Andrew Lewis outlines the data-driven strategies for delivering web and mobile services to audiences at the V&A, the importance of digital asset-management and making data web-portable. Real-life examples are used to illustrate how this strategy is translated into services that adapt flexibly and efficiently to audiences’ rapidly-changing and diverse range of digital devices, and their increasing digital expectations
How can we continue to delight and inspire with our cultural collections and accumulated knowledge as our audiences’ lives, behaviour and expectations are constantly being changed by digital?
Digital is now everywhere in people’s lives. With lots of frequently-changing digital options,and short technology lifespans, people interact and behave in less easily predictable ways. This means investing heavily in specific technologies becomes riskier making service-development more challenging. There is a huge range of diverse consumer devices available and the capabilities of devices like phones, cameras, computers, TVs, and even cars and watches have become blurred. The ways in which it is possible to access our services digitally evolves, or perhaps mutates, at a rapid pace – which is challenging.
Here, I look at some simple strategic principles that can help you deal with the challenges of digital change. The main focus is the strategy of identifying digital assets and making them web-portable as data via APIs (application programming interfaces). The V&A maintains and manages a number of significant repositories of unique information and researched knowledge. Using APIs for data exchange maintains the integrity of that authoritative data and makes it deliverable for easy multiple re-use. This has allowed the V&A to deliver a range of digital services in a way that is efficient, scalable and adaptable.
However, while strategies are essential, they are worthless unless they are delivering your organisation’s mission, so these are illustrated with examples of real V&A data-driven services delivered over the last five years, which demonstrate how these strategic principles have been implemented in practice. I hope to show the benefits you can gain, the challenges posed and the organisational culture and processes that need to be in place to take advantage of them.
Authoritative data made portable
The V&A Digital Media department is responsible for front-facing digital services on the web and mobile. Our strategic development principles can be described simply: Always use the museum’s authoritative data sources as raw materials, make this data portable so that it can be used in multiple ways whilst still maintaining its integrity and then design services to suit the needs of audiences in specific contexts.
Having a reliable source of digital assets as the basis for multiple reuse is not a new idea. Terms like “Create Once, Publish Everywhere” popularised by Jacobson1 and others have been around for at least a decade – e.g. Business Wire2 – and the principles even longer. More recently, the web API – Application Programming Interface3 -concept has established itself as a well-understood and widely-adopted business strategy for the efficient and flexible redistribution of digital assets. Services such Google, Flickr, Twitter, Amazon and Facebook all create and make use of web APIs to distribute their content across multiple platforms and devices.
Figure 1 shows a greatly simplified model of the processes discussed in this article, showing the internal managed knowledge-repositories, some of the range of V&A front-facing digital products and services, and the APIs which are the key to efficient delivery.
A key aspect of this approach is that the creation and management of the museums information and knowledge is devolved throughout the museum to the specialist staff who are experts within each area. This includes curatorial or staff for collections knowledge, V&A Enterprises staff for commercial products or Visitor Services, Learning and Marketing staff for events. These staff create, maintain and manage the resources that are so valuable when made portable for re-use. At the V&A, there is an active Asset Management strand within the V&A’s Digital FuturePlan, an organisation-wide development programme, which is looking at embedding a digital asset-management culture into all aspects of the Museum’s work.
Figure 2 shows a simplified process that represents the process of creating digital assets. In this example the expertise of curatorial and other museum staff maintaining the central collection catalogue accumulates information and knowledge creating standardised structured records that can be exploited later.
The socket on the right is a metaphor for being able to access the data. Without this, these assets are effectively locked in silos. To be re-usable they needed to be made portable which we do by plugging the data into APIs.
Collection records are a common and obvious source of digital assets in museums, but the concept applies to any source of digital records. Data in each of your core authoritative organisation-wide systems should be considered for use as digital assets. This might include event databases, online shop databases, media repositories for audio/video/multimedia, even HR databases as a reliable source of staff names and titles (see Figure 3).
It is a simple, but extremely useful exercise to audit the databases in your organisation. If you’re lucky there will be one system for each core function, and it will have some means of interchanging data ideally with a built-in flexible API. In reality, you may have some information systems like this, but other sources that will have data that will need to be extracted by other means before they can be made portable. If you have systems where this is not possible your data is locked in, it prevents efficient re-use (in Figure 2, this would mean data without a socket). This will effectively limit the potential value of your organisation’s assets. Of course, you may also have some information not even kept in digital formats, which is a bigger challenge.
Digital auditing allows you to identify what you have, and how much of it you can actually use, how you might start addressing any limitations and how your culture and processes need to change to adapt to a managed digital asset culture in your organisation.
The development of portable-data-driven digital services at the V&A
The V&A’s mission is to be the world’s leading museum of art and design, to enrich people’s lives and inspire individuals and everyone in the creative industries, through the promotion of knowledge, understanding and enjoyment of the designed world. This means trying to understand the needs and typical behaviour of our audiences in their lives and giving them what they want, delivered in ways that suit their contexts. Easy to say, but what does that really mean in practice?
In 2008, the V&A Digital Media department (Online Museum at the time) with close collaboration with the Collections Management department, developed its first API, the Collections API4 believed to be the first of its kind by a museum service. This was a standard REST – Representational state transfer- data request service5. This used JSON – JavaScript Object Notation6 – to serve the data, as this is an open format that was judged to be most widely adopted by web developers.
This seemingly obscure event marked a fundamental shift in the V&A’s digital strategy. The Collections API took object records from the internal curator-managed collection database, and turned them into portable web-ready digital assets that could be used to power new services.
The first of these was the V&A’s public collections service “Search the Collections” re-launched in 2009. Newly powered by data from the new Collections API, the number of publically available collection records in Search the Collections grew overnight from 55,000 to more than a million simply by turning the existing catalogue data into portable web-ready digital assets.
Figure 4 shows a typical public record for a chair in the V&A’s furniture collections. All the information on display is taken from the curator-maintained museum collection catalogue.
As an example of the cultural change needed to support the new data-driven service, we introduced a policy that web pages would no longer be about single museum objects created by manually editing in the web content management system. After the re-launch, if a member of the museum staff wanted to publish new information about an object, they simply updated the record and it would be published automatically as a Search the Collections page, ensuring audiences had the most up-to-date information available.
This is a powerful management model. It means that Digital Media content-editing staff resource is not needed to publish basic collection pages. This adds scalability and flexibility. If the museum has an unexpected opportunity to resource more cataloguing, there is no web publishing bottleneck. It just gets published in the routine web update cycle.
Portable digital assets
The new ‘Search the Collections’ interface delivered a huge increase in direct access for researchers in art and design, but the real innovation here was the API itself, which paved the way for other services to be built on the same data and meet other audience needs.
Search the Collections is primarily a research tool. It was designed for the specific user needs of discovery and inspiration. It gave people the means by which they could directly research vast amounts of information about the collections,which was previously only available within the internal V&A catalogue.
In 2010, we released a mobile version of Search the Collections to suit the different user experience of searching on a mobile web-enabled phone. The API allowed this to be developed in about a week by producing a lightly featured mobile display over the API.
This rapid service development is impossible using traditional data entry in a web CMS. A data-driven approach has none of the maintenance costs of maintaining and synchronising the information within two sites. The same data is simply being served differently to suit the different user needs of search on a desktop and a mobile device.
In 2011, we launched a newly designed V&A website, which took the principles of multiple presentation of data even further. The new site used a new Content API to convert the body of rich web articles into portable assets in the same way as the collection records had been previously made portable in the Collections API. Source data from the V&A shop-database and the main events-database was added to the content API. All the articles were classified using a simple data model7 and this classification data was used to aggregate content by subject from various sources and then display it automatically on web pages.
Figure 6 shows these single-source, multiple presentation principles in action on the main Furniture subject-page on the V&A website. The automated presentation of data is on the right hand side of the page highlighted with boxes. From top to bottom, these are: the content API matching other related articles, a grid of images showing items of furniture from the collections data base, a call to action to buy a V&A furniture book from the shop database, and a furniture-related museum event from the What’s On database. The chair previously shown in Figures 4 and 5 appears on this page in the image grid, fed from the Collections API.
Some of the items displayed on the page in Figure 6 are from digital assets from events database and collection database records. These also appear in other places and are displayed in other ways to suit users’ different needs. For example, by 2012 the number of visitors accessing our website via a mobile device had grown significantly and was increasing. We were able to rapidly make the website responsive to screen size for these mobile users. We simply developed our pages to actively change the display to suit the size of the screen being used. The assets from the Content and Collections APIs are the same, but displayed to suit the context.
Here you can see events from the museum’s event management database, re-presented in a mobile-optimised What’s On search result. The top result is the same event that was automatically promoted on the Furniture page in Figure 6.
Digital Labels Interface
In 2012, the new Furniture Gallery opened. One of the challenges was to show object information from the collection for many objects on display whilst taking less physical display space than traditional printed labels would. We developed a new digital labels interface with the project team to display collection information for multiple items of furniture in the displays. These not only took up less space, but allowed remote updating by curatorial staff.
The chair that appeared in the examples above is shown in Figure 8 in situ in the V&A Furniture Gallery, with its corresponding digital label.
The final example is the V&A’s Digital Map, which we launched in a beta version in early 2013. This is an interactive web replacement for the inflexible PDF map. It was built with mobile users in mind, optimised for tablet users, and responsive for mobile phones. It uses very small file-size SVG vector images files for fast downloading. The map is enhanced with information about events in the museum and objects in the rooms, which are yet another way the same digital assets are re-utilised. Figure 9 shows the same chair within the context of visitor orientation within the building. The darker pink room silhouette is a room that has been touched on the screen. This has then caused the map to display the furniture objects that are in that room. The list of items is created automatically by checking the collections API for what is in the room, then serving up the relevant digital assets to display the latest information available.
The examples described so far in this article show the power of this strategy for V&A web services and features. However, the principle of portable web data also applies when we wish to give access to our collections alongside objects from other museums in collaborative 3rd -party services. For example, with our public Collections API, we expect to supply access to our data to aggregating services.
At the time of writing, we are talking to CultureGrid about how they can write a plug-in that converts our API data into the LIDO8 standard. It is hoped in time to allow our collections data to be automatically syndicated into services such as Europeana. Currently this data export has been done by staff manually querying the internal database, which requires a large staffing resource making it inefficient and non-scalable. It also does not update automatically, risking data integrity as we enhance the records in the future.
Summary
It is hoped that the examples here illustrate the effectiveness of the Digital Media delivery strategy at the V&A. We have found significant audience and organisational benefits by applying it. I would strongly recommend others to investigate applying these principles with an API-based approach. Start by identifying authoritative data repositories in your organisation and what you can expect to be easily made portable, in formats that are favoured by web developers. Portable digital assets are the key to creating a flexible and scalable platform that will enable you to deliver new services for your audiences across the many unpredictable ways they choose to use digital.
Andrew Lewis
Digital Content Delivery Manager, V&A
Read more about the V&A’s digital strategy and approach on the V&A Digital Media blog, which is aimed at museum professionals who are developing digital services: www.vam.ac.uk/digitalmedia
References
All urls last accessed 15 August 2013.
1. Daniel Jacobsen. COPE: Create Once, Publish Everywhere.13 October2009
http://blog.programmableweb.com/2009/10/13/cope-create-once-publish-everywhere/
2. iXL and Team of its Alliance Partners Launch Strategic Initiative for Media Delivery to Web, TV and Wireless Applications. In PRNewswire. 10 April 2000. Available at
http://www.thefreelibrary.com/iXL+and+Team+of+its+Alliance+Partners+Launch+Strategic+Initiative+for-a061800408
3. Wikipedia. Application Programming Interface.
http://en.wikipedia.org/wiki/API
4. V&A Collections API – Principles and Getting Started.
http://www.vam.ac.uk/api/
5. Wikipedia. Representational State Transfer (REST).
http://en.wikipedia.org/wiki/REST
6. Wikipedia. JavaScript Object Notation (JSON).
http://en.wikipedia.org/wiki/JSON
7. Andrew Lewis Mixing It Up: Developing and Implementing a Tagging System for a Content-Rich Website Which Uses Aggregated Content from Multiple Sources . In J. Trant and D. Bearman (eds). Museums and the Web 2011: Proceedings. Toronto: Archives & Museum Informatics. Published March 31, 2011. Consulted August 20, 2013. http://conference.archimuse.com/mw2011/papers/mixing_it_up_tagging
8. International Council of Museums (ICOM). What is LIDO (Lightweight Information Describing Objects)
http://network.icom.museum/cidoc/working-groups/data-harvesting-and-interchange/what-is-lido/
