Sponsored in part by... Web Crossing WebCrossing Neighbors Creates Private Social Networks
Create a complete social network with your company or group's
own look. Scalable, extensible and extremely customizable.
Take a guided tour today <http://www.webcrossing.com/tour>

 [F] TidBITS  / TidBITS  / TidBITS Talk  /

Annoyance level of different ad types

[mmatty]mmatty (apparently) - 08:04am Sep 1, 2005 PST
via email

I always enjoy Glenn Fleischman's articles, and although this week's
review of Flash blockers is informative, I don't agree with the
implication that Flash ads are objectionable to a wide swath of web
surfers.

<http://db.tidbits.com/getbits.acgi?tbart=08225>

While text only ads on search engines have been, and will continue to
be, wildly successful and the surge in internet advertising has been
fueled by Yahoo and Google's ad programs. If the search engines were
not constantly introducing new options for advertisers, this growth
would inevitably slow (and stock prices would consequently fall), and
not long ago Google and Yahoo announced increased visual display
options for participants, and Google is actively testing Flash
banners on AdSense:

https://www.google.com/adsense/new

Google has been dropping hints that visual advertising on their
traditionally minimalist search result pages might be available in
the near future. This might be accelerated as Yahoo recently
announced the beta test of its Publisher Network, which will compete
with AdSense:

http://publisher.yahoo.com/

Combining words with moving images, and sometimes with sound, are
incredibly powerful influencers. And consumers don't necessarily have
to like them to be effective (i.e., Mr. Whipple). But Flash ads often
are appealing and interesting to consumers, and industry studies,
such as one conducted by Doubleclick, have proven that Flash ads tend
to generate better click through results:

http://www.doubleclick.com/us/about_doubleclick/press_releases/
default.asp?p=434

As Glenn mentions at the start of the article, advertising is
important to the survival and success of online publications.
Advertising pays for the creation, storage and delivery of content,
and rich media ads can command higher prices than text only or static
ads.

Marilyn


Mark as Read
  (older msg: 16)OutlineAll MessagesOlder MessagesOldest MessagesNewest MessagesNewer Messages

tbutler (apparently) - Sep 9, 2005 7:32 am (#17 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 145
Re: Annoyance level of different ad types

On 9/7/05 at 3:11 PM, mmattynyc.rr.com (Marilyn Matty) wrote:

> On Sep 5, 2005, at 2:56 PM, Edward Reid wrote:
>
> > I agree with that, but I don't buy the argument that the set of
> > models which will work to support online content is identical to
> > the set of models which support print content. What we are seeing
> > so far is that the models for supporting print content don't work
> > very well online. We need to be looking for new models, not
> > insisting on forcing square pegs into round holes.
>
> Without advertising, how will it be possible to give daily global
> news information that is extremely expensive to gather, edit, design,
> archive, research, etc., etc., available for free online? Or to
> create and maintain search engines that are effective? The
> alternative would be a subscription based online newspaper model, or
> subscription based search engines, and leave the rest to bloggers and
> open source.

I think you may be talking past each other here. My impression was not
that he's talking here about alternative models as in advertising vs
non-advertising; I believe he's said some kind of advertising is
probably necessary, and *I* believe it's necessary, little as I like it.
For me, this is about alternative *advertising* models - text ads vs.
graphical vs. rich media.

I know I refuse to buy *anything* linked to a rich media ad, and as far
as I can remember and manage to I will not do business with companies
that do rich media ads. I will seek out and do business with companies
that use text-based advertising to the exclusion of others, when I can;
aside from the ubiquitous Google ads, I can hold up the text ads at the
bottom of the Macintouch web page as an example, or the sponsor links on
tidbits.com. These are the first places I use when I want to buy
something and I don't have a place already in mind.

I will also refuse to use a website where the advertising gets too
intrusive. Weather Underground, which Edward mentioned, used to be my
favorite weather site; however, I've slowly stopped using it in the past
year because its advertising was getting too intrusive for me, and I'm
not willing to reward an advertising offensive by subscribing. Or to put
it another way, I won't let advertising overload push me into
subscribing to a site to get rid of the ads; I'll go somewhere else
first, if I have any choice in the matter.

This isn't to say I reject subscription sites, but it has to be
approached the right way; I'll pay to support a site I like, or to open
up more content, but not to eliminate obnoxious behavior. Overdone
advertising definitely falls in that category.


Travis Butler
tbutlermac.com

mmatty (apparently) - Sep 9, 2005 2:58 pm (#18 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 383
Re: Annoyance level of different ad types



On Sep 8, 2005, at 9:09 PM, Travis Butler wrote:

> On 9/7/05 at 3:11 PM, mmattynyc.rr.com (Marilyn Matty) wrote:
>
>
>> On Sep 5, 2005, at 2:56 PM, Edward Reid wrote:
>>
>>
>>> I agree with that, but I don't buy the argument that the set of
>>> models which will work to support online content is identical to
>>> the set of models which support print content. What we are seeing
>>> so far is that the models for supporting print content don't work
>>> very well online. We need to be looking for new models, not
>>> insisting on forcing square pegs into round holes.
>>>
>>
>> Without advertising, how will it be possible to give daily global
>> news information that is extremely expensive to gather, edit, design,
>> archive, research, etc., etc., available for free online? Or to
>> create and maintain search engines that are effective? The
>> alternative would be a subscription based online newspaper model, or
>> subscription based search engines, and leave the rest to bloggers and
>> open source.
>>
>
> I think you may be talking past each other here. My impression was not
> that he's talking here about alternative models as in advertising vs
> non-advertising; I believe he's said some kind of advertising is
> probably necessary, and *I* believe it's necessary, little as I
> like it.
> For me, this is about alternative *advertising* models - text ads vs.
> graphical vs. rich media.

What I'm trying to say is that it's really not a matter of what you
or I like. What will make a lot of money for companies that own the
search engines, and what will fund, and hopefully begin to turn a
profit for the news services, is what is inevitable.

>
> I know I refuse to buy *anything* linked to a rich media ad, and as
> far
> as I can remember and manage to I will not do business with companies
> that do rich media ads.

The companies who are supplying information that depend on
advertising revenues from rich media, information that is extremely
expensive to develop, archive and deliver. When people use this
information but block the ads, they are denying revenue the companies
need to survive. What is ending up happening is that companies
including the Time Warner, McGraw-Hill, VNU, etc. periodicals will
stop providing their archives for free. And some of the magazines of
these companies have started to give only limited access to archives
to paid print subscribers - those who want full access need to pay
additional fees. As I mentioned, Conde Nast publications like The New
Yorker, Wired Magazine (only the magazine is owned by Conde Nast -
Lycos owns the online news site, and the links to the magazine
articles, etc. is a combination of syndicated and advertising
material), etc. offer extremely limited access to information, while
their food/travel publications, like Gourmet, Bon Appetit, Conde Nast
Traveller offer extensive archives for free because they generate
advertising revenues.

The ultimate result is that because advertising revenues are not
supporting the business model, access to information that was once
free is becoming increasingly limited.

Marilyn

mmatty (apparently) - Sep 9, 2005 2:58 pm (#19 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 383
Re: Annoyance level of different ad types



On Sep 9, 2005, at 10:22 AM, Lucas K. Mathis wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 8.9.2005, space aliens observed mmatty saying:
>
>>> Consider actual research. Jakob Nielsen finds that about 80% of
>>> users
>>> dislike or strongly dislike moving content and blinking; floating
>>> across the screen and playing sound were close behind at about 70%.
>>>
>> Anyone who has such a poorly designed that is so difficult to read
>> and
>> navigate,
>>
>
> I can't see how useit.com could possible be called "difficult to
> navigate". Ugly, I give you that, but definitely not "difficult to
> navigate". Likewise, texts are all black-on-white, and font-size is
> defined to be "100%", so basically, if the text is hard to read on
> your
> computer, it may be that you've configured your browser to show
> text too
> small, which is hardly Nielsen's fault.

The browser defaults are waaaaaay too large to be readable for the
big majority of the population, particularly for articles containing
more than a few short paragraphs of text. These defaults came to be
because of the way Internet Explorer and Windows defined them, not
because they are good. Coding text to 80-85%, or using ems, is the
best way to establish a reasonably readable type size to compensate
for type that is difficult to read for most people and adjustable for
people who have difficulty.

The big majority of internet users do not know how to adjust their
browsers touch their browser defaults, and an even bigger majority do
not want to change them if they do. They also do not want to resize
browser screens when they find line length is too long.

>
>
>
>> and who uses such small samples and restricted methodology in
>> research
>>
>
> Nielsen often writes that it's okay to use small numbers of people for
> incremental usability tests,

It is in his interest to write this. It would be more expensive and
time consuming for him to do things the right way.

> but I've never seen him use small samples
> for such studies. For example, for his "Most Hated Advertising
> Techniques" study, he questioned over 600 people:
> <http://www.useit.com/alertbox/20041206.html>

This sample is way, way, way too small to produce anything resembling
reliable result given the number of people who use the internet. And
the research was not conducted by an independent source, but rather
by eBay and Yahoo, who have a vested interest
in proving that the stuff they said was terrible is bad because it's
not in the portfolio of ad configurations they sell.

>
> For the record, he found that 87% hate blinking ads. Even if his
> studies
> were flawed, it's not hard to believe this number.

If this were actually true, then nobody, especially the large news
organizations, would be selling blinking ads.

But it's also important to draw a distinction between blinking, which
is not necessarily a part of Flash ads, and motion, which usually is.
This article doesn't say anything about Flash or motion per se. I
strongly doubt that it would have since Nielsen went on the
Macromedia payroll in 2002.

Marilyn

tbutler (apparently) - Sep 9, 2005 2:58 pm (#20 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 145
Re: Annoyance level of different ad types

On 9/9/05 at 1:01 PM, mmattynyc.rr.com (Marilyn Matty) wrote:

>
> On Sep 8, 2005, at 9:09 PM, Travis Butler wrote:
>
> > On 9/7/05 at 3:11 PM, mmattynyc.rr.com (Marilyn Matty) wrote:
> >
> > I think you may be talking past each other here. My impression was
> > not that he's talking here about alternative models as in
> > advertising vs non-advertising; I believe he's said some kind of
> > advertising is probably necessary, and *I* believe it's necessary,
> > little as I like it. For me, this is about alternative
> > *advertising* models - text ads vs. graphical vs. rich media.
>
> What I'm trying to say is that it's really not a matter of what you
> or I like. What will make a lot of money for companies that own the
> search engines, and what will fund, and hopefully begin to turn a
> profit for the news services, is what is inevitable.

So you've said in prior messages. What I and others are arguing is that
rich media ads are *counterproductive*; they turn away customers,
instead of attracting them. You have argued that they are more
effective, but this goes against my own experience, the experience of
everyone I have ever discussed this with, several other people chiming
in on this thread, and at least one study cited by someone else on this
thread. Forgive me, but your arguments simply haven't convinced me.

> > I know I refuse to buy *anything* linked to a rich media ad, and as
> > far as I can remember and manage to I will not do business with
> > companies that do rich media ads.
>
> The companies who are supplying information that depend on
> advertising revenues from rich media, information that is extremely
> expensive to develop, archive and deliver.

I think we're still talking past each other. I *agree* that advertising
revenues are probably necessary, and that also seems to be the consensus
in this thread. However, you keep conflating the need for advertising
with a need for rich media advertising and using the first to argue the
second - and this is what I and others dispute. Accepting that
advertising is necessary does not prove that rich media ads are
necessary, and as I said the rich media-specific evidence you've cited
has not convinced me.

> When people use this
> information but block the ads, they are denying revenue the companies
> need to survive.

And this is a perfect example of why I contend that rich media ads are
counterproductive. Are there products for blocking text ads? Do you see
people trading tips for viewing pages without text ads? Is there an item
on the Safari menu saying 'Block text ads' as there is for 'Block Pop-up
Windows'? Text ads do not offend people, at least not to the extent of
rich media ads, and therefore do not suffer the backlash that rich media
ads do.

Also, you're again conflating the issue of 'Companies need advertising
to survive' with rich media ads. The two issues are not the same, and I
can accept the need for advertising while continuing to fight
advertising that is intrusive to the point of being offensive.

> What is ending up happening is that companies
> including the Time Warner, McGraw-Hill, VNU, etc. periodicals will
> stop providing their archives for free. And some of the magazines of
> these companies have started to give only limited access to archives
> to paid print subscribers - those who want full access need to pay
> additional fees. As I mentioned, Conde Nast publications like The New
> Yorker, Wired Magazine (only the magazine is owned by Conde Nast -
> Lycos owns the online news site, and the links to the magazine
> articles, etc. is a combination of syndicated and advertising
> material), etc. offer extremely limited access to information, while
> their food/travel publications, like Gourmet, Bon Appetit, Conde Nast
> Traveller offer extensive archives for free because they generate
> advertising revenues.
>
> The ultimate result is that because advertising revenues are not
> supporting the business model, access to information that was once
> free is becoming increasingly limited.

All of this is unfortunate, yes. But it still only proves the need for
advertising in general, not the need for rich media ads.


Travis Butler
tbutlermac.com

chuck goolsbee (apparently) - Sep 9, 2005 2:58 pm (#21 Total: 36)  

Reply to this message
via email - chuck goolsbee  

Photo of Author
Posts: 433
Re: Annoyance level of different ad types

>
>>And the search engines need advertising dollars, and in order to keep the
>>revenues growing, they will have to come up with methods and avenues to
>>deliver ads.
>
>The only reasons I know of that a search engine needs to increase revenues are
>
> 1) keep up with growth of the web

Which has *significant* beind the scenes cost. Significant CAPITAL and EXPENSE cost. A "search engine" is, at the core a physical asset consisting of server farms in datacenters. My undserstanding is that Google runs many many such datacenters.

Think about:

* Bandwidth costs.

We run a "small" Internet Datacenter. That is, around 1000 servers. Our bandwidth bills are in the 10s of thousands of dollars per month.

Google's server farms are likely thousands of times larger than ours, and frankly probably exceed our bandwidth usage by a significant margin. Yes, bandwidth gets cheaper per-megabit the more you buy, but the numbers don't get smaller.


* Electricity and cooling costs

Ditto on the above scales. Our facility has 75 tons of HVAC capacity. Our power consumption is measured in 10s of thousands of Kilowatt hours per month. We maintain a backup power system that is measured in Megawatts. Diesel fuel, even non-road taxed Diesel is well over $2 a gallon and our genset sucks down 40 gallons per hour (when we run it.)

Again, scale these numbers by thousands if you are contemplating an operation the size of Google.


* Real estate costs

Datacenters must be located near reliable power sources with easy access to fiber optic lines. Such real estate is usually expensive to acquire or lease. Datacenters require a minimum of $1500 a square foot to build (and that assumes an existing structure, more if you start from bare real estate. I know, I have built two of them in the past 5 years.) The more redundancies you put into them the cost just rises at, usually at an exponential rate. UPS units, Generators, Power Distribution Units, HVAC and Humidity controls, cabling, ladder rack, server racks/cabinets, FM200 fire suppression equipment, Routers, Ethernet switches, etc. Labor costs for construction, electrical, and UPS installation. Replacing thousands of UPS batteries every three years.


The costs of these items is not trivial.


--
Chuck Goolsbee V.P. Technical Operations
_________________________________________________________________
digital.forest Phone: +1-877-720-0483, x2001
where Internet solutions grow Int'l: +1-206-838-1630
*** celebrating eleven years of service 7/12/1994 - 7/12/2005 ***
12101 Tukwila International Blvd Fax: +1-206-838-3749
Suite 410 http://www.forest.net
Seattle, WA 98168 email: cgforest.net

tekelenb (apparently) - Sep 9, 2005 9:10 pm (#22 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 263
Re: Annoyance level of different ad types

At 14:58 -0700 UTC, on 2005/09/09, Marilyn Matty wrote:

[...]

> The ultimate result is that because advertising revenues are not
> supporting the business model, access to information that was once
> free is becoming increasingly limited.

No, the *ultimate* result is that 'products' like information will very
simply be funded differently[*]. Ad-based funding is just one model. Another
model is for consumers to just pay directly, instead of indirectly though
ads; with the added bonus of actually getting what you pay for. A real world
variant on that is tax-based funding, which allows for the information to be
available to not only the lucky few who can afford it. (Has an inherent risk
of being manipulated by the organisation overseeing that cash flow, but
that's no different with ad-based funding, where plenty of 'information' is
adjusted to better suit advertisers.)

Yeah, yeah, there's probably somebody yelling "commie!" right now :) Goes to
show how informative ad-based information is ;)


[*] Provided there is a market of course. But if not, the funding model is
irrelevant anyway.


--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

tekelenb (apparently) - Sep 9, 2005 9:10 pm (#23 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 263
Re: Annoyance level of different ad types

At 14:58 -0700 UTC, on 2005/09/09, Marilyn Matty wrote:

[...]

> Coding text to 80-85%, or using ems, is the
> best way to establish a reasonably readable type size to compensate
> for type that is difficult to read for most people and adjustable for
> people who have difficulty.

If those who have difficulty with small size can adjust it, why can't those
who have difficulty with large type? Where's the logic?


--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

edward (apparently) - Sep 12, 2005 3:39 pm (#24 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 269
Re: Annoyance level of different ad types

At 02:58 PM 09/09/2005 -0700, chuck goolsbee wrote:
>The costs of these items is not trivial.

Chuck,

I'm not disagreeing with this in the slightest. My emphasis was on
proportion. I expect that both Google's expenses and their revenues will
grow roughly in proportion to web use, and that this will happen without
changes in their advertising models. This was in response to a claim that
they had to change their advertising models because of growing expenses,
which did not take into account that the same factors leading to increased
expenses also lead to increased revenues.

The proportion is certainly very rough. Perhaps Google-type ads have
already passed their sweet point and the revenues will not grow in
proportion to web use. I tend to think it's still short of that sweet
point. But I don't think anyone can say for sure.

Edward
Art Works by Melynda Reid: http://paleo.org

LKM (apparently) - Sep 12, 2005 3:39 pm (#25 Total: 36)  

Reply to this message
via email - Lucas K. Mathis  

Photo of Author
Posts: 80
Re: Annoyance level of different ad types

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 9.9.2005, space aliens observed Marilyn Matty saying:
>Coding text to 80-85%, or using ems, is the best way to establish a
>reasonably readable type

It's also the best way to make sure that those who changed their default
font size can't read the text on your site.

People set the font size to what they like. Claiming that what people
like isn't readable for them is, well, certainly an interesting
standpoint.


>>>and who uses such small samples and restricted methodology in
>>>research
>>Nielsen often writes that it's okay to use small numbers of people
>>for incremental usability tests,
>It is in his interest to write this. It would be more expensive and
>time consuming for him to do things the right way.

I suspect you don't read what he writes. He advocates using small
samples if you don't have the budget to hire him. He himself uses larger
samples for his tests.

Having done usability tests myself, I can attest that he's right. Even
if you only test with 5 people, you'll find most of your usability
problems, and you're very likely to find all of the serious ones.


>>but I've never seen him use small samples for such studies. For
>>example, for his "Most Hated Advertising Techniques" study, he
>>questioned over 600 people:
>><http://www.useit.com/alertbox/20041206.html>
>This sample is way, way, way too small to produce anything resembling
>reliable result given the number of people who use the internet.

You obviously missed your statistics classes. The amount of people on
the internet is irrelevant in this context. If you have a sample of 600
people, that's more than enough to get statistically valid results, no
matter how big the source you took your sample from is.

The probability that you survey 600 randomly selected people from a
source that, in general, loves blinking ads, and then find out that 87%
of the people you surveyed hate blinking ads is extremely small. The
size of the source you select these people from has no influence on
this. I haven't calculated what the confidence level is when surveying
600 people and finding out that 87% of them hate blinking ads, but
similar studies have confidence levels of 95% while surveying less than
300 people of a very large population.


>>For the record, he found that 87% hate blinking ads. Even if his
>>studies were flawed, it's not hard to believe this number.
>If this were actually true, then nobody, especially the large news
>organizations, would be selling blinking ads.

I think you've said it yourself: They'll sell whatever makes them money.

lucas

- --
THE LAW OF ENTROPY:
The perversity of the universe tends towards a maximum.

-----BEGIN PGP SIGNATURE-----
Version: PGP SDK 3.2.2

iQA/AwUBQyKtYc0zN2kKTjB0EQLUdQCfePid2TRpGTo/TJTrgD+HCZ/6qCYAn3e+
96rKn4SR3tOr3PLK43tg3JAF
=iBOm
-----END PGP SIGNATURE-----

mmatty (apparently) - Sep 12, 2005 3:39 pm (#26 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 383
Re: Annoyance level of different ad types

On Sep 10, 2005, at 12:10 AM, Sander Tekelenburg wrote:

> At 14:58 -0700 UTC, on 2005/09/09, Marilyn Matty wrote:
>
>> Coding text to 80-85%, or using ems, is the
>> best way to establish a reasonably readable type size to compensate
>> for type that is difficult to read for most people and adjustable for
>> people who have difficulty.
>>
>
> If those who have difficulty with small size can adjust it, why
> can't those
> who have difficulty with large type? Where's the logic?
>

The advantage to using ems or % in style sheets is that the font size
can be adjusted either larger or smaller by just about any modern
browser. Here's the W3C recommendations:

http://www.w3.org/TR/REC-CSS1#font-size

And a more "plain English" explanation:

http://usabletype.com/styles/sizes/

Using terms for designating font size will also work, but it's not as
reliable across platforms and browsers. There's a technique for
creating alternative style sheet here:

http://www.alistapart.com/articles/relafont

The 16 point browser default (which was pretty forced on the browser
developers by Microsoft, who has never been good with type or
understood that legibility and readability are two different things)
is way to large to be readable by the vast majority of the
population. And most people do not know how to, or want to, adjust
their browsers. he number of general visitors who who will take the
time and effort to adjust their browsers. People who are not
comfortable with reading will leave, and often they will not even
know why or what's making them uncomfortable.

Marilyn


tekelenb (apparently) - Sep 12, 2005 3:39 pm (#27 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 263
Re: Annoyance level of different ad types

At 13:00 -0400 UTC, on 2005/09/10, Marilyn Matty wrote:

> On Sep 10, 2005, at 12:10 AM, Sander Tekelenburg wrote:

[...]

>> If those who have difficulty with small size can adjust it, why
>> can't those
>> who have difficulty with large type? Where's the logic?

The question remains.

> The advantage to using ems or % [...]
>
> http://www.w3.org/TR/REC-CSS1#font-size

I know all about em, ex and %. CSS has nothing to do with this. What this is
about is whether or not to mess with a user's default font size *at all* and
why. (Obviously, my view is you should leave the main text for what it is (no
need even to specify 1em) and let the user decide what font size is
comfortable for him.)

If you think IE users are incapable of figuring out how to define a default
font-size that is comfortable to them, you can use this bit if WinIE
proprietary code in the HEAD of your HTML documents:

<!--[if lt IE 7]>
<STYLE>
BODY {font-size: small}
</STYLE>
<![endif]-->

It will mess up things for IE users who *did* configure their browser to use
a font size they can read, and those for whom the default is confortable. But
I suppose if you can afford to lose their money, that's fine.

At least this way you don't screw up things for non-IE users. People who use
a different browser probably made a more conscious choice and are thus more
likely to be able to figure out how to configure their tools. Even if they
need help, they're likely to know where to find tha1t, given that they were
able to find enough information to choose to use something better than IE.

> And a more "plain English" explanation:
>
> http://usabletype.com/styles/sizes/

Not interesting. It makes claims without pointing to sources ("The default
font size for browsers has been standardized to the value of 16 pixels"); it
simply assumes it's the Web designers's task to 'fix' what really is a
browser problem; it completely ignores the fact that not everybody's eyesight
is that of a healthy 12-year old - it even flat out claims that "9px
[...is...] legible" ; it at one point says many changed their default font
size and at another that many don't - which underlines my question, which
your answer doesn't address.

[...]

> The 16 point browser default (which was pretty forced on the browser
> developers by Microsoft, who has never been good with type or
> understood that legibility and readability are two different things)

Agreed.

> is way to large to be readable by the vast majority of the
> population. And most people do not know how to, or want to, adjust
> their browsers.

If Web developers would not mess with user's default font sizes (for the main
text) it would take those users no more than a few minutes to figure out they
can change that default. By hiding bad browser defaults (and bugs), Web
designers only contribute to the mess. They make users think there is nothing
wrong with their browser, which in turn means there is no incentive to the
browser developer to fix things. Which means the Web designer will 'need' to
continue with this stupid practice of breaking things in order to hide other
breakage.

[...]

> People who are not
> comfortable with reading will leave, and often they will not even
> know why or what's making them uncomfortable.

Right. So the question remains: why is that different for those who don't
know how to make the default font size smaller and those who don't know how
to make the default font-size bigger?


--
Sander Tekelenburg, <http://www.euronet.nl/~tekelenb/>

edward (apparently) - Sep 12, 2005 3:39 pm (#28 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 269
Re: Annoyance level of different ad types

[Quotes trimmed to reminder size only, to make following the discussion easier]

==================================================

Section I: Motion, color, perception, interference

At 02:58 PM 09/09/2005 -0700, Marilyn Matty wrote:
>What I'm trying to say is that it's really not a matter of what you or I
>like. [...] And as I mentioned, advertising doesn't have to be loved for
>it to be effective - like the Mr. Whipple ads I mentioned.

And what I keep saying is that it's not a matter of like or dislike, but a
matter of interfering or not interfering. You keep bringing up Mr Whipple,
but Mr Whipple didn't interfere with people seeing the TV show. Had Mr
Whipple been overlaid on the TV show, he'd have been sent out to right
field. Moving ads DO INTERFERE with reading the content, and they interfere
greatly.

As Travis and others here have pointed out, I don't see anyone around here
saying advertising is inherently objectionable. It's objectionable when it
interferes with access to content, and that interference will be
detrimental even to advertising in the long run anyway because the
advertiser isn't cooperating with the content provider.

>But it's also important to draw a distinction between blinking, which is
>not necessarily a part of Flash ads, and motion, which usually is.

Blinking and moving have very similar effects on peripheral vision. There's
no need to draw any distinction in a discussion of interference with
perception. A large area blinking is just the extreme of motion.

>Color is extremely distracting - it causes chemical changes in the brain
>and in the eye and elicits physical reaction equal to that of movement.

Oh, heavens. No, color outside the fovea is nowhere nearly as distracting
as motion. Color IN the fovea has a strong effect, but outside the fovea
it's motion that's far more powerful. Only the fovea has an abundance of
the cone cells to detect color. This effect is extremely well known: if you
want to detect small movements, don't look directly where you expect them.

And guess what's focused our fovea when we're reading? The words! The
content! Once we look at the ads, then yes they are on the fovea and the
color is very important. But while we have the content on the fovea, it's
motion that's far more powerful on the rest of the retina.

>This sample is way, way, way too small to produce anything resembling
>reliable result given the number of people who use the internet.

The samples are plenty big enough. Most people don't know enough statistics
to realize that samples which seem intuitively rather small can give "good
enough" results when the margin of error doesn't have to be extremely
tight. Even for tight margins of error, the needed sample sizes for a large
population seem intuitively very small. The sample sizes used are good
enough for the kind of data involved, which isn't precise in its nature.

==================================================

Section II: Revenue models

>The ultimate result is that because advertising revenues are not
>supporting the business model, access to information that was once free is
>becoming increasingly limited.

That information has been free for only a few short years. Perhaps making
access a bit more difficult will help public libraries. It's hardly the end
of the world. Placing moving ads on the page with the information makes the
information inaccessible anyway, so if the only choices are free
information with moving ads or no free information, it's six of one and
half a dozen of the other. We'll only get what we think is valuable enough
to pay for. What a concept.

>If this were actually true, then nobody, especially the large news
>organizations, would be selling blinking ads.

Your faith in the ability of the publishers to know what's good for them is
touching, but I find it badly misplaced. Lacking good data, they are simply
accepting money and placing the ads requested. They have no information on
how many users avoid their site because of particular forms of advertising.
(Not even Nielsen claims to know this AFAIK, and it's a much more difficult
problem. It's easy to watch users leave a site, but finding out where they
never go to begin with is much harder.) Heck, these are the same publishers
who originaly put their content on the web for free with no idea how to
make it profitable.

>Without advertising, how will it be possible to give daily global news
>information that is extremely expensive to gather, edit, design, archive,
>research, etc., etc., available for free online? Or to create and maintain
>search engines that are effective? The alternative would be a subscription
>based online newspaper model, or subscription based search engines,

Gosh, horrors. Then I might have to pay for Diabetes Care, Backpacker, The
New Yorker, Adventure Cyclist, CACM, Consumer Reports, Queue, Science,
Clinical Diabetes, IEEE Spectrum, American Scientist ... oh wait a minute,
I'm already paying for all of them. And most of those (9 of 11) get their
revenue primarily from subscriptions and memberships rather than from
advertising. Actually some of these represent models of almost-free content
in print, due to the fact that academic researchers basically donate their
work to scientific journals, and even pay for the privilege.

But I have to say it again: no one here has expressed any desire to
eliminate advertising -- only advertising which interferes with access to
content. Sometimes this is expressed as "annoying" (though Glenn didn't use
that term in his article), but when you probe you find that the real issue
is interference with access to content.

>This is comparing apples to pizza. Neither of the above [Weather
>Underground or MyComicsPage] is exactly taking the market by storm

Almost nobody is taking the market by storm. There are a few 800 pound
gorillas on the web, but past those few, it's all the rest.

My point is not the current balance but the direction of motion. Five years
ago people were saying you could never charge for online content, and I'm
showing that there is motion -- albeit slow -- toward being able to charge
for content. The fact that the LA Times failed last year doesn't mean it
would fail again a couple of years in the future.

A huge issue is that at present, most paid online content requires a rather
hefty subscription. One reason Weather Underground and MyComicsPage can
charge is that the annual fee is small.

But consider another example often discussed here: iTMS! Why is it so
successful? After all, iTMS is charging for online content, even though
it's "keeper" content. One reason for the success is that no commitment is
required (a modern site ;-), with a one-dollar minimum. If I'd had to shell
out $20 for 25 songs, I wouldn't have made an iTMS purchase yet, but in
fact I've spent nearly $25 there.

Suppose the LA Times, instead of requiring annual subscriptions, had been
able to provide an option for daily purchase for, say, ten cents or 25
cents, enough that an annual subscription would be a hefty price break but
still less than buying a print paper. Would people have tried it out? Sure
they would! Whether this would translate into a sufficient revenue stream
remains an open question, but it fits how people approach content in the
print world: buy a single copy at a newsstand, then maybe subscribe, or
maybe just continue buying occasional single copies.

So to my mind, a really critical issue in paying for online content is
developing the ability to pay readily for small chunks.

But again: there IS motion in the direction of willingness to pay for
online content.

>And more and more publications are having to charge for archives,
>including most of the Time Inc. publications, Business Week, etc.

So what? Until recently, the only way to access these archives was to visit
a public library or pay to order back issues. Recent years have been
experimentation. If we end up with online access to archives for a fee,
that's more we had before. We still have libraries; they have suffered but
haven't died. Print periodicals still sell back issues.

>Conde Nast won't give away much editorial from the New Yorker

Take a look at the latest issue online of The New Yorker,
http://newyorker.com. They're giving away about 2/3 of the current content
(but no archives, and the only ads are their own). I only wish they had the
September 12 cover online -- it's the saddest one I've ever seen.

>It [the new NYT plan] will not be for all online archives

What they are saying on the web site is "We ... will be offering our
TimesSelect service this fall which will include access to our Archive. We
will be finalizing the details of TimesSelect in the coming weeks". That
seems to say all archives (presumably back to 1996, beyond which it's a
different game), but obviously this is not a firm commitment, so we'll just
have to wait for the details.

>They are doing this [selling subscriptions] because they can command much,
>much higher prices by delivering very targeted ads.

Then doesn't this indicate where the money is? In delivering ads that
people actually want, rather than trying to force them using motion to look
at ads while they are trying to digest the site's content? It goes back to
the fact that productive advertising is a cooperative endeavor between the
advertiser, the consumer, and the publisher, and if the advertiser doesn't
respect the consumer then the ads will be far less effective.

>Because so much information, entertainment and services are available for
>free, people do not want to pay for content.

True, people have never liked to pay. But people do like convenience.
That's why public libraries never killed bookstores.

There are other examples. Software is one. It's possible to run a computer
very well with free (as in beer) software, but most people find that many
commercial software packages (not all) provide features, convenience, or
reliability that's worth the price.

If/as the best content moves to small charges, people will get used to
paying. Again, if we can establish the ability to purchase online content
in small chunks, that will greatly accelerate acceptance.

>You are very unusual in your willingness to pay for online content.

But again, the general public reluctance is due to seeing free
alternatives. If the free alternatives drive out the paid, then it's
because the public (with its finely honed critical skills and appreciation
of fine work ... hmm) has judged the free a preferable alternative.

People have been willing to pay for content for decades (and to a much
lesser extent, for centuries). Most popular print publications have always
been supported by a combination of subscription and advertising revenue.
Various aspects of the web have distorted this willingness to pay in
various ways.

But consider yet another example: NetZero. OK, it's not content. But
NetZero offered free Internet access in return for on-screen advertising.
It flopped -- not in favor of free unhampered access, but in favor of paid
access. People found it preferable to pay. Yes, NetZero still offers this
deal, but only for ten hours/month, which is probably more of a loss leader
than a profit center. They have morphed into just a niche ISP with a catchy
name. They went from ad-supported to user-paid because that's what people
preferred. People can prefer to pay when they judge all the factors.

==================================================

Section III: Fairness

>When people use this information but block the ads, they are denying
>revenue the companies need to survive. [...] When you look at an online
>newspaper or magazine and block the ads, you are denying that company
>revenue that funds the content it produces and delivers. And because most
>newspapers and magazines [lose] money on their online content, you're
>denying them an important avenue that could lead to profitability.

I'm not blocking anyone else from seeing the ads. I'm not blocking the
company from assigning an ad to the page they serve me.

(In fact, depending on how they count, I suspect that ads are often counted
as displayed once the page linking to them is sent to the user, whether the
user's browser actually downloads the ad. Possibly you can answer whether
this is correct.)

I don't owe the company revenue or survival. I'm paying them exactly what
they asked me to -- same as I do with print publications, both the
partially ad-supported and the wholly subscription-supported. They asked me
to consider reading the ads they offered; I considered, and I read some ads
and declined to read certain ads.

Finally, I repeat: I'm not aware of any situation in which fairness demands
that I actually pay attention to any particular ad. I'm not morally
obligated to read every billboard on the highway. I'm not obligated to read
even one ad in any newspaper or magazine that I subscribe to. I'm not
obligated to read the TidBITS sponsor ads in the newsletter. I'm not
obligated to read every ad in the online versions of the Times or the
Times-Picayune, the Post or the Post-Intelligencer. I do read some of the
ads, as some have value for me, and some are amusing. But I can't see where
the method by which I choose to skip particular ads is (morally) of any
concern as long as that method affects only me.

>The consumer can leave if they don't want to view the ads, or they can not
>refer to a publication they find offensive. If it is that offensive, there
>will not be enough of an audience to sell ads to. But in the case of the
>publications we've been discussing, there are big, viable and growing
>audiences that are very interested in the content.

And those audiences are trying hard to reach the content by blocking
certain kinds of ads which interfere with reaching the content. Not all of
them are fighting the interference, but a lot of that is because the
majority don't yet know it's possible to fight it. You point to the growing
audience, but as I've pointed out repeatedly, we have no idea how much
faster that audience might be growing, or how much larger it might be now,
without some of the inhibiting factors currently present (of which moving
ads are only one factor).

>While it's easy to block ads, the question is whether it is ethical to do
>so, esp. when receiving content for free. Magazines and newspapers are
>having a very difficult time online.

I didn't ask for the content to be free. That's just how it came. Perhaps
it's useful to review how this situation came to be:

On the first day, Vint created the Internet and Steve created AOL, and
darkness was on the face of the wall between them.

On the second day, the universities found the Internet and a small slice of
the public found AOL, but the darkness between them remained.

On the third day, the wall between AOL and the Internet was breached, and
brother and sister could once again speak to each other.

On the fourth day, Tim invented the World Wide Web, and the universities
started making content available via the WWW while everybody else ignored it.

On the fifth day, AOL and the advertisers saw that the WWW was a Profitable
Thing, and the public gained other means of access to the Internet, and
light was on the face of the net.

On the sixth day, the commercial interests saw that the WWW might be the
Next Big Thing but didn't know what to do with it, and entered into
contracts with whatever angel or devil they could find to help them partake.

On the seventh day, nobody rested, the users because of information
overload and the commercial interests because they still hadn't figured out
the New Demon they had made a contract with (most of the presumed angels
having turned out to be in camouflage).

So that's where it stands. Magazines and newspapers are having a hard time
online because they went online not knowing (since no one knew) how to deal
with publishing online. They offered their content free because it was the
only thing they could figure out how to do online which would draw visitors
(and at the time they were probably right). But it WAS their choice to get
online and offer free content.

>Most card stock is for subscription solicitation - in-house advertising.

A lot of it yes, but far from "most". If you'd like, I can save a month of
mine for you to analyze.

>And if you're ripping it, you're looking at it anyway.

You obviously don't know me. ;-) Oh, of course I looked at it once, but
then I knew what was in it, and I stopped looking. Perhaps I get a glimpse,
but it's maybe 5% or less of the exposure that I get to ads printed on the
same pages with the editorial content of articles I read. I also rip out
thick multipage advertising supplements from magazines after looking only
at the first page, and dump the advertising inserts in newspapers usually
without examining the stack at all.

I see no moral difference between avoiding an ad 95% and avoiding it 100%.
If you see a problem with my blocking certain kinds of online ads at a 100%
rate, then I have to conclude that you'd see a problem with my blocking
certain kinds of print ads at a 95% rate. And if you see no problem with my
blocking certain kinds of print ads at 95%, then I don't see how you can
object to my blocking certain kinds of online ads at 100%.

==================================================

I seem to notice a significant circular motion in this discussion, a
slippage toward content-free in place of free content, so I'm probably
going to let it lie no matter what response I get.


[Yes, I think it's time to wind this down, please. -Adam]


Edward
Art Works by Melynda Reid: http://paleo.org

Dan Frakes (apparently) - Sep 13, 2005 9:58 am (#29 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 878
Re: Annoyance level of different ad types

On 9/9/05 2:58 PM, "Marilyn Matty" wrote:
>> For example, for his "Most Hated Advertising
>> Techniques" study, he questioned over 600 people:
>> <http://www.useit.com/alertbox/20041206.html>
>
> This sample is way, way, way too small to produce anything resembling reliable
> result given the number of people who use the internet.

This is not necessarily true. A sample size of 600 Internet users has the
potential to be *quite* accurate, even given the large number of people
using the Internet. For example, here's a layman's sample size calculator:

<http://www.surveysystem.com/sscalc.htm>

For example, given a population of 200 million people, if you want a
confidence level of 95% and a confidence interval of +/- 4%, you need a
sample size of, conveniently enough, 600.




>> For the record, he found that 87% hate blinking ads. Even if his studies were
>> flawed, it's not hard to believe this number.
>
> If this were actually true, then nobody, especially the large news
> organizations, would be selling blinking ads.

The true measure of an Internet ad's effectiveness is not how many people
dislike it, but how many of the people that *don't* dislike it actually
click through it. If a particular ad on CNN is hated by 95% of readers, but
the other 5% *always* click through, that would likely be a very profitable
ad ;-)



mmatty (apparently) - Sep 13, 2005 11:04 pm (#30 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 383
Re: Annoyance level of different ad types



On Sep 9, 2005, at 5:55 PM, Lucas K. Mathis wrote:

>>
>>
> I suspect you don't read what he writes. He advocates using small
> samples if you don't have the budget to hire him. He himself uses
> larger
> samples for his tests.

This is fine for small projects for small sites, but the article you
mention uses a small sample to project up for the entire universe of
web users. And 600 people in an unscientifically conducted study by
interested parties is nothing that can be considered in any way valid.

"God speaks to Brady, and Brady tells the world world!"

>>> but I've never seen him use small samples for such studies. For
>>> example, for his "Most Hated Advertising Techniques" study, he
>>> questioned over 600 people:
>>> <http://www.useit.com/alertbox/20041206.html>
>>>
>> This sample is way, way, way too small to produce anything resembling
>> reliable result given the number of people who use the internet.
>
> You obviously missed your statistics classes.

[Let's avoid the snarky comments and stick to substantive positions, mm'kay? -Andrew ]


I not only worked extensively with statistics for many publishing and
media companies for many years, including Nielsen, Scarborough, MRI,
SMRB, and Erdos & Morgan. I also worked directly with Paul Erdos on
about 7 projects for 3 magazines. When the consumer electronics
industry began to take off in the late 70's/early 80's with the
advent of video recorders, etc., Mr. Erdos was impressed enough with
me from my work at Ms. and Texas Monthly to ask me to be part of an
advisory panel on questions to be included in their magazine study.

Marilyn


mmatty (apparently) - Sep 13, 2005 11:04 pm (#31 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 383
Re: Annoyance level of different ad types



On Sep 13, 2005, at 12:58 PM, Dan Frakes wrote:

> On 9/9/05 2:58 PM, "Marilyn Matty" wrote:
>
>>> For example, for his "Most Hated Advertising
>>> Techniques" study, he questioned over 600 people:
>>> <http://www.useit.com/alertbox/20041206.html>
>>>
>>
>> This sample is way, way, way too small to produce anything
>> resembling reliable
>> result given the number of people who use the internet.
>>
>
> This is not necessarily true. A sample size of 600 Internet users
> has the
> potential to be *quite* accurate, even given the large number of
> people
> using the Internet. For example, here's a layman's sample size
> calculator:
>
> <http://www.surveysystem.com/sscalc.htm>
>
> For example, given a population of 200 million people, if you want a
> confidence level of 95% and a confidence interval of +/- 4%, you
> need a
> sample size of, conveniently enough, 600.

For a survey to be considered close to accurate, it must be at least
+/- 3%; 4% is unacceptable. Given a population of internet users of
about 200 million US internet users (I could be wrong about this, but
I'm fairly sure I read this recently), you need a sample of about
1,500. For the study to be global, the number would have to be much
larger.

There are other factors to consider as well - how the sample was
compiled (and the survey Nielsen quotes was taken by eBay and Yahoo
(I'm assuming that they are teaming up with eBay in regard to Yahoo
auctions and stores, but there are no details about it, so we can't
be sure), who are not uninterested or independent parties, how the
questions were asked, the data collected, etc. The article doesn't
give any information about where the data actually comes from.
>
>
>>> For the record, he found that 87% hate blinking ads. Even if his
>>> studies were
>>> flawed, it's not hard to believe this number.
>>>
>>
>> If this were actually true, then nobody, especially the large news
>> organizations, would be selling blinking ads.
>>
>
> The true measure of an Internet ad's effectiveness is not how many
> people
> dislike it, but how many of the people that *don't* dislike it
> actually
> click through it. If a particular ad on CNN is hated by 95% of
> readers, but
> the other 5% *always* click through, that would likely be a very
> profitable
> ad ;-)

This is true if you assume a direct response model, but image
advertising is becoming increasingly important online.

Marilyn

bitreader (apparently) - Sep 13, 2005 11:04 pm (#32 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 120
Re: Annoyance level of different ad types

On 9/13/05 at 9:58 AM, DanFrakes.org (Dan Frakes) wrote:

>This is not necessarily true. A sample size of 600 Internet users
>has the potential to be *quite* accurate, even given the large
>number of people using the Internet. For example, here's a layman's
>sample size calculator:

><http://www.surveysystem.com/sscalc.htm>

>For example, given a population of 200 million people, if you want
>a confidence level of 95% and a confidence interval of +/- 4%, you
>need a sample size of, conveniently enough, 600.

There is at least one condition required to make the comments above regarding sample size valid which I haven't seem mentioned. That is the sample must be a randomly selected sample from the population it is intended to represent. For example, if I wanted to determine how the population using the internet felt about some particular item, I would be seriously misleading if I thought responses from colleagues (other engineers employed by Boeing) would be representative regardless of sample size.

edward (apparently) - Sep 14, 2005 10:25 pm (#33 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 269
Re: Annoyance level of different ad types

At 11:04 PM 09/13/2005 -0700, Marilyn Matty wrote:
>For a survey to be considered close to accurate, it must be at least +/-
>3%; 4% is unacceptable.

Please specify the type of survey you are talking about. There's simply no
fixed cutoff point for acceptability. When you're talking human psychology
(as here), there's no such thing as 3% accuracy; the underlying mechanisms
are too variable for 3% accuracy to be meaningful. You're lucky if you can
get repeatable results within 10% -- not due to chance variation but due to
the inherent variability of the phenomena. So asking for 3% is worthless;
asking to limit the chance variation to 10% will keep it well below the
inherent variability.

Edward
Art Works by Melynda Reid: http://paleo.org


mmatty (apparently) - Sep 14, 2005 10:25 pm (#34 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 383
Re: Annoyance level of different ad types



On Sep 14, 2005, at 12:43 PM, Edward Reid wrote:

> At 11:04 PM 09/13/2005 -0700, Marilyn Matty wrote:
>
>> For a survey to be considered close to accurate, it must be at
>> least +/- 3%; 4% is unacceptable.
>>
>
> Please specify the type of survey you are talking about. There's
> simply no fixed cutoff point for acceptability. When you're talking
> human psychology (as here), there's no such thing as 3% accuracy;
> the underlying mechanisms are too variable for 3% accuracy to be
> meaningful. You're lucky if you can get repeatable results within
> 10% -- not due to chance variation but due to the inherent
> variability of the phenomena. So asking for 3% is worthless; asking
> to limit the chance variation to 10% will keep it well below the
> inherent variability.
>


Confidence level and margin of error are very important
considerations, and the advertising and journalistic industry
standard is a confidence level of 95% and a max +/- 3%

I found some formulas here:

http://www.resolutions.co.nz/sample_sizes.htm

It is also imperative to consider any possible bias that may have
entered the study, and I do question how viable a study conducted
directly by eBay & Yahoo would be, esp. since the article mentioned
gives absolutely no details about what was measured and how the
questions were asked, and who the respondent were.


Marilyn

edward (apparently) - Sep 14, 2005 10:25 pm (#35 Total: 36)  

Reply to this message
via email  

Photo of Author
Posts: 269
Re: Annoyance level of different ad types

At 03:04 PM 09/14/2005 -0400, Marilyn Matty wrote:
>Confidence level and margin of error are very important considerations,
>and the advertising and journalistic industry standard is a confidence
>level of 95% and a max +/- 3%

That level simply makes no sense. Almost everything measured in advertising
and journalism doesn't even have a definition that precise. This is not a
criticism of those industries, just a statement of how little is known
about human behavior.

>It is also imperative to consider any possible bias that may have entered
>the study, and I do question how viable a study conducted directly by eBay
>& Yahoo would be, esp. since the article mentioned gives absolutely no
>details about what was measured and how the questions were asked, and who
>the respondent were.

Most web studies, and certainly all of Nielsen's studies that I've read
about, do suffer from selection bias. It's very difficult to ascertain the
effect of this bias, but the selection is unquestionably non-random. This
however is a problem quite distinct from sample size.

I've love to see an adequate study of web user behavior. AFAIK it doesn't
exist. Nielsen is the best we have -- he does good studies within the
restraints, and his arguments are very sound based on what we do know of
human psychology and behavior.

Edward
Art Works by Melynda Reid: http://paleo.org


LKM (apparently) - Sep 14, 2005 10:25 pm (#36 Total: 36)  

Reply to this message
via email - Lucas K. Mathis  

Photo of Author
Posts: 80
Re: Annoyance level of different ad types

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 14.9.2005, space aliens observed mmatty saying:
>For a survey to be considered close to accurate, it must be at least
>+/- 3%; 4% is unacceptable. Given a population of internet users of
>about 200 million US internet users (I could be wrong about this, but
>I'm fairly sure I read this recently), you need a sample of about
>1,500. For the study to be global, the number would have to be much
>larger.

Again, the size of the population is irrelevant when calculating the
margin of error. I actually looked up the formula, it's

    1 / (sqrt(number of people in sample))

Of course, this assumes that you select the people you survey randomly,
and that you phrase the questions in a way that allows them to answer
without being influenced by the question, but considering that it would
be in eBay and Yahoo!'s best interest to sell more expensive moving ads,
I assume it's not in their best interest to skew the results in favour
of static ads.

For 600 surveyed people, this formula gives a margin of error of 4%.
This would be bad if you came up with something like "51% love moving
ads, 49% hate them", because then it might just as well have gone the
other way. If you have numbers like "70% hate moving ads, 30% don't",
then it's more than good enough.

lucas

- --
"The reward of suffering is experience."
  -- Aeschylus, 525-456 BC, Greek Dramatist

-----BEGIN PGP SIGNATURE-----
Version: PGP SDK 3.2.2

iQA/AwUBQyiG2s0zN2kKTjB0EQL/5gCg8lvCbXXjVub3MLVVb2g1T5XdvaQAn0ZH
NJkCbigwfyLuBhF7jX9xk9VM
=57gQ
-----END PGP SIGNATURE-----



  OutlineAll MessagesOlder MessagesOldest MessagesNewest MessagesNewer Messages


 [F] TidBITS  / TidBITS  / TidBITS Talk  / Annoyance level of different ad types




Add a message

To add a message to this discussion, you must be a registered user. Enter your email address below. If you have an account associated with the email address you enter, you will be prompted for your password. If not, you'll be able to create a new account with no fuss.

Enter your email address:

Submit