Welcome to WebmasterWorld Guest from 220.127.116.11
Forum Moderators: buckworks
The loading time was second only to pricing in making people decide whether to purchase from a particular site.
4 seconds is half the time most shoppers were willing to give in the last survey, presumably because broadband connections have upped people's expectations for how quickly a site can appear.
[edited by: lorax at 1:58 pm (utc) on Nov. 12, 2006]
[edit reason] corrected to match report [/edit]
The only way to accuratly find out how long people will wait is to set up computers with different load times, and video tape people using them. Tell them the purpose of the experience is to find out how customers find the lowest price so they are not thinking about speed.
I'll bet you will find surfers are more patient than people think and only the longest load times will chase people away.
People expect more and more cool content so all these people saying they need to keep home pages under 5k are living in the past. Creating sites with dial-up users in sacraficing the growing majority for the shrinking minority.
Also, read the report carefully and pay close attention to the terminology being used. They use words like load and render in the same sentence. There is a distinct difference between the two of them. A site can render in a second and easily take 3-4 seconds to load behind the scenes.
But I have no trouble seeing a drop off starting at 4 seconds, especially on smaller "unknown" sites that I'm clicking through to from SERPS, especially for broadband surfers.
Here's my completely unscientific list of things that piss me off and drive me away from an e-commerce site (NB: this list is entirely made-up of things that slow load times.)
#1 Flash. May I repeat - Flash. Flash is for multimedia content sites. It sucks for selling stuff. Get over it. Really.
#2 Java. Man, I see that coffee cup in the corner of my browser while a page grinds away loading, I'm gonzo.
#3 >40% photos above the fold. I'm kinda odd this way, I know. But I'm an info oriented guy. If all I'm seeing is a mass of pics with no info to accompany it, it screams to me "all gloss - sizzle and no steak."
#4 Too Busy - Not enough white space. Focus your customer in on a few key identifiables. Don't shotgun him/her with everything and the kitchen sink.
#5 Too much blinky/moving stuff. Animated gifs scream "low rent"
Also NB: you can re-order that list any way you want. They're all equally annoying.
All those things add to page weight and/or grind out your browser to slow load times.
If you say to yourself "I'm doing that and my site is doing great anyway..." You're not asking yourself the right question.
What you should be asking yourself is "How much better would my site be doing if I cleaned it up and got load time down?"
If I can see the page is functioning, fine. But if it's really sluggish after 10 seconds, images staggering to be dispayed, I'm outta there, especially if the internal product pages are as slow.
Once you have the IP to the site, the product pages should be 1-2 seconds tops.
[/quote]You think he knows something unknown to Amazon or Walmart.com?[/quote]
Hell, I know something Walmart doesn't know, the drunken monkey that designed their site causes the page to hang up in the browser waiting for the data to render.
When we talk "load time", if I can get the HTML on the screen in 1 second, people will wait for images to trickle in, but if I'm sitting here watching the page bounce around wheezing like Walmart I might just puke. Amazon isn't so bad, for the volume of garbage they spew out in every page it's amazing it loads as quick as it does, especially compared to Walmart's junky site.
I just checked one of my old customers sites, which have some fairly bulky catalog/category pages, about 3-4 seconds and the product pages about 1 second flat, and that includes looking up all the product sizes, colors and product suggestions on the fly.
This reminds me a bit of the time when I had the quasi-monopoly to sell a very hot produkt in a big country. It took 5 days to proccess an order for a product we had on stock. People needed the stuff so they waited in line.
But all of a sudden the monopoly was gone and we had 500 competitors. Guess what: within 4 weeks we had changed the whole process to deliver every order that came in before 2pm at the same day (it was 20 years ago, so accouting software was less sophisticatd than now...)
Buttom line: people do not want to wait AT ALL. Everything over 1 second feels like waiting. So IMHO this is the goal.
You're making the assumption that most people are like you, but what if they're not? What if you're just one of the 25% who'll wait?
We put Google analytics code in a 1x1 div at the bottom of the page, advertisements in iframes, etc. Works well for us.
Admittedly, our site is informational content and community, not ecommerce.
I can't say this is totally unexpected. We keep expecting the gas mileage of our cars to improve. Hell we keep expecting beer to have less calories! What we're dealing with is expectation versus reality. It's a no brainer. If you can meet the expectation, you may win big. But it's a definate that if you don't meet the expectation, you will lose business.
[edited by: lorax at 12:14 pm (utc) on Nov. 11, 2006]
ok, I'm a bit cranky at the moment but w.t.f.? 4 seconds is unreasonable, as most of us seem to agree.
Most of us? I don't think so. Look at the second hand on your watch ticking away for four seconds. It's not a long time. Now let's analyse this process.
1. You click a link on the screen or enter a URL in the address bar.
2. As you wait for the new page to appear something may be displayed on the screen or it may just go blank.
3. You brain observes this process and after one or two seconds you are really brassed off so you decide you are not waiting any longer.
4. You decide to hit the back button.
5. You move your mouse pointer to the back button.
6. You click the back button.
The truth of the matter is that this whole process takes a finite amount of time. In practise it is probably impossible to do it in less than two or three seconds. People who say that they won't wait four seconds for a page to load are denying themselves a large part of the content of the Internet.
People who say that they won't wait two seconds either don't understand the question or they are telling porkies.
I am absolutely amazed that a professional webmaster forum is having this debate, a quick site is a no brainer to me.
Ummm....no one is saying that a slow site is needed. The debate is what qualifies as "fast."
Why not have a 4 milisecond site then, since a quick site is a no brainer?
I think I am a typical surfer, and unless the site is really important to me, I won't go back unless it loads real fast ( less than 8 seconds ).
when I design a site, I base the page load time on 6 seconds with dial up. then I can increase the load time on the back pages by 1 or 2 seconds.
another trick is to make sure my images load up in the progressive mode ( it makes them think that everything is fine and they don't wait forever ).
When I design a site, I base the page load time on 6 seconds with dial up.
I go about this a bit differently. I design a site and pay close attention to rendering. I make sure that I'm using my divs correctly and I do source ordered content (SOC). If I'm able to give the visitor the meat of the page first while everything else is loading, that works for everyone.
Performance not only applies to the website but also the server. And then you have the visitors performance issues to contend with. ;)
Anything (electron, photon, bit, whatever) needs about 60 milliseconds from London to Sidney - these 60 msec are enough to make Google spread datacenters all over the world. Ok, they want fault tolerance and terror safety and the like, but it's about speed as well.
60 msecs do matter for google.
And we discuss if we should aim for 2, 4, or 8 seconds?
Back when I used to do websites for a major government entity some of the research that was given to our agency suggested that the attention span of the average adult surfing the internet was about 8-10 seconds. Our goal was to load dynamic pages in 4-8 seconds and bring in static pages at 2-4. Considering in 97-2001 there were a lot more static pages, it wasn't all that tough. It was a rule of thumb in our circles that the frontpage of any of our sites should load at around 2-3 seconds.
These days you can force a lot more down the pipe in a lot less time, but the data and code driving a site spends more time on the server while the download timer is running.
Still waiting for someone to cite a high revenue/high profit E-COMMERCE site that fully loads under 4 seconds at 56K
Actually, the thread title indicates we're talking about broadband load speeds.
ncix dot com (are the mods gonna let that stay? I have no affiliation with that site). Loads in just a smidge over 4 seconds on broadband, when their servers are busy. When traffic loads are light (in the middle of the night), it comes in around 3 seconds.
The more established competition in the same market, memoryexpress dot com, has been around longer. MemX is losing market share online. MemX's front page takes about 7 seconds to load.
MemX has better prices, and operates out of a province in Canada with lower sales taxes. The money factors are all in MemX's favour. But they're still losing market share.
Do the math.
Look at the way those two sites have their landing pages set up, and you'll immediately see why the one loads faster than the other. You'll also see which one has the better "classical" marketing strategy, with good use of white space, less clutter and confusion, better focus, etc etc.
The site with the better load time wins on ALL these points.
If you aren't big and established, and able to rely on significant brand recognition, then you have to be leaner, meaner, faster.
4 seconds IS possible. Maybe not for every page, but you should be able to do it for your key landing pages. Just get rid of the extraneous crap and focus, focus, focus.
That is HOW you take out the big boy competition, and become the next big fat target for some upstart cowboy to take a shot at.