7 on XP and Vista Firefox on XP and Vista Safari on Mac OS X 10.4 and 10.5 Firefox on Mac OS X 10.4 and 10.5 IE6 on XP At a minimum, I test what I develop and how it renders in IE7 on a Windows machine (sometimes I even test in Vista *and* XP), Safari on a Mac, and Firefox (sometimes I test in Vista, XP, and Mac OS X). Firefox does great job with standards. I cannot say the same about IE7. Sometimes IE has some quirks I have to code for. However, more often than not, what I develop works in Firefox and Safari and has few if any problems with IE. How do I know? I test. Why do I test? I want it to work for my audience.

steven.taylor
steven.taylor

As someone who has struggled in the past writing commerical web applications that are compatible with all browsers (including MACs), I know the frustration with this. But it is important. Now, developing and maintaining an intranet for one OS/Browser is a vacation. I've had to add up to 50% into dev schedules just becuase of this. However, the ASP.Net platform is very will suited for writing code one, and having the platform itself generate code based on the client. It works much better than anything done manually -- at least saves mucho time, which translates into mucho donaro.

G...
G...

We do not use a lot of time finding out which browser our visitors have because we "know" that there will be MSIE, Firefox and Safari. So we always develop and test on those 3 platforms. And, in my opinion, every site should.

edmicman1
edmicman1

Yes, I do check site statistics, to see what our users are using. But I don't really use that much in development, at least not in the sense of catering to the majority. When I develop web code, be it intranet or internet, I try to make it as much as possible to render the same or similarly in ANY browser, no matter what. This may be partially true standards compliant, but I won't purposely break things so they won't look right in IE, either. I do most of my development in Firefox, but I really want anything I do to display and function the same in Firefox, IE, Opera, Safari, or whatever. Minimalism is good. I've never really understood the need for extra controls, plugins, or things that might marry my sites to any specific technology.

khuryk
khuryk

Absolutely, a user's browser preference will impact their experience on your site. So if your goal is to attract and retain visitors you need to build your site to work with the most commonly used browsers. For every site there are 20 at least with the same or very similar content. Users have the option to stay or go to a competitor. If I am having a difficult time getting the content I want, I move on to the next site in the google search. It would be helpful if IE would get up to speed with trends for site building, better CSS recognition would be helpful. I browse with firefox but my audience trends with the norm, 80% IE users. So I need to adjust to IE despite it's shortcomings because that is what my current audience decides to browse with. Web builders are in the customer relations business as much as any other media outlet, maybe more so.

Deadly Ernest
Deadly Ernest

I know lots of people who use Firefox but have it set to tell the world it's MSIE. They do this for two reasons. 1. The company IT has a think about using only MSIE although the policy allows the use of FF. The IT people monitor statistics on all Internet use - in and out. They set FF to say it's MSIE and the IT people are happy thinking everyone is using MSIE as they say they don't support FF. 2. Some web sites check what your browser is before they really start to send you the files. Some of these will only send certain parts if they find you've got MSIE as they say 'it's best in MSIE and then set out to prove it.' By telling it you're using MSIE you get the full set up. I've seen this demonstrated with some web sites where people are pushing their tech expertise as MS IT shops. I also find it interesting that MSIE it's almost impossible to turn off the bit that tells the world what browser you are using, while most of the others it's easy and some of the stats gathering systems ignore the ones that don't identify. I find these sort of stats about how ubiquitous MS software is when we have so many major organisation have switched to Linux and Unix in recent years and countries like China where Linux software is very prevalent - well, according to a lot of the Chinese IT people it is. Add in all those Macs out there, the figures just don't seem to add up fully.

powellc
powellc

I use the same approach as Mr. James. I try to maintain standard HTML4.0 / CSS 2.0 code as much as possible, then do minor tweaks as needed if something really botched up between FF/Linux and IE/Win32.

mattohare
mattohare

I will be doing it a bit more when I get the next version of my site up though. I have to say, the one area I don't look at, at all, though is what software they use to browse the site. I like to see what they're doing on the site, so I know what to make better.

thewebmaster
thewebmaster

He said in a nutshell. I don't go over board but every 6 months check out how many are using what. I use Statcounter. It's right on the target.

TechLizard
TechLizard

Just this morning I received an email from MS about IE8. Check this out: http://support.microsoft.com/kb/956197 Compatibility Mode? Ey caraumba! Just when I thought we were over the "This site is best viewed in..." crap! Weighing in on the discussion: Yes, I test in both IE and FF for two most recent versions as well as Opera. I dont feel like I'm delivering a quality product unless I do. With the availability of sites like BrowserShots (http://browsershots.org/), it's just lazy not to!

ndavies
ndavies

This smart thinking. At the end of the day web sites have customers whether they come for info or to buy products they consume. It is therefore essential to meet the customers needs. This is great advice - look at all the stats available and tailor the site accordingly.

nerdizer
nerdizer

As long as people are discussing testing... As an older user my biggest problem with browser applications is that they are not tested with larger fonts. A large number of users are getting older, -something that does not come up in simple browser statistics. Check other demographic figures like the census etc. Large fonts are required by older eyes, we just can't see that 9 or 10 point default font. Larger fonts overwrite other areas of the screen, cause buttons not to work, or just cause part of the text or buttons to disappear. Some sites are written well and others are not. Please test with larger fonts also.

r.salazar
r.salazar

I think this is fairly off-topic, since this is more about server-sided work rather than client-sided. Now, .Net throws a huge amount of garbage to the client. It actually follows IE's behavior (which is really not the best one), so you're effectively targeting IE users. All this depends on the kind of work you do. If you're talking about a really "straight-forward" design (maybe even table-based) and letting the "controls" act through JS, well, yes, you have a point here. Otherwise, I think you're a bit off. Newer versions though ?incidentally hard to be able to work with (because of bureaucracy)? allow for a more cleaner and effective code on the client side, but that also means some more work for you, at least, at the beginning, and not many developers bother to do it.

steven.taylor
steven.taylor

While MAC users are a smaller percentage, don't forget them. That makes for more platforms than three...because the browsers behave differently on MAC than on Windows. Speaking of compatibility, not even this web site (tech republic) is 100%. Every time I log in, I get a javascript error, so the application does not remember me...so every time I post or do anything that requires it, I have to log in manually.

chris
chris

Knowing the percentages don't really help unless you (your client) is ok with excluding visitors (IE: presenting them with a less than optimized experience) and just wants to know about how many people they are doing that to. For a proprietary app, sure thing, I get that, but for a public site? Bad idea (lazy idea).

nkozi
nkozi

I'm getting sick of IE-bashing by people wearing blinkers. MS had a monopoly (whether it's good or bad) in browsers for many years. Anything they added to the W3C standards therefore does not mean it is non-standard, it became the de facto standard. The fact that W3C did not keep up, does not make FF a better browser because it only conforms to these out-of-date so-called "standards", it only means that it is an out-of-date browser. Unfortunately FF has a 17% usage share, so we have to take it into account and develop to the lowest common denominator. Pity.

pointzerotwo
pointzerotwo

It is certainly possible that MSIE is over-represented in web server logs for the reasons you describe. That's not a reason to ignore the data altogether. Most web developers have limited testing resources, but unless you're willing to alienate a lot of potential visitors, you should make sure your site is at least readable and functional (if not pretty) for the top 99 percent of their browsers. Paying attention to browser usage might be the easiest way for many sites to increase sales and/or ad-revenue by a few percentage points. Any skew in the data towards MSIE means other browsers are under-represented, and only reinforces the need to pay attention to the tail of the curve.

ndavies
ndavies

Having worked as a desktop developer for many years there were two hurdles I had to overcome when learning to develop for the web. 1. Platform 2. Training In desktop apps you can limit the platforms you develop for, many companies do this and it's no great issue. you can also provide training. In the world of web sites the platform can be any one of a number of combinations and it;'s essentials we know what the most popular are. Sites on the Internet (not intranet) are intended to publicise information to a wide audience - even if it is field specific content. Nothing bugs me more than being told by a site what browser I must use - this is wrong! Web developers need to know what the main options are so that sites can be developed and tested to work on all the options delivering the same message in the same style. This is where standards come in, the idea being that all browsers handle the same code the same way. As this is not yet the case (and probably never will be) we need to know that the work we produce functions correctly on all platforms, to know that we need to know what those platforms are. To keep up with that I simply use the W3Schools stats and monitor the specific stats on the sites I am responsible for. I mentioned training above and this is only tenuously linked so I won't twitter on for too long. On the web we don't have the chance to meet all the end users and so the design has to be obvious the navigation clear. This has to be true regardless of where the user is accessing the site. Which ties in with the point above. If a user is familiar with the site in FF and then shows a friend/colleague on IE6 and it doesn't look the same then they get a bad feel from the site, the potential new user will not graduate to full user and you run the risk of being 2 users down. So that's my opinion, fairly long winded so in summary we must know the trends and we must cater for them properly.

mattohare
mattohare

between this tread and some conversations I've been part of in Belfast. It's really easy to get on a narrow idea about statistics or anything. We've become so conditioned to looking at what's above the fold, we forget there's another half a page (or more) of text just waiting below the fold. In the case of Webalizer, it can go on for several screens.

mattohare
mattohare

I avoid writing pages that depend on certain dimensions for text areas, or even certain resolutions. So may pages (including CNet pages) can depend on large resolutions on large screens with small text. Not all of us have large screens. And many of us with large screens want to have more than one window visible at a time. (Otherwise, why not go back to the old single-task character based days?)

johja
johja

nerdizer, Good point on including larger fonts in testing. This is especially true with the push for greater accessibility.

deepsand
deepsand

"[i]MS had a monopoly ... Anything they added to the W3C standards therefore does not mean it is non-standard, it became the de facto standard.[/i]" That MS added non-compliant browser functions was the very means by which they maintained their monopoly; such were [b]not[/b] added for the purpose of contributing to the good of all, but solely for the benefit of MS. "[i]The fact that W3C did not keep up[/i] ...," owes to the fact that there were and are those who refuse to be coerced by a monopolistic power into submitting to the will of such. "[i]does not make FF a better browser because it only conforms to these out-of-date so-called "standards", it only means that it is an out-of-date browser.[/i]" To the contrary, any browser that is standards compliant is a better browser as it allows for an application to be more easily designed so as to have a predictable behavior. "[i]Unfortunately FF has a 17% usage share, so we have to take it into account and develop to the lowest common denominator.[/i]" The real LCD here is that subset of standards which MS has implemented, not the browsers that are fully compliant. This situation is one of Microsoft's own making, not that of the developers of other browsers.

Justin James
Justin James

Yeah, back when 17" - 21" CRTs and non-widescreen monitors were out there a lot, I pushed "eleastic design" and "liquid layout" very hard. Now, with the wider screens, it is actually quite difficult to read text without losing your place, so I've started to recommend fixed width on the content past a certain point (like having a minimum size and a maximum size, but no defined size). J.Ja

mattohare
mattohare

Anyway, I guess I can code my systems to look for such monitors and put some extra pictures there. Maybe pics of the Giants Causeway, Naomh Padraig's Cathederal and Fort Sumpter? Actually, I've wondered what I should do with my text intense pages. I'd rather you did didn't have to read 22" wide paragraphs. At my 'day job' we even put a bit of a width limit on the pages.

Justin James
Justin James

... some of us have big screens (I have a 22" and 24" monitor on my desk, both wisescreen), and I really hate in-elastic pages that effectively throw away $200 worth of screen space. :) J.Ja

Editor's Picks