Honoured Guests - Towards a visitor centred web experience
A website is a work in progress. Unlike an interactive, an exhibit or an exhibition, it offers an experience which can be changed once completed, in fact an experience which the user expects to be changed. It is dynamic and it is organic. This offers managers and evaluators of websites an opportunity not often afforded to the developers of physical exhibitions ñ to respond dynamically to the needs and expectations expressed in visitor feedback and visitor studies.
In addition to being organic, a website holds the potential for interactivity. Interactivity is the foundation on which the web has been built, springing, as it did, godlike out of the thigh of the collaborative environment at CERN. Therefore, even when interactivity extends only to a visitor feedback form, Internet culture is such that museum website visitors are not shy in expressing their opinions and offering suggestions where development or usage is concerned.
Finally, websites generate raw usage data. Web servers can record not only the time of requests and the exact files requested, but the countries, domains, in some cases even cities from which these requests come, along with a variety of extra information.
It is evident therefore that the web is a medium which offers a unique opportunity to involve users in its development. It provides the means of collecting both subjective and objective data, solicited or unsolicited and the possibility of responding to it both summatively and formatively.
The flourishing of museum visitor studies over the last decades of the twentieth century has reflected an ongoing commitment to evolving a more visitor-centric experience of the museum environment. Ames (1993) states the purpose of museum visitor studies clearly, "Visitor studiesÖ help attract larger, more diverse audiences by exploring the positive and negative perceptions of museums among visitorsÖthereby enabling museum staff to work on strengths and to correct weaknesses". My contention in this paper is that museum website development teams must take a more professional attitude towards website visitor evaluation in order to ensure we make use of the resources available to us to receive and assimilate the user experience to the benefit of our sites and thereby of our visitors. They are honoured guests whose input should be highly valued.
Brochures ñ a dirty word
The current stage in the development of museum websites still owes much to the early nature of the web. Early web development can be likened to that of early television. In the early years of television popular programming included among other things plays. Plays are a visual medium, they tell a story and engage the viewer, broadcasting a play was a good idea. However, these plays were actually broadcast as theatrical plays using still camera shots and stage sets. The camera angles, the visual effects, the music, the lighting and most importantly the removal of the stage itself, those things we now understand to be the vocabulary of television plays, had not been explored. What the viewer saw was a stage broadcast with all camera shots taken from the front as if watching the play at the theatre.
It is hardly surprising therefore that the earliest websites, including museum websites, were based on a pre-existing form of media, that being the brochure. The web originated as a pure information publishing medium with its roots firmly in document sharing. Visually-enhanced, physical, information documents are brochures. This makes the brochure form a natural form to apply to the Web given the characteristics already outlined. Adding to this the highly personal nature of the web, the possibility it affords for document publishing with very little knowledge or experience, the slightly gnostic atmosphere which surrounded its most enthusiastic adherents and it becomes obvious that the first mass approach to the Web would be to see it as a ëpersonal propoganda machineí. In using the word ëpersonalí I mean to refer both to the individual character and the communal personality of institutions and companies. In short early websites were online brochures and including websites of museums.
It may seem that the nature of the information on museum websites would preclude them from being classed in the same way as other, information-light websites, after all the majority of museum websites have always included information about our collections and many used the web to tell stories about those collections from the earliest stages. However, it is in the structuring of information that the brochure formula was most evident. The Science Museumís own pages show this phenomenon and its development.
This image of the original home page of the Science Museumís website from 1997 demonstrates the point clearly. We had thought internally about our audiences and arranged our page to our liking. Therefore, Exhibitions On-line is at the top because we were, justifiably, proud of the work we had done there, exploring the possibilities of new technologies in setting context and interpreting collections.
Exhibitions and collections are next. Despite the fact that many visitors think of the two and, as we have discovered through subsequent visitor studies, have different needs and interests in approaching both, in the Museum they are developed by one department, therefore we grouped them together. Further, we see the influence of an enthusiast on the front page in the presence of the Science & Society Picture Library which, although it is an important arm of our commercial division, is no more important than any of our other arms such as exhibition development or press and therefore does not warrant a presence on the front page of our site. Finally Visitor Information is the fourth item down. From research at the Museum and from common sense we know that many visitors to our site have come specifically to find visiting information, however, we have not highlighted it at all. The final point to note is that there is no way into the site other than by using the Museumís personal divisional navigational preference. It is clear that the structure and navigational options mark this site out clearly as a ëbrochureí site, one developed with reference to the Museumís needs and not with reference to the visitorís expectations.
Later developments moved Visitor information up to the second spot on the navigation bar grouped the Science and Society Picture Library with other commercial operations and split Exhibitions and Collections. Further highlights were offered which enabled visitors to find content not only through the divisions on the navigation bar but also through exhibition highlights on the front page.
Yet we are still in a position of offering content based on our own activities, until recently the Research pages of the site were more focused on the research we do here than on the resources we have for researchers. The education, entertainment and information remit of museums remains and visitors expect and authoritative voice as much in our virtual as our physical incarnations, but they equally expect to find that information easily and clearly.
What a difference three days made
ëMuseum visitor studies has always been a subject with an agenda ñ to improve the lot of visitors.í (Bicknell and Farmelo, 1993). It is important to remember that the end aim of all museum visitor studies is to ensure site development follows to improve the lot of the visitor. This will obviously benefit the visitor ñ making content more relevant, easier to find, more attractive ñ but in doing so it will also benefit the museum.
The latest estimate of the number of domains on the Internet (July 1999) is 56,218,000 (Internet Software Consortium http://www.isc.org/). This is up from 43,230,000 in January of the same year, an increase of c.26%. As the number of domains continues to grow museumsí sites will increasingly be judged against standards set by the commercial sector in entertainment and education websites. The quality of information we provide, the extent to which our web presence contributes to their remit for education and access, ease of navigation, visual attractiveness and so on will all become as essential virtually as they are physically. Indeed governments have already introduced formalised performance indicators relating distinctly to a museumís websites they are therefore already having a direct impact on many museumsí funding.
Museums were amongst the first to recognise the enormous potential of the information and educational opportunities which the web can provide. In Britain the governmentís drive towards access can be directly fulfilled through the immense reach of the web. Therefore it is in our interests to provide a web service. However, it is important to remember that the web is a self-selecting and fast moving medium. It is not a ëpushí medium, it is a pull medium and our aim is not to get visitors to visit the site and then leave, but for them to experience the content we have prepared and to build community and bring their own experiences to it. By listening to our visitors we strike more of balance between serving our needs and serving the needs and of our visitors.
There are three main forms of data which have been used to date and can be used simply by researchers to conduct visitor studies: server statistics, qualitative studies and unsolicited comment and existing studies of museum visitors such as MORI reports and internal knowledge of audiences and needs.
One of the great benefits of the web is that content providers can determine in a more accurate way than any other visual media, how many people have seen their information ñ whether that be an advertisement, a resource or a whole website. Statistical file analysis is used by almost all websites and this provides a bedrock, an expectation, that the uses of museums sites should be subject to scrutiny a priori. A good thing in itself. Statistics packages can analyse the hits to our servers and offer general indications on the popularity and use of our sites. This is particularly true in analysing such areas as the most popular entry and exit pages and paths through the site. By analysing which sections of the sites are most used or most linked to museum web teams gain indications as to the type of content most valued by the users.
Good statistical packages are easy to come by and need not be expensive, or even cost anything. Reliable statistics can be obtained by packages such as Nedstat (http://uk.nedstat.net/) which compiles data for sites with up to 10,000 visits per day or an eXTReMe Tracking account (http://www.extreme-dm.com/tracking/?reg) which does much the same thing. There are also packages which can be downloaded and installed on your server or a remote machine (if you do not have access to the server) such as Analog which is free for both Mac and PC (http://www.statslab.cam.ac.uk/~sret1/analog/). For more graphic representations of data there are several packages for which you can pay such as HitList (http://www.pwi.net/hitlist.html) (which creates large database files and so should only be used on machines with large memory and disk space) or Webtrends (http://www.webtrends.com).
An example of useful visitor information which can be gathered from server files is the types of browsers being used by your visitors.
Lies, damned liesÖ
There are however pitfalls. As Disraeli so wisely said, ëThere are three kinds of lies: lies, damned lies and statisticsí. Whilst it tells us about visitor habits and gives us the raw data of their location, time spent onsite and so on, analysing quantity of hits without analysing the quality of user experiences is a mistake. There is no way to know whether a site is confusing as visitors access the home page; can visitors get straight to the information they need; is the reason for the apparent popularity of one section of the site directly linked to its place on the navigation bar rather than its content?
An example of the need for caution can be seen in comparing the quantitative data for accesses to the /on-line directory of the Science Museum site with qualitative data which resulted from a museum visitor study conducted during June 1999. This directory houses the majority of the Museumís interactive content on the website including exhibitions, games and special content developed for children. The statistics from the server files demonstrate the a large number of hits to this section of the site over a period of 5 months.
From the graph above a number of assumptions can be made:-
These are the assumptions which can be made from the statistics taken in isolation. However, if we look at these statistics in conjunction with a more qualitative analysis of the site we find out far more about our visitorsí actual experience of the site.
Getting our hands dirty
In the summer of 1999 a Science Museum team conducted an on gallery evaluation of the home page of the Science Museum website. The study was carried out to evaluate the design and navigability of the Museumís homepage. 112 visitors, selected at random were questioned over 3 days. 60% of respondents were male, 40% were female this split is representative of the general UK internet audience of which 38% are female (Cyberatlas, http://cyberatlas.internet.com/big_picture/geographics/article/032359112684100.html). The UK audience is the largest for the Science Museum website.
Respondents were shown a print out of the Science Museum home page and asked questions about it. As the aim was to test the design and navigational text on the page a print out of the page was used (see figure 2) in order to avoid creating an artefact of method. Users were then asked a series of questions about the page, how they would navigate around it, what they might expect to find in certain sections.
The popular /on-line directory is labelled ëOnline Featuresí. This title had been chosen because we felt that it best described the kind of content was housed within that directory and because we felt that it was a more general title than ëOnline Exhibitionsí offering therefore a better indication of the information presented within it. Note that these conclusions were drawn without any form of visitor study.
Visitors were asked what they would expect to find if they had clicked on the button ëOnline Featuresí, they werenít prompted in any way. Their responses shed new light on the efficacy of our signposting as can be seen from the table below:-
It is evident from these responses that the title which we had deduced to be so clear from the statistical evidence is actually confusing. A better option might be to use the title ëVirtual Exhibitioní as although it suggests virtual reality to 18% of respondents there is far less confusion about the kind of content and the expectation of interactives and games is fulfilled.
Using the same survey looking at the ëunpopularityí of education pages it is apparent that in this case the navigational text used is correct for the purpose. 55% of visitors expected to find information for the education community (i.e. teachers and students), 10% felt that there would be information for children on education, 4% felt that there would be information for the further education community.
These results can be further combined with those from a survey conducted in a slightly different way, via the Internet, in 1998 which focused on use of the education pages. The full report of findings can be found at (http://www.sciencemuseum.org.uk/eval/index.htm).
Science Museum website - education section survey
This survey was conducted as follows. An invitation was placed on the Science Museum web pages and volunteers emailed their details to the Museum. These were responded to by Solomon Research with a message offering a choice of completing a text questionnaire with the e-mail or an on-line form located on the Solomon site.
Of 216 responses 49% came from those in education and of those 46% had looked at the education section. A healthy proportion, indicating that the section is speaking to the right audience. The study also concluded that a fair proportion of the visitors in the education section were looking for educational resources. The specificity of information in the /education directory is the main reason why its hits are so much lower than the /on-line directory. The information it provides is right, the target audience is smaller and has a more specific agenda.
Further benefits of getting your hands dirty
Thus far I have demonstrated the importance and usefulness of qualitative data to enable summative adaption of web pages or web experiences in line with visitor needs. It is also useful for formative approaches to new website content. Often studies done in oneís own museum can be looked at in the context of other international studies.
Chadwick and Boverieís investigation into the characteristics and patterns of behaviour in visitors to a museum website (1999) revealed interesting data about the intentions of visitors to the New Mexico Museum of Natural History and Science museum website which can be extrapolated as being appropriate to all museum sites. The main purpose for visiting, whether in a group or alone, was found to be learning, with 66.20% of lone visitors and 91% of group visitors stating this as to learn. In contrast, 44% of lone visitors and 9% of visitors in a group stated that their main purpose for visiting was to simply browse the site.
In the Science Museum there is a high volume of specific enquiries to curators relating to the collections. This combined with the popularity of our search engine and with Chadwick and Boverieís results led us to conclude that we could do more to target visitor needs through providing a science - subject oriented navigation alongside our needs oriented navigation. Such a form of navigation would enable visitors to reach specific information more rapidly and provide a useful focus to the general educational interests demonstrated in the results of the Chadwick and Boverie study.
This too was subject to a visitor study. In order to evaluate a) whether the current navigation system was working or could be improved and b) whether a new navigation system would get visitors to the content they required more easily we undertook the aforementioned study of June 1999.
I have already referred to the print out of our home page which examined existing navigation and design. The second half of the evaluation concentrated on two topic navigation systems, one focusing on three broad topic headings and the other on ten more specific topic headings. These were developed in consultation with staff on the basis of our collections and of other commonly used subject systems. The first menu comprised three subjects: Science, Technology and Industry. The second menu comprised nine subjects: Science, Health and Medicine, Power, Chemistry, Communications, Engineering, Transport, Physics and Computing. These menus were put on A5 sized cards.
One selection of collections headings was chosen as follows: Astronomy and Maths, Experimental Chemistry. Air Transport, Civil Engineering, Space Technology, Bioscience, Agricultural Engineering, Materials Science and Communications. An optimum division of collections headings into subject areas was drawn up. The collections headings were then placed on small cards such that they could be placed on top of the subject menu headings.
Visitors were selected at random and shown one of the two subject menus, they were then asked to place the collection headings onto whichever of the subjects menus they felt they fitted best into. They were informed that they could place more than one collections heading on each subject menu, they were also informed that they could place a collection heading on more than one subject menu.
It was found that whilst the three broad subject headings led people to place their collection heading cards in the right general area, a substantial number of visitors placed headings in totally irrelevant categories. For instance, 30% placed Experimental Chemistry in Medicine, and 32% placed Materials Science in Technology.
Tests on the nine more specific subject headings showed very high recognition for a greater number of collections areas such as 91% of visitors placing Air Transport in Transport as opposed to 83% placing it in Technology under the three broad subject headings. This led us to conclude that, with some tweaking of collection heading names to ensure that all were clear, the option of using around nine more specific subjects was a more useful option for the visitor coming both to learn and browse.
Studies like this are simple to arrange, do not require technology or portable systems yet offer invaluable information to museum website teams about their visitorsí needs. When taken in context with pre-existing visitor studies and quantative analysis they enable us to form a clearer picture of our visitors.
We are still part of the museum
Finally, it is important to set the site in the context of our museum visitors, our museum audience. Marketing departments and personnel employ tools which can be used to offer useful information about Museum visitors and website visitors.
The biannual quantative MORI survey at the Science Museum reveals interesting data on internet access and the Science Museum website. Six out of ten visitors have internet access; the overall figure in the general public is 21%. The table below shows the percentage of visitors who have visited the website and compares it will the percentages who have visited other websites.
The sites listed are very popular and well known sites. While these figures do not suggest that we compete with them on a national scale, amongst our visitors we are evidently providing a useful service.
Data can also be gathered from simple self-selecting surveys on information leaflets. For instance, the Science Museum currently runs a questionnaire on visitor information leaflets which includes amongst other questions two about the Internet. Of 170 people responding to the questionnaire 62% have internet access and of those 37% have visited the site. Whilst this tells us nothing about the user experience it does place our efforts to provide relevant information in context and will enable the monitoring of growth and decrease in interest in the site. The important things to note is that the opportunity to ask more pertinent questions about the site is obviously open.
Marketing expertise offers many more resources than simply existing or forthcoming survey opportunities. The experience of marketing and indeed other internal museum professionals can contribute towards the development of a more visitor-centric site in other ways. Visitor evaluations start from premises built on experience and knowledge, from valid theories which we are not arrogant enough to leave untested. Approaching content from a visitor focused point of view it is well to gain inspiration from those who deal with visitors on a day to day basis such as the marketing, education and research departments. In fact it is essential to consult with as many of the forward facing museum professionals as possible to ensure that as many groups as possible are considered and assessed. It would be as much a mistake to present only detailed research information on a museum website as it would be to present only marketing information.
Simple internal exercises can easily provide new ideas for user profiling and suggest visitor groups. Exercises of this kind have been run at the Science Museum National Railway Museum and National Museum of Photography, Film & Television as part of the redesign process of these sites. In the case of the Science Museum it has resulted in plans for a radical redesign and restructuring of the site in the case of the other two museums it resulted in a re-direction of the existing site and detailed content development.
Firstly it is important to assess who are the intended audiences. Very often these will overlap with existing museum audiences. For instance in the case of the Science Museum our list of audiences was as follows:-
Web Browsers, People wanting visitor information, those in formal education, museum professionals, opinion formers, self-directed researchers.
Further to this the group were asked to offer suggestions as to the needs of these various groups, specifics which could then be broken down into general categories of needs as follows:- Authoritative voice, Visit planning, General museum information.
Bearing in mind audiences and needs the group were then asked to create scenarios for visitors to the site which helped to demonstrate content and navigation paths. For instance, an opinion former such as a journalist visits the site, finds press releases easily and also notes forthcoming exiting exhibitions which will provide a good story in two months time; a researcher comes to the site enters the research section searches for all resources on a given topic, finds two exhiblets, one online exhibition, various papers and a curatorial contact. They then contact the library, link into the library contact and make an appointment to visit.
Often such exercises throw up new and challenging ways of structuring and presenting information. They must then be looked at in the context of current statistical and qualitative data and of formative evaluation.
For museums the imperative to be accessible and understandable is stronger now than ever before. In our eagerness to benefit from the communication and narrative possibilities the Internet has to offer to us we must be careful to take into account the intended audience. We must be aware of the fact that we are still in the early stages of a revolutionary medium and not berate ourselves where we fail. However, we must also avoid the complacency that comes with our undoubted contribution to the Webís development over the last six years. It is in asking questions Do we actually provide visitors with clear information as well as interesting information? Are we enabling our visitors to reach it as easily as we can or are we setting out our own agenda? That we will find new ways to engage with those visitors which will take us beyond the museum-centred websites of the past into the more balanced and rich sites of the future.
It may seem a daunting prospect - self-examination should result in change - yet we are amply provided with information and opportunities to gather information in the new statistical evidence that web sites produce by their very nature, in the legacy of visitor evaluation studies done physically and the pools of knowledge which reside in the different departments of the museum already. It is in combining all the resources available to us, however small or large that we will be able to move towards more visitor-centric museum websites.
Ames, P (1993) Views on the value of various evaluations. In Museum Visitor Studies in the 90s S. Bicknell and G. Farmelo (ed) pp.47-49 London: Science Museum
Bicknell, S and Farmelo, G (1993) Introduction. In Museum Visitor Studies in the 90s S. Bicknell and G. Farmelo (ed) p.7 London: Science Museum
Chadwick, J and Boverie, P (1999) A Survey of Characteristics and Patterns of Behavior in Visitors to a Museum Web Site. Museums and the Web 99: Selected Papers, D. Bearman and J. Trant (ed) pp. 154-162 USA: Archives and Museum Informatics, 1999