Can dropping IE6 support be considered a realistic performance indicator?

Question

As part of the new year, my manager has required me to put forward a number of performance goals that will make up my performance review at the end of the year.

One of the indicators I'm considering pushing is dropping support for IE6 from our (large) public website.

The site currently has a representation of 12.1% in IE6. However nearly all our internal users are forced to use IE6 as it is the company standard (like so many other corporates out there).

Given that my team spends probably a good 20% of their time just getting things to work correctly in IE6, I can see a quantifiable cost benefit there, plus it would mean that our external customers would be "encouraged" to enjoy a more secure browsing experience by upgrading.

But is this an appropriate performance goal and should we be forcing the "mum & dad" users to abandon something they're familiar with?

Answer 1

Performance is vague here. Do you mean just a faster website and less errors and no loss of money? maybe so. But if you mean you'll lose $20,000 so you can gain something that requires a computer to measure the speed with, I'd leave it. If you can take the monetary hit, and you will increase your revenue soon to justify the loss, AND you can measure either now or in the future by some metric that whatever "performance" means is accomplished then yes. do it. I think you need to demonstrate you'll make more money in the future to justify the current loss.

Answer 2

But is this an appropriate performance goal and should we be forcing the "mum & dad" users to abandon something they're familiar with?

From a moral standpoint, the answer to the latter is definitely YES. In the same way that it would be a good thing to "force" home owners to stop leaving the front door key under the welcome mat. (But lets face it. Even if you do something the stop the site from working for IE6 users, you are only forcing them to choose between using a different browser or not using your site!)

But from a business standpoint, it is not your call to make. Leave this decision to the people whose responsibility it is to make business decisions; i.e. your management.

Either way dropping support for IE6 is (IMO) not an appropriate performance goal. It is really a matter of what your job is, not how well you are performing it. (I'm assuming that you are developer/support staff, not management.)

However, if you could (single-handedly) undertake to convince your company to change its IT policy on internal use of IE6 ... that would be a worthy performance goal. ;-) ;-)

Answer 3

It depends what you mean by “getting things to work correctly”. Of course your site should run with IE6. Partly because turning away 12% of your current userbase is insanity (and business suicide), but mostly because if your site requires a load of complicated styling and scripting to be usable at all you're doing it wrong and it'll surely break on other browsers, be they really old ones, obscure ones, or newly-invented ones.

But no, I wouldn't spend much time today on making everything render in exactly the right place on IE6. I haven't for some time and many other sites don't; IE6 users are going to be used to seeing layouts and images that are just a little bit wonky by now. They may also be used to sites popping up an “IE6 isn't very good, you know. Why not upgrade?” nag, which is another weapon you can try without completely breaking the site.

If an IE6 user has to live without some of the swankier animations, or loses some transparency cleverness, or ends up with columns slightly the wrong size, I don't care. But if they can't use the site at all I've done a blunder and haven't been doing my progressive enhancement properly.

Answer 4

I agree with the above answers and comments in general, but I think we've overall skipped over the 'performance goal' aspect. I think of performance metrics as positive, specific things, not implementation questions. Thus I might suggest, if you believe that dropping IE6 support is the correct thing, metrics such as the below with appropriate time frames and including results that are measurable in your organization:

• Work with customer teams, information technology, and other stakeholders to determine appropriate browser support, and supported time frames, for internal and external users;
• Isolate metrics that accurately reflect development and support effort versus both browser-type and revenue;
• Target a 20% higher fix rate, or whatever it is you are doing;
• Increase site security and interoperability by conforming to appropriate standards.

These goals aren't as numeric as I'd like, but they give you the idea of some specific, achievable targets. One way to achieve them is by discontinuing IE6 support, but you aren't necessarily limited to that; perhaps you can achieve them by targeting defects to revenue versus utilization. Or maybe you can drop support for Firefox. Or whatever else your manager tells you to do that you can't argue down.

Anyway I'd say the precise -implementation- is not the right thing to discuss as a performance evaluation, but the -business goals- are. Once everyone is on the same page, you can work up a proposal to drop IE6 maintenance - or maybe after discussing with customer ops you'll find that it'd be better to do something else, like hiring 20% more contractors to support IE6 while the rest of the world moves on. :)

Answer 5

"Performance goals" is an ambiguous term. I'll put forward 3 possibilities.

If you want to improve the performance of your application by not having to parse CSS for IE6, then you should get rid of support for IE6.

If you want to improve your own performance on the job, and removing IE6 support makes the CSS less bloated and rendering noticeably faster, and if you know you can do it with minimal effort, then you should get rid of support for IE6. You'll look like a hero for minimal work.

If you want to improve the performance of your business, then you need to see the impact of those 12.1% of folks who use IE6. If your's is a free site where users don't pay, where there is no ad revenue, then you will get reduce operating costs for your web site and improve page rendering performance for people using IE7+ and other modern browsers. This is good for business performance If those 12.1% of users are paying customers users from whom you derive advertising revenue, then getting rid of them will cause your revenues to drop by about 12.1%. This is not good for business performance.