MW-photo
April 9-12, 2008
Montréal, Québec, Canada

Beyond Launch: Museum Videos on YouTube

Chris Alexander, San Jose Museum of Art; Allegra Burnette, The Museum of Modern Art, New York; Daniel Dark, The Indianapolis Museum of Art; David Hart, The Museum of Modern Art, New York; Jennifer Rossi, Hirshhorn Museum and Sculpture Garden; and Nicole Minor, The Exploratorium, USA

Abstract

Professionals from a variety of museums that have started YouTube channels discuss how to build and maintain these Web video initiatives. After briefly reviewing key discoveries from the different institutions, the panel will be opened for discussion and questions. Topics covered include: using YouTube as a marketing tool (do YouTube visits translate into Web site or physical visits to your museum?); YouTube as an educational tool; reaching and expanding your audience and community on-line and offline; managing and maintaining the initiative after the excitement of the initial launch dies down; in-house versus outsourcing; working within constraints of time, money, and resources; future trends and current features; and how not to waste time chasing your own (long) tail.

Keywords: video, YouTube, communities, 2.0, education, marketing

Introduction

The rapid growth of Internet video, and YouTube in particular, has opened up the possibility for museums to take part in the larger on-line visual discussion through the contribution of museum content. Previous papers and presentations at Museums and the Web have addressed how to create video (Incandela, 2007) and how this resource can be part of a larger community-building effort on-line (Caruth and Bernstein, 2007). In this paper, we focus on the results of experiments with video, particularly as featured on YouTube, that have been undertaken by several museums, including how these initiatives have changed since the initial phases of development and how these institutions intend to move forward.

All statistics featured within this paper are as of mid-January 2008, and do not reflect changes in content, visits, or any other factors after that date.

Included Institutions

Our sample group consists of YouTube video channels that were created within the time frame of July 2006 through July 2007. The institutions included are the Exploratorium, the Smithsonian's Hirshhorn Museum and Sculpture Garden, the Indianapolis Museum of Art, The Museum of Modern Art, and the San Jose Museum of Art. This sample reflects diversity in institution size, focus, geography, and culture. All participating institutions completed a survey outlining background information on their projects as well as statistical data from YouTube about their videos and channel.  

It should be noted that many other institutions in the United States and internationally are creating outstanding content on YouTube and other video sharing sites, including the Brooklyn Museum and the Dallas Museum of Art. We encourage readers to explore these and other sites outside of the scope of this study.

Working Groups

Most of the museums originated their initiatives with their marketing or communications departments, usually in conjunction with one other department representing some segment of the museum Web team (i.e. Digital/New Media), or in the case of SJMA, the combined department of Museum Experience and Education. At the Exploratorium, the initiative began within the Moving Images department. These core working groups do not seem to have grown in size, though a greater range of departments seem to be featured or to have provided minor input. For example, at MoMA, the Education department has begun to help contribute guidance and input, and at the Exploratorium, there has been greater collaboration with the departments of the Exploratorium Teacher Institute, PIE (Playful Invention and Exploration), and Life Sciences.

The story often told about digital video is that it takes fewer people to produce content than previously with analog video, and the results from our group prove this. In one institution, the production from end to end was created by one intern with one Web staff member posting the materials on-line; in two other institutions the working groups consisted of two staff members; and in one institution the working group was listed at four people, though not all of these staff would be working on the projects at the same time, nor were they dedicated only to video projects. Since the launch of these projects, none of the institutions reported an increase in the number of staff assigned to these projects, with only minor additional input from internal groups.

Technology

It appears that most of the productions were initially done using off-the-shelf Apple iLife software and consumer-grade digital cameras (still and video), sometimes with borrowed or rented equipment. A few of the institutions have since upgraded to more advanced software, more refined microphones, and HD cameras.

Production

There seem to be three types of production, which we’ll refer to as agile, adaptive, and traditional. The agile type consists of video produced quickly using a small crew and with little pre-planning or post-production. Examples of this include the Exploratorium’s videos demonstrating many of the installations within the museum space. This type allows for quick turnaround with little overhead, and results in the type of short and to-the-point video content normally associated with YouTube. The adaptive type consists of some pre-planning and an intended audience or message, but often captures extra footage and then refines the video in post-production editing. Examples of this include many of the “artist in action” videos which merge footage of artists, in the studio or creating an installation, with audio from the artist, curator, or visitor. As Chris Alexander from the San Jose Museum of Art explains:

In regards to our technical methods, we try to interview artists and let them do all the talking.  We have learned to try and keep this short and to the point; otherwise we have a lot of material to present.  The editing process whittles the interview down into nuggets that become the basis for the video.  We try to avoid doing a lot of writing because it is so labor intensive.  Also, we have toyed around a lot with the length of the videos, but have not come up with the perfect length yet.

The great majority of the videos tended to fall into the adaptive category. This may be due to the small set of people working on the videos and the need for rapid deployment for videos to coincide with events, or it could be due to the impermanent nature of YouTube traffic and the technical challenges of compression for the site. In the same way that many audio tours start with interviews with artists or speakers and then are later refined into a concise set of statements, it appears that many of the videos follow this production cycle. The traditional type includes a scripted message, a high level of preplanning, and a sizable crew. Few institutions have taken this approach in regards to YouTube, but have used this type more widely for content hosted on a museum’s own site.  

Production time (from shooting to editing to posting) was listed as between two days and a few weeks to up to a month or longer, depending on what the project is (size and length) as well as cooperation from within the museum.

Video Types

Videos can be created in multiple ways, and also for varying purposes. The institutions in this group were asked to categorize their videos within the following types, and could mark each video as belonging to as many as applicable.

  Exploratorium Hirshhorn IMA SJMA MoMA TOTAL
Total  42 7 57 37 31 174
Promotional (for program, exhibition, service, etc.)  30 3 38 2 21 94
Archival  38 1 22 10 22 93
Artist Interview  0 4 17 28 4 53
Employee Interview  2 0 26 0 2 30
Other Interview  1 2 15 10 1 29
Demon-stration  25 0 5 0 0 30
Audience Submission  0 0 0 0 0 0
Artist submission  4 0 0 0 3 7
Presentation/Public Program  16 2 6 0 10 34
Other 0 0 0 27 0 27

Table 1: Video Types

As might be expected, most reported that a good number of videos were of a promotional nature or for archival purposes. In other areas, the focus and mission of the institution are evident; for example, the Exploratorium features a number of video demonstrations, as might be expected with the overall focus of the museum. At the IMA, one of the main goals for the YouTube initiative was to change the reputation of the institution and reach out to the community, so several of their videos are employee profiles which help expose audiences to the diverse human side of the institution. SJMA uses many of its videos to support the exhibitions (with some of these available to visitors within the museum), so the artist interviews help visitors hear what the artists have to say about their works and processes. Their tagging of 27 videos as “Other” refers to exhibition tours; in these cases, the video functioned as its own entity, giving greater knowledge of the artist, but also offered greater insight into the exhibition that it accompanied. 

Community Interaction on YouTube

Much has been hyped about the community aspects of sites like YouTube and the opportunities for audience interaction and “user generated content." How do museum channels reflect this ability, and how much interaction occurs?

All of the museums allow comments and video embedding for outside sites, and all but one allow video responses. Half of the museums require a review before comments are published, and the other half maintain an informal review on comments as they arrive at the site.

Comments 

For the combined 174 videos on the five channels (as of mid-January 2008), there were 625 comments, and the per-video average was four comments. However, those figures may be inflated due to seven videos having over thirty comments each; when those seven videos are excluded, the total drops to 247, and the average number of comments drops to two. It should be noted that most of the institutions are quite aggressive in filtering out spam messages so the totals are reflective of actual comments and not of spam.

It appears that overall, museum videos generate a small number of comments, and that the dialogue is not universally as active as the term “social Web” suggests, at least not at this stage and in this format.

Video responses

For the combined 174 videos on the five channels, there have been a total of forty-two video responses, but only through two of the channels; additionally, the majority of the responses have come from one user. So it does not seem that users are actively submitting videos to these channels in direct, unsolicited response. The occasions on which responses have been effective seem to be when a direct call for submissions has occurred, as in the case of the MoMA project with The Residents art group (http://www.youtube.com/profile? user=theresidentsmoma) or the Brooklyn Museum’s calls for Visitor Videos for a competition (http://www.youtube.com/ view_play_list?p=3083AC21010A8F2E); the Walker has recently asked for submissions (http://www.youtube.com/walkerartcenter) and it will be interesting to see the results of this project.

Views per Video

A great emphasis is often placed on the number of views per video for YouTube; however, this focus tends to be unrealistic for museums and can be slightly distracting from an institution’s goals. Museums are probably never going to produce the same level of traffic as other types of YouTube videos; of the institutions surveyed, only one video broke the 100,000-view barrier, and some of that traffic might have been due to the fact that it involved celebrities, a well-known artist, and an extensively promoted exhibition.

Certain trends are noticeable within the traffic, however. The following charts show total hits for each video for each institution, based on their own traffic scale for the Y axis; therefore the slope of each institution is relative only to itself. Further, the amount of traffic is cumulative, so earlier videos (to the left) may be higher due to results over time.

The results imply that institutions tend to experience an initial, uncharacteristic spike near the beginning of the channel launch, and traffic thereafter settles into a more reliable average with occasional spikes. 

Figure 1

Fig 1: Views per video for institutions

Views by Time

There are many theories about how long the ideal length of videos on YouTube is, and what effects occur based on length. In our group, we’ve correlated time with views to see if the time has a clear impact on the overall views.

Overall, 124 (71%) of the total 174 videos were under 5 minutes in length and had 650,481 views (81%) of the total 806,426. The highest number of views and highest average fell in the 1:00 – 3:00 range, but good values were present at other levels.

VIDEO LENGTH 0-0:59 1-1:59 2-2:59 3-3:59 4-4:59 5:00-9:59 10 minutes and greater
# of videos 22 32 28 25 17 41 11
Total Views 806426 47153 206994 243898 96649 55787 103468 52477
Average Views 24605 16977 11055 10280 3407 4859 3070 4801

Table 2: Views by Video Length

Institutions did not seem to have consistent ranges for their most and least “successful” videos (in terms of views):

VIDEO LENGTH 0-0:59 1-1:59 2-2:59 3-3:59 4-4:59 5:00-9:59 10 minutes and greater
Lowest View Count for Institution   1 1     3  
Highest View Count for Institution 1 1 1     2  

Table 3: High and Low Views by Video Length

Further research needs to be done before making any specific correlation between length and views. Institutions are experimenting with different video lengths, and the results from these experiments have varied for each institution and across the entire group.

Traffic Comparison

The following chart compares the traffic for each of the institutions, comparing their own estimated statistics for Web site traffic and physical traffic with the number of views for their YouTube channels.

The first column shows the total number of views as of mid-January 2008. The second column shows the total number of views as a comparison to the average annual visits to the institution. The final column shows the views as a percentage compared to average annual Web site traffic as reported by the institution. 

  YouTube views Views as a % of physical visits Views as a % of Web traffic
Exploratorium 58664 11.1% 0.326%
Smithsonian 3066 0.4% 0.139%
IMA 64118 14.2% 16.030%
SJMA 59095 118.2% 47.276%
MoMA 621483 27.0% 8.878%

Table 4: YouTube views compared to physical visits and Web traffic

As seen, the relationship between the size of views for an institution and its other types of traffic is variable. For example, the SJMA did not have the highest number of video views of the group, but their number of views in comparison to their physical traffic is highest. MoMA had the highest number of video views, but in comparison to total normal traffic to the museum's Web site, it had a lower rate than other institutions. This also highlights differences in traffic types within institutions. The Exploratorium has a significantly higher number of on-line visitors than physical visits to the institution, so in comparison to total Web traffic, the influence of YouTube seems to be lower than in comparison to physical traffic.  

It will be interesting to see how this metric develops and if this type of monitoring is valuable in evaluating the success of social Web projects.

Outcomes

There is much discussion about possible outcomes from the use of a social video sharing site like YouTube. However, after a year or more of participating in this experiment, do institutions believe that the potential has been realized? Below is the feedback assembled from the panel on the overall impact of the institutions' YouTube initiative.

"Our YouTube videos___" Strongly Disagree Disagree Neutral/Not sure Agree Strongly Agree
Have increased traffic to our Web site (DIRECT clickthrough)   1 3   1
Have increased traffic to our Web site (not from a direct hyperlink from YouTube)     2 3  
Have moved traffic away from our site (as users click on video or related videos) 1 2 2 1  
Have increased physical traffic to our museum     2 3  
Have helped introduce new audiences to our museum and resources       3 2
Have helped our core audience(s) get a closer connection to our programs/services       3 2
Have helped to explain a difficult topic/concept/artist in a way that other resources couldn't       3 2
Have increased revenue (in our store/products/etc)   2 1 2  

 Table 5: Outcomes of using YouTube

In the end it appears that posting video content on YouTube benefits the institution, and that the potential risk of lost direct Web traffic has not necessarily dissuaded the participants. The purported benefits of generating income are not strongly agreed with, but the increase in traffic appears to be bearing fruit; however, this statistic is hard to quantify specifically. Visiting information for museums, and even more so for museum sites, tends to be influenced by a large number of variables. One way to track increases in traffic specific to YouTube is to track links on YouTube that are returning to the generating institution's site via site metric software. YouTube provides traffic information on where links are coming from, and this can help identify sources of traffic from an institution site as well as other sites and blogs.

Often, it is more qualitative feedback that will prove useful. For example, at the SJMA, the videos are featured on YouTube and iTunes (like the other institutions) but are also featured on video iPods that are available for checkout within the galleries. Visitors have often made comments about wanting to see the artists whose videos they saw on-line, so the assumption can be made that physical traffic is being influenced by the on-line video initiatives. In another case at the SJMA, a catalog for an artist featured in their videos, including those on YouTube, was the first catalog to sell out at the institution during its exhibition run.  Many of the sales were made over the phone, where the purchasers commented about seeing the video of the artist on-line.

The chance to present the institution to visitors in a new and fresh way both on-line and in person cannot be understated, and all of the institutions featured highlighted this potential. For the IMA, one of the great draws of YouTube was combating their local image as a “stuffy institution that did little to reach out to the community.” As part of the larger mission, YouTube was a good way to for them to reach out, market, and educate in a very non-traditional way.

Challenges

So what are some of the challenges this group has encountered with their YouTube initiatives?

  • Internal/external restrictions. For example, the Smithsonian has institutional limitations on the production and usage of video. At many of the art institutions, rights issues are a concern for works featured, and many videos have to be abandoned or only posted with specific limitations of time and placement. These limitations can result in less emphasis being placed on creating new video content and devoting any budget or staff, thus limiting an initiative's ability to grow and develop fully. 
  • Compression and resolution. YouTube has a very specific type of encoding, and video producers need to experiment to get the best results. Often, the suggested formats and specifications on the site will not produce the best results. Video producers should research specialty sites such as Ken Stone’s Final Cut site (http://www.kenstone.net) for more information.
  • Managing Comments. Many of the featured institutions lamented the amount of time and effort required to keep comments on topic, or at least spam-free. The      comment management tools on YouTube are insufficient to deal with this issue, and the amount of spam is daunting.
  • Getting traffic. As we’ve seen, there is no easy formula for getting traffic to your channel or to a specific video. Producers need to make sure that others are aware of expectations, based on previous traffic, and be realistic. Consider also the quality of the traffic; if your goal was to produce a video for use by teachers and you’re hearing from the target audience, then the number becomes less relevant.
  • Learning on the job. Few museum professionals come to the table with an understanding of dealing with ambient noise, correcting white balance, or other film and video production techniques. Luckily, the tools are relatively easy to learn, and with experience you become better and better. If you can afford to purchase the better equipment, it will make your life a lot easier. Costs might be daunting initially, but in the long run it will save you hours of tweaking sound and video.
  • Time. As with any on-line initiative, time to create and edit content is valuable, and a luxury that many museum staff do not have in excess.

Advice

  • Find communities of interest for your content. Realize that most visitors to YouTube aren’t inherently going to search for the term “museum,” and are instead interested in topics and looking for contents around those topics. For example, when MoMA released videos of presentations by the artist Swoon, contact was made with the blog Wooster Collective to possibly promote it; when another blog was featuring videos on the nature of painting, links were sent for videos with artists discussing this question. For the IMA, one of their most successful series featured calligraphy artist Hirokazu Kosaka. Based on the comments (incidentally, their most commented-upon videos), the people watching this were not searching for "museum" content; they were searching for "calligraphy" content. While these searches for calligraphy content may not result in direct in-person museum visits to IMA, they do provide the IMA with good exposure, and provide the IMA staff with the information that people are going to YouTube to find "How to..." videos very specific to their interests. Because of this experience, they’ve been able to consider other audiences (such as art teachers, college students, and donors) for later features and have adapted production techniques to focus on these audiences. For the Exploratorium, some of their most successful videos featured people solving Rubik’s cubes: it began with a solid concept that was informed by the audience interest in videos of that type that were circulating on-line (including a few by film director Michel Gondry) and thus had a better chance of gaining momentum through viral marketing. SJMA has frequently targeted niche bloggers to get mention of their videos, thus driving traffic to their YouTube channel.  In one such case with underground artist Camille Rose Garcia, a cartooning blog named Drawn (www.drawn.ca) and the well-known blog BoingBoing (www.boingboing.net) accounted for a large part of traffic to that series of videos.
  • Integrate with your other sites and services. The ability to embed videos opens up the possibility of easily adding content within your sites in a number of different ways. For example, MoMA has posted exhibition-related videos on exhibition pages, as well as trailers for films screening at the museum, within their on-line calendar. IMA has embedded their videos within a specific area of the site, but has placed visual links in other areas of the site to drive traffic. SJMA has posted YouTube players as well as links to podcasts within their exhibition descriptions. Use other e-news, RSS, or blogs that are traditionally connected to your audiences to promote this connection and to gain exposure.
  • Build an internal base, and help your institution understand the project and process. Inside the museum, work with people whom you feel comfortable with and who will be flexible and willing to experiment. Start quietly with a core group until you have the project in hand, and then build a broader coalition. There will be many people who have not heard of on-line video sites and/or do not know their capabilities and limitations, so be sure to spread the knowledge around. 
  • Start small and grow as needed. Most of the institutions featured began with very little equipment and took a chance.  One of the SJMA’s first videos was a promotional video for an artist named Il Lee who creates large-scale abstractions with ballpoint pens.  For the video, producers shot a close-up of a hand scribbling on paper, and later, during the editing process, interspersed still shots of the artist working and his artwork.  For audio they found some legally available contemporary music that worked well with the “flurry of scribbles” going onto the page. MoMA launched with a series of videos created by artist Doug Aitken to promote an upcoming show, but continued to create videos using one camera, one laptop, one staff member, and one intern, and then built up production and technology as growth required. At the Hirshhorn, the videos were shot and edited by an intern.
  • Be creative: try out new approaches to presenting the materials while using your institution’s strengths. The capabilities of video are quite dynamic, and the production times are much shorter with digital equipment, so innovation and experimentation can come easily. Use methods and processes that create a unique look and feel for your videos, but watch what other institutions are doing for inspiration. Take advantage of your institution’s strengths; for example, SJMA has a close connection in the shared areas of Education and Experience, so it is possible to offer the videos in multiple areas and include the content in planning for exhibitions. At the Exploratorium, there is a great sense of experience and play in the layout and execution of the institution, so the on-line experience can be fluid and dynamic and capture the sense of excitement,
  • When possible, assign one person responsibility. As with any social initiative on-line, the daily work of maintaining the presence can take time. Determine what you want a YouTube channel to do for you, and dedicate a person or small group to the YouTube channel. This way, one person can make sure that the initiative continues, keep track of institutional goals, and build relationships while keeping up with comments, responses, etc. Keep your page up to date and add new content only when it makes sense for your goals and when video is the correct media for it. If you are adding more than one file at the same time, space them out over days or weeks - often only the last video posted will get traffic, and users will not backtrack to earlier entries in the series.

References

Dance, K. and S. Pau. (2007). "Vodcasting: 5 Easy Steps to Film an Interview and Get It On-line in a Day!" In J. Trant and D. Bearman (eds). Museums and the Web 2007: Proceedings. Toronto: Archives & Museum Informatics, published March 31, 2007, at http://www.archimuse.com/mw2007/papers/dance/dance.html

Caruth, N. and S. Bernstein. (2007). "Building an On-line Community at the Brooklyn Museum: A Timeline”. In J. Trant and D. Bearman (eds). Museums and the Web 2007: Proceedings. Toronto: Archives & Museum Informatics, published March 31, 2007, at http://www.archimuse.com/mw2007/papers/caruth/caruth.html

Incandela, D., et al., (2007). "Starting a Digital Revolution". In J. Trant and D. Bearman (eds). Museums and the Web 2007: Proceedings. Toronto: Archives & Museum Informatics, published March 31, 2007, at http://www.archimuse.com/mw2007/papers/incandela/incandela.html

Madsen-Brooks, L. (2007) Percolations: Museums and Social Networking Sites, Part IV. July 8, 2007. consulted January 14, 2008. http://www.museumblogging.com/2007/07/percolations-museums-and-social_08.html

Cite as:

Alexander, C., et al., Beyond Launch: Museum Videos on YouTube , in J. Trant and D. Bearman (eds.). Museums and the Web 2008: Proceedings, Toronto: Archives & Museum Informatics. Published March 31, 2008. Consulted http://www.archimuse.com/mw2008/papers/hart/hart.html