Comments on: Measuring time spent on a web page

By: CLOMIDONLINER

CLOMIDONLINER — Fri, 09 Sep 2011 01:42:14 +0000

[b]clomid worked before but follistim doesn’t[/b] clomid and follistim

By: roger tomson

roger tomson — Fri, 13 May 2011 00:29:25 +0000

All of the vendor’s important submit furthermore there spouse!

By: Web Writing Must be Simple, Short and Scannable | Kevin's Notes

Web Writing Must be Simple, Short and Scannable | Kevin's Notes — Mon, 27 Dec 2010 08:59:00 +0000

[...] page is competing with millions of others. On average, users scan a page for only 33 seconds. They quickly move between [...]

By: fajar

fajar — Mon, 12 Jul 2010 02:33:31 +0000

Thank you for the link

By: Fred

Fred — Tue, 02 Feb 2010 13:03:19 +0000

A network problem
clicked from google to go to this page
http://www.mondaynote.com/2009/05/24/measuring-time-spent-on-a-web-page

And found myself redirected a security scam page
You might check your site for a virus

I checked my machine and it seem to be clean
Ron

By: The web’s design problems | Monday Note

The web’s design problems | Monday Note — Sun, 22 Nov 2009 18:48:50 +0000

[...] 10 scrolls divide the probability by 100 or more. As discussed in an earlier Monday Note (see Measuring time spent on a web page ), a full 24% of all banners are not seen at all; a placement on the footer rates as low as 10% of [...]

By: SIT3 » Hello, goodbye.

SIT3 » Hello, goodbye. — Mon, 27 Jul 2009 05:03:05 +0000

[...] amount of time users spend on web pages. The average time spent on a web page today is about half a minute. And I would venture to guess much of that time is spent determining which piece of eye candy to [...]

By: SIT3 » Hello, goodbye.

SIT3 » Hello, goodbye. — Tue, 21 Jul 2009 15:06:44 +0000

By: Jim Smyth

Jim Smyth — Tue, 09 Jun 2009 13:29:57 +0000

As a former electronic publisher, and after spending over two decades studying this, the “time spent” model has very little corollary with the value of the content delivered. There are better measures, like the initial search, but even these need to be modified as technology changes and improves.

By: Frederic Montagnon

Frederic Montagnon — Mon, 25 May 2009 22:11:43 +0000

Frederic, thanks for this great post. Do you have some cost per 1000 minutes figures comparing TV to Web?

By: Laurent Nicolas (Alenty)

Laurent Nicolas (Alenty) — Mon, 25 May 2009 14:12:31 +0000

Elias,

I think that the technical issue that you mentioned is solved by Alenty. The user’s presence is guaranteed by its mouse-keyboard activity, with a limit (one minute by default). On top of this, duration is also limited by the visibility (of the page, of the content in the page or of the banner, etc.). So, it is fairly comparable across sites and works very well for ajax-based sites!
The same kind (at page level) of measurement is used in user-centric tools (NetRatings). The good news is that it is now available in site-centric and ad-centric tools.

Regarding the conceptual issue, I agree with you that some sites have a media model and sell ads by CPMs, and some don’t. There is no such thing as a universal metrics because the goals may be very different. Traffic generation is a different model, so it requires different metrics (clicks or time spent on the target site!).
But as far as media are concerned, it is fair to compare if and how long banners are actually seen.

By: Elias Bizannes

Elias Bizannes — Mon, 25 May 2009 09:27:33 +0000

Frédéric – great post. I too was excited about the time spent metric as being the future, covering up the bottomless pit of online metrics that’s killing online advertising.

But after some research, I discovered there is a technical and conceptual problem with time spent. The technical issue is that in practice, it is difficult to accurately measure. That’s because if a user only commits to one action (which initiates the counting of their time), they need another action to close off on that viewing period. Therefore, the metric collapses unless there is another action to mark the period.

The conceptual issue is that it doesn’t fairly value all sites. I think time spent is the more accurate measure available due to its comparability (ie, video and text are treated equally), but what about a site that has an in-and-out model. So something like techmeme.com – users don’t stay on it for long as they are clicking on news articles. I suppose the answer to this issue, is that time spent will never be a universal standard as it biases certain types of experiences (like media consumption).

But it’s the technical issue that’s a problem. It will just become one of the dirty secrets in the new media business, because it will lead to inflated engagement.