🚀 go-pugleaf

RetroBBS NetNews Server

Inspired by RockSolid Light RIP Retro Guy

Article View: bit.listserv.nettrain
Article #187

Re: web use/response time

#187
From: JQ Johnson
Date: Wed, 05 May 1999 00:00
40 lines
3006 bytes
Kathy Gill quotes Nielsen:

> The 15-second guideline in the previous paragraph was derived from the
> basic set of response time values that have been known since around 1968.
> System response needs to happen within about 10 seconds to keep the user's
> attention

I did some human factors work in the mid 70s and looked closely at that
research at the time (it was still fairly new then!).  The results were all
over the map even then.  One big issue was the size of a transaction that
was being responded to.  10s is wildly overstating most of the results --
with a typical form-oriented transaction (user fills out a screen on a 3270
terminal and presses send) one could find that 10s was acceptable, but most
of the studies found that there was severe degradation in attention and
user satisfaction at as little as 3s.  The studies also found that
predictablilty (variance in response time) was much more important to user
satisfaction than actual response time -- people were willing to wait if
all transactions took 10s, but if half of them took 1s then any long
transaction was very frustrating.

For other kinds of response, e.g. keyboard echo time in a full duplex
typing environment, the acceptable response time was much lower.

What would you think of a guideline that required that web pages load in
3s?  Not very practical given current techology.  Or how about a tool that
delayed all loading to a minimum of 10s to reduce variance?

I'm not convinced that we can learn anything useful to the web from these
old studies.  Although one should respect Nielsen's intuitions on this, I
think that's all they are.  His practical recommendations for UI design are
very well thought out, but one should not accept a magic number here, be it
3s, 15s, or 30s.

What the acceptable response time is for web sites remains an open and
interesting question.

JQ Johnson                      office: 115F Knight Library
Academic Education Coordinator  email: jqj@darkwing.uoregon.edu
1299 University of Oregon       phone: 1-541-346-1746  -3485 fax
Eugene, OR  97403-1299          http://darkwing.uoregon.edu/~jqj/
Technical Details
Message-ID: <2467511870.925885480@metriclient-3.uoregon.edu>
Path: arc1-ca.pugleaf.net!unknown!not-for-mail
References: <199905050406.VAA24796@darkwing.uoregon.edu>

Show Full Headers
Complete Article Headers
X-Google-Language: ENGLISH,ASCII-7-bit
X-Google-Thread: 10449f,f8eaa3d2ff4be0e8
X-Google-Attributes: gid10449f,public
From: JQ Johnson <jqj@DARKWING.UOREGON.EDU>
Subject: Re: web use/response time
Date: 1999/05/05
Message-ID: <2467511870.925885480@metriclient-3.uoregon.edu>
X-Deja-AN: 475004543
Sender: Internet/BITNET Network Trainers <NETTRAIN@LISTSERV.ACSU.BUFFALO.EDU>
Content-Transfer-Encoding: 7bit
Approved: NETNEWS@AUVM.AMERICAN.EDU                              *
References: <199905050406.VAA24796@darkwing.uoregon.edu>
Comments: ********************************************************
Content-Type: text/plain; charset=us-ascii
Originator-Info: login-id=jqj; server=darkwing.uoregon.edu
MIME-Version: 1.0
Reply-To: JQ Johnson <jqj@DARKWING.UOREGON.EDU>
Newsgroups: bit.listserv.nettrain
Content-Disposition: inline