Museums and the Web 1999

Best of the Web

Archives & Museum Informatics
2008 Murray Ave.,
Suite D
Pittsburgh, PA
15217 USA

Join our Mailing List.

Published: March 1999.


Promoting a Museum Website on the Net

Giuliano Gaia, National Museum of Science and Technology, Italy

The Museum and Its Website

The Museum

With its twenty-eight sections, from information technology to engines to astronomy, some 40 000 square meters of displays, and a massive 15 000 pieces in its collection, the "Leonardo da Vinci" National Museum of Science and Technology in Milan is one of the most important technical and scientific museums in the world.

The Museum was inaugurated in 1953. It is made up of three separate buildings: the Monumental Building, which is a former Olivetan monastery whose construction dates back to the early sixteenth century; the Rail Transport Building, which reconstructs the environment of an art nouveau Railway Station, and the Air&Sea Transport Building, which contains two of the Museum's most sensational pieces: the sailing ship Ebe and the bridge of the transatlantic liner Conte Biancamano.

One of the main attractions of the Museum is its display of over one hundred models illustrating Leonardo's machines. The relation with Leonardo is particularly significant because the man lived and worked in Milan during the central part of his life, and the great majority of its drawings are still in the city.

Beside its exhibition activities, the Museum is an important centre of educational initiatives, with a varied program of activities, meetings and conferences. Particularly important are the school activities, with a rich program of interactive laboratories, designed for pupils aged 3-19, and staffed by a team of museum guides. The laboratories include a number of hands-on exhibits whereby young visitors learn science through play.


The Website

Like every Museum deciding to open a website, we found ourselves facing the dilemma "virtual visit or not?". This problem is quite well pointed out in a recent Italian publication (Forte et al. 1998), which can be summarized as follows:

There are three ways of making a virtual museum:

  1. "simulated museum", trying to re-create the experience of visiting the actual museum (virtual tours, and so on..) without adding any information; according to this approach, to visit the museum or to visit the website should be the same.

  2. "information": the website is an instrument to use before or after the visit; it offers a lot of information not available at the actual museum. According to the authors, this one is often the European way of thinking a web museum.

  3. The real "virtual museum", a website in many parts independent from the actual museum, with many sections and exhibitions residing only on the net. This kind of virtual museum is not narrowly focused on the actual museum. This seems to be the American way of thinking.
At the Science Museum we decided to put ourselves between the second and the third approach - we were not trying to simulate any real visit, but we wanted the website as an instrument to prepare and deepen the actual visit - while we had also some sections independent from the actual museum.

We decided not to simulate the real visit because we felt that we would have inevitably lost one of the key factors of an actual visit to the Museum, its unique atmosphere, deriving from the fact that most of the Museum's collections are housed in the 16th-century "Edificio Monumentale", once a monastery complete with cloisters and halls of great beauty. This is a rather uncommon setting for a museum of science and technology which, by its nature, zeroes in on both topical and future issues. Such a blend of old and new gives an aura of great fascination and magic to the whole Museum.

We felt that if we had tried to build some sort of "virtual visit" this peculiar atmosphere would have been difficult to maintain. So we decided to highlight only few objects in the Museum, suggesting that many other could be discovered during a real visit. We selected seven subjects (ships, aircraft, cars, motorbikes, trains, radio and energy) and highlighted two objects for every subject, describing them very deeply. We deliberately chose also objects which were not very spectacular but of valuable historical content, to suggest to the visitor that even less well known objects are worth a deep study. Moreover we decided to avoid any image of the object in the Museum, preferring drawings and photos of their real life and in their real context, trying to re-establish the relationship between the object and the real world so often broken by the freezing in a Museum.

Detail of the Leonardo section

Fig. 1 - Detail of the Leonardo section (

The Leonardo section is quite different. Leonardo's machines are the main attraction of our Museum, and certainly the most international one; so we decided to make a reference section, reporting ample descriptions and original drawings of 40 of his machines - we intend to describe 100 machines by spring 1999. This huge section was also important to create a "core" of the website in order to make the promotion easier. We also decided to use Leonardo's machines as key graphic elements of the Home Page.

Apart from the actual Museum-related sections, we opened some virtual-only sections, trying to transform the website in a real "web center" of various activities. The first virtual section to be opened was the one about how the Internet works, with an interactive navigation histories "traceroute" and some spectacular examples of Internet maps.

In autumn 1998 we made an important decision: to become "permeable" to the Net, making our website borders much less defined than they were before. On one hand, we started to host mirror sites of exhibitions we found valuable. The first one was a 38-page website made by an Italian school about steam engines, with nice animations - hosting their website was also a way to say that on the Net even unusual authors like schools can create valuable scientific content.

On the other hand we started to open pages on different websites, in which the Museum offered the editorial content, while the IT staff of the host actually made the pages - the most remarkable "outside section" we made was the one about the sea legends opened on the website of Martin Mystère, a well-known Italian comic character.

Detail of the

Fig. 2 - Detail of the "Sea Legends" Website (

As for the interaction with the public, the interaction via e-mail with the Museum is strongly promoted all over the website; we have also a rather lively guestbook and a News section in which visitors cam subscribe to receive news directly by e-mail (350 subscribers so far). We also have started an experiment of e-commerce selling books in association with Amazon.

Our aim is to greatly increase the interaction level by opening discussion forums on our and on other websites; a mailing list devoted to discuss IT issues in Italian Museums has also been started.

The Promotional Campaign

Once the website was ready, on 26 January 1998, we started the promotion campaign. We decided to use as domain a composition of two Italian words, (equivalent to instead of only few letters (like, for example,, because we felt it would be easier to remember and recognize for Italian people, who were our main audience.

Direct Means:

The first audience to be targeted, and the most natural, were the Museum visitors. Large posters were placed at the entrance and at the bookshop, the front-desk staff was instructed to give the URL to anyone requesting it, and gradually all of the official mailing of the Museum reported the URL. This are quite natural things to do, but their importance is often under-valued.

On autumn 1998 thousands of schoolteachers were addressed with a letter signaling the existence of the website and the opportunity to receive its news directly by e-mail. This was an important move due to the strong ties existing between the Museum end the schools - some 120,000 of its 300,000+ visitors are schoolchildren. Although Internet in Italian schools is not very widespread yet, we had a good response especially in people signing our News mailing list.


Creating a news item from our website was not an easy task because in 1998 the opening of a Museum website was no longer interesting in itself, even for a well-known Museum like ours. It was certainly easier to get attention by the technical magazines, always searching for quality websites. The first thing done was of course targeting a number of magazines with short press releases. Then we offered articles about the site - some sort of "inside views" of the project - and we succeeded in having a paper on Internet News, the largest-circulation Internet magazine in Italy, only a month after the opening of the website, on February 1998.

As for the major newspapers, we joined our efforts with the Museum Press Office, and officially presented the website during a national event, the Scientific Week, in which all Science Museums in Italy proceed to various initiatives to spread the scientific culture. By doing this, the news of the opening was reported by many national newspapers. Also important was the News mailing list, subscribed by many journalists, especially of local newspapers, as a good mean of news retrieving.

Another important way of diffusion of the website was its printing on CDs attached to Computer-related magazines - something we did four times during 1998, having our website printed for free in over 200.000 copies. We also succeeded in having back several thousand unsold copies - We distributed them to the classes visiting the Museum. The magazine people were very satisfied with the operation, since we provided them good free editorial content, while for us it was undoubtedly a very good promotion, and also a way of rewarding our visitors (by the distribution of the unsold copies).

On The Net

Submitting the website to the largest number of search_engines and directories was of course the first move. Being quite difficult to rate high in search_engines with all our subjects ("ships", for example), we decided to use as main keyword "Leonardo da Vinci", which appears also in the official name of the Museum. We used also massively a specific instruction of many search_engines, the "link" option, which reports all the pages linking a specific URL. We found this instruction useful to identify websites most likely to be interested into linking the Museum. We selected some websites quite similar to ours (i.e. other Italian science museums, sites about Leonardo, etc.), and contacted all the websites linking them, proposing our link. We also contacted all the Italian websites of certain categories, such as museums, tourism sites, sites about Milan, institutions, universities and sites about the various sections (Leonardo mainly, and ships, trains, etc.). Our task was eased by the huge "link" section we had prepared in our website, reporting the major websites on most subjects. This section not only was greatly appreciated by the visitors, but also allowed us to address the major websites with a message like this "We have already linked your website - could you please return us the favor?". Exchanging a link is always much easier than asking a single-sided one.

In general we had a very good response from our link campaign - we sent out hundreds of e-mails, received dozens of links and never got any complaint for spamming (sending unwanted messages). Where we failed almost completely was in getting attention from the websites of public institutions - with the exception of universities and museums, nearly no ministry or government website accepted to put the link to the Museum (which is a public institution itself). This can be explained with the high degree of bureaucracy existing in the Italian state, even on the Net - one of this public websites asked me to send the link-exchange request with an official letter by traditional mail!

Evaluating the Effect of Promotion

To evaluate the impact of the promotional campaign on the website audience it is necessary to carefully examine the server logfile, which reports every request received from external visitors.

Particularly important is to study the "referrer" - the referrer log is a part of the log file that reports the page from which the visitor clicked to get the current page; if I click the link to the Milan Science Museum from the Boston Museum of Science link page, the referrer reports the Boston URL. It's a precious tool to know where you get your visitors from - when no referrer is reported, in most cases it means that the visitor directly entered the URL into a browser, without clicking any link. So, he already knew the URL either because he already went to the site, or because he knew the URL from press, friends, etc.

Here is the referrer log for a 7-month activity of the Science Museum:

Where Do the Museum Visitors Come From

Fig. 3 - Where Do the Museum Visitors Come From?

We will now see in detail how, analyzing this three main groups of referrers, we tried to evaluate the effects of our promotional campaign.

No referrer

It's quite difficult to evaluate exactly the impact of all the direct means, such as press, posters in the Museum and so on, because they all end up in the "no referrer" group, and you cannot tell exactly which move took them in, nor you can distinguish the new visitors from the returning ones (although using cookies you can approximately determine this last figure).

One case where you can guess with a good degree of approximation the impact of an event is when you have a peak that you can put in relation with a single event, especially a short-timed one (e.g.. an article on a daily newspaper; usually its effect will not last more than 1-2 days.).

As for CDs reporting the site, they don't improve much the visits at the actual sites, at least not in the short term; to get people from the CD to the online site, we left some links opened to frequently-renewed page (such as the News page). On the last CD we also left a javascript on the Homepage, which connected itself to an online counter; in this way we registered some 300 visitors, notably, the ones who used the CD while having an Internet connection opened; of course they were only a small fraction of the people who actually ran the CD.


Analyzing the referrers from search_engines is very interesting because they tell you what your visitors were searching when they found your site: most search_engines report in the URL what query was made by the visitor. This is a valuable tool to know what your visitors are searching in your website, and is also a good indicator on how you're placed on the search_engines, so you can do something if you feel that you're receiving fewer queries than expected on a specific subject (for example you can change your META tags, the HTML tags that instruct the search_engine about how to index the page).

Search_engines and directories are very important for promotion also for another reason: they're more and more becoming "web portals" that can promote your site in a very effective way. A look at our first months' log will be convincing enough:

Visits 26 gen - 31 mar 1998

Fig. 4 - Visits 26 gen - 31 mar 1998

The spectacular peak of 18 February was entirely due to the fact that Virgilio, the most important Italian search_engine ( made us "Site of the Day" and put us on the Home Page with good evidence. We also had significant peaks when Yahoo! Italy put us on the "Best of the Week", and again when Virgilio made "specials" about the Italian Museums online. So, even if Altavista is the most used search_engine (at least by our visitors), national search_engines and directories proved very important in promoting our website, and were also much easier to contact than the big international directories.

Also interesting is the study of how many pages are viewed by the single visitors. We have decided to use this data as an indicator of the quality of the audience: the more pages the visitor looks at, the more interested he is. Of course this is not always true, especially for people with a very specific interest, who want to get as quickly as possible to one page and then read it carefully, but being quite impossible to use the time as an evaluating method, for the methos of use varies greatly from person to person (some get the page and then read it offline, others open more than one page at the same time, etc.), this one seems to be the only reliable method.

Here you can see the interest of the visitors, expressed in average number of page requests per visit, coming from different search_engines, compared with the total average and with the "No Referrer" visitors.

Average number of page requests by referrer

Fig. 5 - Average number of page requests by referrer.

As you can see, while the "No Referrer" visitors are quite of the same interest (6 pages per visitors), Yahoo! and Virgilio are significantly higher (7 and 8), while visitors coming from Altavista see only an average of 4 pages per visit.

This perhaps can be explained with the different nature of the three search_engines: Yahoo! and Virgilio are directories, while Altavista is a "pure" search_engine; this can have two effects: that Altavista is more used for specific searches, and brings people directly to the page they were searching, while directories tend to link only the Home Page; on another hand, directories usually describe better the website content, and by doing this they tend to select more interested people.

Link from websites

Working on the websites is a bit more difficult because the number of the referring websites is much greater than that of the search_engines. In the graphic below you will see the top 40 referring websites, divided into categories.

Top 40 Referrers From Websites

Fig.6 - Top 40 Referrers From Websites

The sites which proved themselves more useful in bringing visitors were the ones about Milan; in this category we include virtual guides of the cities aimed at tourists, as well as community networks visited mainly by Milanese citizens. In the "Special Interest Groups" category we include all the websites runned by individuals or associations about a specific subject related to the Museum, such as aircraft, trains, astronomy, etc. An interesting source of visitors were the small Internet providers (in Italy we still have a very fragmented Internet providing situation); we plan to target these small providers with a specific linking campaign during spring 1999.

The Logo-Diffusion Campaign

One important step of our promotional campaign on the Net was the logo-diffusion campaign. We prepared two little buttons and asked the webmasters to help the Museum by putting on their website the button; in exchange we reported on the "contact" page every site showing our button.

The Two Banners with the HTML Code to Download

Fig. 7 - The Two Banners with the HTML Code to Download. (

This was intended as a community-building tool: we gave webmasters the appearance of being "Science Museum supporters". This initiative worked particularly well with small websites and personal Home Pages, which are much more inclined to put buttons on their websites than big websites, which usually have more graphical and image constraints; we gave also webmasters the feeling that even little Home Pages with just tenths of visitors were important for us, and really they were, because summed up they made a good audience even from the quantitative point of view.

The advantage of being linked with a graphical icon instead that a text link are various: firstly, you're much more visible, especially if you're in a long directory; secondarily, you create a sort of brand image, which makes you recognizable on the Net, especially if the visitor sees your logo on two or more different websites.

But there's also another very important advantage. If you manage to convince webmasters not to download the buttons to their servers, but to use just the HTML code, leaving the buttons on your server, this can be very important for two reasons. The first is that doing so you have the possibility to change the buttons without having to re-contact all the webmasters, and the second is that with a good log-analyzing tool you can know not only how many people clicked the buttons, but also how many people actually saw them.

So you can monitor also the "external visibility" of our logo, as showed in the graphic below.

Visits and External Visibility

Fig. 8 - Visits and External Visibility 1 mar - 31 Oct 1998

There's another advantage you can get from monitoring the traffic on a group of various pages independent from yours: the possibility of feeling seasonal changes with much more accuracy. For example, the external visibility goes very low during summer, while the visits at the Museum site do not show the same decrease (this was due to the fact that the Museum site received an important promotion from a search_engine during August).

Additional Resources About Log Analysis

  • A good reference site for discussing standards and technical problems in an understandable way is the Internet Advertising Bureau (

  • Some interesting technical discussions can be found on

  • As for the books, I found very useful Rick Stout (1997), Web Site Stats, Berkeley, McGrawHill.

  • A low-cost but full-featured log analyzing software is WebTrends, (

  • There's also a very interesting free counter which reports referring websites and search_engine queries: it's the Extreme Tracking Counter (


Forte, M. & Franzoni, M.(1998). Il museo virtuale: comunicazione e metafore., Sistemi Intelligenti, Il Mulino, Bologna, anno X, n. 2, agosto 1998