museumsandtheweb.com - Comments for "Impact of Museum Websites - a comparison using Alexa"

Thanks

rmiller@alphabetadesign.com — Thu, 11 Jun 2009 19:25:08 +0000

I did want to add my thanks to dbear for the thought and work on this exercise, one where more understanding is needed--the relationship between real and virtual museum experiences.

give me a break

Shelley Bernstein — Thu, 11 Jun 2009 19:17:25 +0000

David,

Sorry, but I don't buy it. I think you could have a done a lot more to educate your audience here and instead took a shortcut. You could have easily put in some gotchas/pros/cons in the body of the text and still generated plenty of discussion. The fact that you didn't is disappointing.

The discussion is the value

dbear — Thu, 11 Jun 2009 18:51:01 +0000

Dear Shelley,

My point in posting is to generate discussion, not to promote Alexa or any other source of data. From my perspective, what is interesting is not to compare data sources - none of which give a 'true' picture - but to look at what any source can tell us with intra-source comparisons. Those of you in a specific museum will, of course, want to look more in depth at all the sources that together can tell you something about your own institution, including in house logs and door counts and post-visit surveys. For both these purposes, I'm quite happy that the posting is bringing out issues with data and what can best be done with different sources.

David Bearman

thanks for the correction!

jtrant — Thu, 11 Jun 2009 17:13:17 +0000

thanks Perian for spotting the de Young / Fine Arts Museums of San Francisco correspondence that i missed. [it bothered me a bit that 'the thinker' wasn't there. but i'd forgotten the other name!]

correction / update coming shortly.

/jt

j. trant archives & museum informatics www.archimuse.com

Scoob, I expect more from you...

Shelley Bernstein — Thu, 11 Jun 2009 14:46:00 +0000

Scrappy Doo, here...
I have to say I think it's pretty irresponsible to publish a list like this without looking at more data and making sure that questions about the validity are in the post itself, not the comments. You should be listening more carefully to your own presenters - people like Seb Chan - who will always caution you to look at all sides of an issue before throwing data out there.
First, using Alexa for traffic ratings and not cross-comparing with Google as a second source is problematic (or any second source, for that matter). You also need to note (or at least link to) the pros/cons of grabbing stats this way. Alexa is notorious for being off the mark, Google too - that's why you should at least look at more sources or at the very least explain this. Frankly, Frankie Roberto probably has better data on the sites he's been working with because he's going to source and perhaps you could have used his data to at least provide a baseline as to how accurate these figures are, but I digress. Let's just say there's a lot to be desired here.
Second, yes, you do need to look at physical visitor stats to cross compare. It's another metric that needs to be factored in and there will often be a direct correlation. A institution that has a popular exhibition schedule one year will dramatically change their site traffic.
Third and most importantly, you need to look at how many pages a site is offering. A site with fewer pages is going to wildly impact these ratings just because there less out there to find. Take the Powerhouse, for example, which has over 72,000 object records online. I love the fact that there is no explanation that si.edu consists of something like 14 museums under one domain. On the other hand a site with far less (New Museum) content can get you thinking even more. Had you done some type of analysis about what sites are offering, it could give you all kinds of interesting information to better explain the metrics.
All that said, I'm a proud mama to be at #46, but I expect more thorough consideration out you guys, darnit!

all traffic sources are fraught with bias

jtrant — Thu, 11 Jun 2009 14:00:00 +0000

Hi Rohn,

we realise that all sources of Web traffic data are prone to biases -- and that traffic itself isn't a very good measure of success for museum Web sites. [there have been many papers at MW about this subject; try a site search on analytics for example.]

this was quick probe using a publicly accessible source that we felt useful enough to share. it probably should be repeated using other sources and the results compared to see how different they are.

the 'list' is interesting to me, not for its actual numbers, but for the relative positioning of institutions on it, and for the comparisons it makes possible with other rankings.

AND for the comments it elicits about methods and sources.

thanks for your input.

jennifer

j. trant archives & museum informatics www.archimuse.com

Top 50 Museums -- Questions about Alexa data

rmiller@alphabetadesign.com — Thu, 11 Jun 2009 13:33:00 +0000

I'm Strategic Director for IconNicholson, an interactive design firm in NYC that has worked with the National Gallery of Art, the Metropolitan Museum of Art, the Chicago Humanities Festival and other cultural non-profits. Our work is grounded in research, including Website metrics. I appreciate the effort you've gone to here---and the point you make about virtual museums standing on relatively equal footing with real-world institutions is an interesting one.
I do want to point out that Alexa, a free source of Website metrics, is not considered to be a particularly reliable source. They collect information from a group of users who have agreed to install the Alexa tool bar on their browser (or have had it installed with other software) There can be a wide disparity between the behavior of this small sample group and the actual Internet population, especially given the large amount of traffic generated outside of the United States for major cultural institutions such as the Louvre or the National Gallery of Art.
For example, I work with a large corporate client here in the United States. Alexa says they receive around 100,000 unique visitors a month. I know from viewing the actual server logs that number understates the actually traffic by around 250,000 unique visitors--a factor of 3.5 times.
Compete.com is a similar free resource of Internet metrics, but is in my opinion more reliable because they also receive anonymous information from Internet service providers (ISPs) from their servers. Compete.com lists the National Gallery of Art (www.nga.gov) as the 14,690th most popular Website on the Internet, vs. Alexa.com estimate that it is the 50,030th.
These free metrics sites drastically understate the amount of traffic coming from schools, universities, and from outside the United States which are all large sources of traffic for cultural institutions.
Please excuse my going on about this point, which stands beside your point about virtual vs. real-world institutions. But I have more than 20 years of leading strategic planning for Websites and Internet strategy, and I know from working with my clients that this list is going to get forwarded on to colleagues, board members and management. I would not represent this information as accurate even as relative estimates of traffic. The methodology behind these metrics makes it likely that errors vs. actual traffic are so significant as to make the traffic estimates meaningless.
Other than saying there is a general likelihood this is a list of the top 50 museums online (give or take 10 or so) you can't say much about its accuracy.
Comscore and Neilsen are the only two companies that gather balanced and statistically significant samples of Internet traffic, and unfortunately subscriptions to their data cost more than $25,000 per year.
If anyone wants to discuss the research methods of any of these firms, or about Website metrics in general, please feel free to contact me at rohn@millerusa.com and I'd be glad to engage further.
Thanks for an otherwise thoughtful exercise and post. I'll say that in our work we regularly survey the landscape of cultural non-profits on the Internet and the Rijksmuseum in Amsterdam has always stood out as one art museum that engages online visitors in a wide range of new ideas on their Website. As many of you know the museum itself is undergoing extensive renovation until 2013, which makes the Website at times the only "place" where the museum exists.
Cheers-RJ

Comparisons that might be valid

dbear — Wed, 10 Jun 2009 20:51:30 +0000

The 'rank' comparison could be useful even if the experience of visiting in person and online are very different. After all, rank is about degree of recognition. Nevertheless, I'd agree that we ought to look more to comparisons within the online experience or within the in-person experience.

What does it mean, for instance, that the Newseum online visits average 3.2 minutes while the Smithsonian and the Metropolitan Museum of Art average 2.6 minutes? Those are statistically highly meaningful differences. Are the three institutions attempting to create different kinds of user experiences?

Similarly, it is interesting (and surprising) that the Newseum gets 46% of its visitors from the USA while the Smithsonian gets 63% and the Met draws 61% Americans. Why is the Newseum drawing so much of an international audience when the two premier institutions are seemingly parochial? What does it mean that MOMA has 52% American visitors (or, I ask, that archimuse.com has less than 37% US based visitors)?

Exploring comparisons within the online visit will probably prove fruitful.

David Bearman

Internet usage is English language skewed

dbear — Wed, 10 Jun 2009 20:25:00 +0000

So this list also reveals something we know - that the Internet is language skewed. It isn't an absolute of course - the Russian museums site came up #4 and some French language museums do appear in my list, but clearly the English language institutions do better in Web data.

David Bearman

list construction and related validity issues

dbear — Wed, 10 Jun 2009 20:18:39 +0000

Thankfully I didn't have to look through 500,000 sites - only about 1000. I did a search for 'museums' in Alexa setting the limit at those ranked 1 to 1M. Then I eliminated from that list those that are returned by the term museum but aren't museums (search engines, wikipedia, commercial sites, etc.) and I constructed this list with those ranked up to 500,000.

The method has quite a few problems that I know of.

First, as you will find if you replicate it or search any one of these museums, the rank is a rolling three month average so it changes daily - not much but some, and there are definitely those going up and down in the ranks. More importantly, Alexa data are based on what Alexa can track, and thus are different from other measures such as Google search rank etc.

Second, the term 'museum' searched in Alexa excludes some organizatiuons that ICOM would consider museums, such as some 'sports halls of fame 'and 'zoos' that would have been in the top 500,000. For example, I found (on 09/06/09) using 'hall of fame':
National Baseball Hall of Fame    http://www.baseballhalloffame.org    156,844
Pro Football Hall of Fame    http://www.profootballhof.com     181,150

and using 'zoo' :
San Diego Zoo    http://www.sandiegozoo.org    32,597
Zoologischer Garten Berlin    http://www.zoo-berlin.de    202,003
Woodland Park Zoo    http://www.zoo.org    215,784

The third issue is that there are lots of associated statistics in Alexa that are more meaningful - length of visit for example, that are ignored here.

But in a gross way the data are still intriguing, and I ithink museum web staff will look deeper into the data on their own. It is one way of discovering whether the changes we make to our sites are having an impact.

David

compare Alexa rank to attendance figures

jtrant — Wed, 10 Jun 2009 15:45:00 +0000

i thought i'd try and probe my own question about how web visitation relates to on-site visitation, one that was prompted by the Art Newspaper / Economist statistics for the most visited museums in 2008.

Here's how they compare.

Museum	Approximate Number of Visitors (estimate from bar graph)	Rank in Visitors	Rank in Alexa 'museums'
Louvre, Paris	~8.5 m	1	13
British Museum, London	~5.9 m	2	22
National Gallery of Art, Washington	~5.0 m	3	19
Tate Modern, London	~5.0 m	4	9 [all Tate]
Metropolitan Museum of Art, New York	~4.8 m	5	2
Vatican Museums, Rome	~4.3 m	6	[not present]*
National Gallery, London	~4.2 m	7	32
Musée d'Orsay, Paris	~3.0 m	8	52
Musée d'Art Moderne de la Ville, Paris	~2.9m	9	[not present]
Museum of Modern Art, New York	~2.8m	10	7
Prado Museum, Madrid	~2.6m	11	[not present]
Centre Pompidou, Paris	~2.6m	12	[not present]
National Art Centre, Tokyo	~2.4m	13	[not present]
State Hermitage Museum, St. Petersburg	~2.3m	14	38
National Palace Museum, Taipei	~2.2m	15	[not present]
Tokyo National Museum	~2.1m	16	[not present]
Victoria and Albert Museum	~2.1m	17	17
de Young Museum, San Francisco	~1.9m	18	~~[not present]~~ 43 thanks Perian
National Portrait Gallery, London	~1.8m	19	[not present]
Reina Sofia, Madrid	~1.8m	20	138

* not present = not in top 500,000 web sites worldwide listed above

nine of the top 20 most visited museums don't show up in David's list above. curiously, one – the V&A – has the same ranking in both lists.

there certainly do seem to be some institutions having a far greater impact online.

/jt

j. trant archives & museum informatics www.archimuse.com