3 ms·

True, but it took a LONG time - IEs 6 and 7 were released in 2001 and 2006 respectively, so assuming your '4 years ago' stat is correct, that's still 3 years of

True, but it took a LONG time - IEs 6 and 7 were released in 2001 and 2006 respectively, so assuming your '4 years ago' stat is correct, that's still 3 years of waiting for a sweet spot of upgrades, by which time IE6 was 8 YEARS old. IE7 still gives me enough traffic (~5%) to require support, so that's a total of 4 active versions I need to cater for. I really hope adoption of latest IE speeds up, but I've been hoping that for a while now, and I just don't see it. Given that XP introduces a cut-off on IE8, I doubt we'll see the situation changing for a while.

nbouscal 14y ago
I still can't wrap my head around this concept that 5% requires support. Tim Ferriss isn't right about everything, but his bit about firing your worst customers is dead on. The Pareto principle suggests that 20% of your customers cause 80% of your workload, and the sensible thing to do is drop those customers. If it's only 5% instead of 20%, it seems like a no-brainer to me.
- eurleif 14y ago
  It really depends what kind of business you're building. If you rely on the network effect, for example, 5% of your user base helps to lock in the rest of your users, and so losing it may have a greater overall impact than just 5%.
coldtea 14y ago
I don't think 5% requires support. 1) Opportunity cost. Supporting them takes time and makes your code more bloated and less able to use cool new stuff without hacks and workarounds (and that cool new stuff can give you a competitive advantage to the other 95%). 2) This 5% is only gonna go down. 3) 5% is too small anyway. 4) There might be 5% of traffic, but how much is quality traffic? Depending on the site, a guy with the latest laptop, OS, browser and everything might be more likely to spend money, compared to some XP using guy that might have gotten there accidentally. This needs more search into your user data to tell. 5) Even if you want to still cater to IE6, you don't have to support it per se. You can just make it generally sure that its user can click around and see the pages, even if they lose a lot of formatting and cool options. I mean, don't even go for "progressive enhancement" stuff. Just let them be able to at least see the page, even if it looks like crap without the necessary css etc support.
acdha 14y ago
IE7 is old enough that I would point users at all of Microsoft's “please upgrade” pages — if you deal with financial or personal information, it's worth doing just for the security benefits.

nbouscal 14y ago

I still can't wrap my head around this concept that 5% requires support. Tim Ferriss isn't right about everything, but his bit about firing your worst customers is dead on. The Pareto principle suggests that 20% of your customers cause 80% of your workload, and the sensible thing to do is drop those customers. If it's only 5% instead of 20%, it seems like a no-brainer to me.

eurleif 14y ago
It really depends what kind of business you're building. If you rely on the network effect, for example, 5% of your user base helps to lock in the rest of your users, and so losing it may have a greater overall impact than just 5%.

coldtea 14y ago

I don't think 5% requires support. 1) Opportunity cost. Supporting them takes time and makes your code more bloated and less able to use cool new stuff without hacks and workarounds (and that cool new stuff can give you a competitive advantage to the other 95%). 2) This 5% is only gonna go down. 3) 5% is too small anyway. 4) There might be 5% of traffic, but how much is quality traffic? Depending on the site, a guy with the latest laptop, OS, browser and everything might be more likely to spend money, compared to some XP using guy that might have gotten there accidentally. This needs more search into your user data to tell. 5) Even if you want to still cater to IE6, you don't have to support it per se. You can just make it generally sure that its user can click around and see the pages, even if they lose a lot of formatting and cool options. I mean, don't even go for "progressive enhancement" stuff. Just let them be able to at least see the page, even if it looks like crap without the necessary css etc support.

acdha 14y ago

IE7 is old enough that I would point users at all of Microsoft's “please upgrade” pages — if you deal with financial or personal information, it's worth doing just for the security benefits.