views:

554

answers:

10

Our product management folks are in the process of deciding which browsers we will officially support for our upcoming web front-end application. Where can I find reliable statistics on browser market share, broken down by agent and version, particularly for business users?

Edit
I should have added:

  • We are building a web application for, and served from, an embedded device. Since it isn't a web site, we cannot collect statistics.
  • The constrained resource is testing time. Although a well coded app will theoretically run just fine on any modern, standards-compliant browser, we simply do not have the time/money to thoroughly test our application on more than a few. The question then becomes: which few will get the official stamp of approval? The rest will go through cursory testing to check for obvious problems.
+3  A: 

w3schools has a browsers section on their site, with a lot of browser and other related stats

http://www.w3schools.com/browsers/default.asp

Midhat
These stats are not good because mainly webdevs go there and the stats are only for that one site
redsquare
Agreed--unless you are targeting web developers the numbers are not particularly useful.
Stu Thompson
+3  A: 

Wikipedia has data from several popular stats sources.

And reliable is quite a loose term, especially when it comes to statistics that is sampled from some arbitrarily collected data set.

NeARAZ
Probably very reliable, but unfortunately doesn't break the stats down into IE7, IE6 etc
Sam Wessel
+5  A: 

Bear in mind that you can't really rely on any set of stats to be accurate. Ultimately, every survey is from a self-selected population, and thus doesn't necessarily give you a true representation of market share. The best you can do is try and find stats for the same market segment and type of site as your own.

Nick Johnson
+1  A: 

I would venture to say that you'll need to aggregate several data sources. They're each likely to be a little self-selecting in their own way. You can also look at Hitslink, which we use to an extent, but the most useful stats of course come from our own users. Obviously you can't rely on this data, but you could also look at this question, and I think 1 or 2 others already asked here.

cori
+1  A: 

The problem is different market segments have different statistics (and business users may be too broad); e.g. Slashdot.org probably has a much higher rate of FireFox users than say MSN.com.

Your best metric is probably any existing online apps you might have targeting a similar audience as the new app. You can use Google Analytics for free; this will give you the usage statistics you are looking for, plus a lot more.

Giovanni Galbo
+1  A: 

The only really reliable browser statistics are your own. External data can be a good place to start when you have no data of your own, but it won't be even nearly as useful as looking at your own community and what browsers they use.

You may find, for example, that there are a lot of people using mobile phone browsers which may need some extra consideration in your site construction, especially if you're looking at business users.

Probably the most important bit of advice I can give is to try your best to support every modern browser, and degrade gracefully so you don't entirely exclude people who might be using something you didn't expect. Although this will increase your build time a little, it's significantly easier to do it from the start, than trying to add support later.

Dan
A: 

It depends very much on your audience.

For example I find the following stats (last month) on the sites I run:

  • Firefox 8% - Safari 2% - Sailing club
  • Firefox 20% - Safari 0,5% - Scouting group

All the rest is Internet Explorer (and some small browsers like Konqueror and Chrome).

The sailing club is targetted to men, aged 30+. Most use their business computer.

The scouting group are young, aged 20-, boys and girls. Using their home/hobby/school computer. Not in a enterprise environment.

The boys and girls from the scouting group, cannot affort an expensive Apple, so they do not use Safari. Some men from the sailing club, like to show off, and carry an iPhone, Mac etc. So you see Safari hits on that site.

GvS
So, umm...where do us FF on Mac people sit? :P Be careful how you interpret your data...
Stu Thompson
Hey, this is simplyfied data. I also have OS-stats.
GvS
+1  A: 

Support the standard. Make sure your site validates and looks OK on Firefox, IE and Safari, then don't sweat it. It'll look just fine on almost everything else. Beyond that, you must make sure you only use standard HTML widgets, or at least degrade gracefully so that all functionality can be reached with Javascript disabled. For a business case, consider that cell phones with embedded browsers account for a huge portion of web usage growth, and that you can't expect them to support any given plugin, especially not those only available on Windows.

Yes, it's a lot of work to make a pretty, functional site that works on both desktop Firefox and OddBrowser (tm) on a cell phone, but you don't do it, your competitors will. Don't let them steal your business by making it hard for your potential customers to use your site.

Just Some Guy
A: 

Instead of investing the resources to go cross-browser yourself, outsource it! If you're at all capable of choosing a toolkit that solves the cross-browser support issue for you, then go for it. Toolkits like GWT, Dojo, or ExtJS allow you to focus on the business logic, instead of getting bogged down in browser bugs.

Also, degrading gracefully is nice in theory, but in practice it's not feasible to build an interface that scales from the lightest cell phones to the heaviest ajax desktop. Recognize that you may need to build more than one front-end. Gmail has a mobile html interface, a basic html interface, and an ajax html interface. There's a reason they have three separate front-ends.

You'll need to understand your customer base to know what types of front-end you need.

Joeri Sebrechts
+4  A: 

I would suggest monitoring the W3C Global Web Stats which are published on a monthly basis.

GateKiller