Paul Kramer

1. Metadata workflows

Metadata workflows

Six months ago, before my boss volunteered me to join a working group that was developing the metadata standard for the publishing industry's growing business of content syndication, I thought that XML was a clothing size, somewhere between Large and Extra-large. Now, half a year and a half-dozen PRISM meetings later, I know that XML has something to do with file markup. But coming from the editorial hemisphere of the content-providers’ world, I will mercifully steer my discussion away from the technology minefield of DOIs, DTDs, RDFs, and URIs, and aim instead at a general description of industry practice, present and future.

After 12 years as an editorial production director, nine of those at American Vogue, I have seen how a typical magazine publisher conducts business. And I have an idea of how a content provider’s aggregation, manipulation, and syndication practices could be conducted faster, smoother, more efficiently, more cost-effectively, and with a greater return on the investment of resources, with properly enabled tools.

Currently, the only types of assets at Conde Nast that are stored for possible reuse are the hi-res image files and page-layout documents that are used to print the magazines. There is no mechanism at this time for preserving individual text elements--articles, photo captions, recipes, sidebars--as separate, reusable assets. The only metadata stored with these two types of assets--apart from file descriptors such as file name, file type, file size, creation date, etc.--are Issue, Issue Date, and Page Number. With our present system, the searching for a particular asset is done manually, and the GUI is the printed page. One flips through multiple issues of multiple titles, finds the story or picture needed, and submits the "locator" metadata to the asset repository (a call is placed to our pre-press vendor, who retrieves files from a production archive, searching against issue, date, and page number).

When it comes to syndicating published content to our foreign licensees, such as GQ Japan, Paris Vogue, and Vanity Fair UK, our partners have a limited-time option period to try and negotiate the rights to reprint, reuse, or repurpose content from the US. With our current system, a bound copy of the US magazine is sent to the foreign publication, where it is reviewed; then individual requests are submitted to our rights and permissions departments. for clearance. The time it takes for this selection/submission/negotiation/clearance process reduces the already narrow option window. The faster these transactions can take place, the more time the buyer has to make selections, and the more revenue-generating exchanges can be made for the seller.

Clearly, a better system needs to be deployed.

When aggregation, asset management, and distribution tools are PRISM enabled--when they can automate the embedding of standard core metadata, and preserve this metadata throughout the content lifecycle--our discovery, retrieval, and identification of digital assets will be prompt and painless, accessed via a robust set of descriptive metadata. Beyond the corporate firewall, these assets will be made available to partners, players, and the public via Web-enabled, searchable databases. Paris Vogue, CNN, or an individual consumer will not only be able to mine our content for specific types and sets of assets, but will also be able to access the necessary rights and cost information associated with each. Costly, labor intensive inquiries fielded by the contracts, rights, and permissions departments will be reduced in number to those that will most likely result in a return of revenue.

Additional efficiency will be gained upstream as well, as material--both commissioned and acquired--will be brought into the workflow complete with predefined, core metadata from partners and vendors that are using PRISM-enabled tools themselves. This will not only avoid the expenditure of resources on tasks already performed by others, but will enable creators to make more timely and in-depth searches of external sources of material. In the great wheel of an asset's life, PRISM facilitates its acquisition, use, resale, and eventual rebirth as someone else's content. Let the circle go unbroken.

But looking beyond the adoption of a publishing metadata standard, content providers will have to examine the upstream costs and downstream values inherent in the capture and aggregation of metadata as a function of the value of the asset itself. Many will be called, but few will be chosen: For every Richard Avedon, there are five hundred generic-looking product shots; for every John Updike essay, there are five hundred fluff-filled sidebars. Metadata clearly adds to the value of an asset, but this value is best profited from at the point of syndication. There needs to be an efficient system of assigning potential reuse value to an asset, and of routing it to increasingly robust--and therefore costly--annotation processes. Today it seems that companies choose between two extremes: tag all assets with a minimum set of mostly-automated metadata at the lowest possible cost, or tag all assets, regardless of reuse rights and likely future demand, with maximum descriptive metadata. The former limits the possibilities of post-publishing revenue, while the latter counterbalances all future resale profits with a heavy up-front investment of time and resources.

In a PRISM-enabled world, the attachment of standard core metadata will be automated for the faceless masses of digital assets and, through the application of extensible metadata triggers, the elite class of assets will pass through annotation workflow filters for an efficient allocation of resources for in-depth tagging.

An analogy that illustrates this present pair of extremes, and the PRISM-enabled Middle Way, is the "Wedding Gift Scenario":

You and your partner plan to have a wedding for the sole purpose of receiving gifts.

Your partner suggests inviting only a few special (rich) acquaintances, and holding the wedding in an exotic locale, the rationale being that the more money spent on a guest, the more expensive will be the gift given.

You suggest having a pot-luck, come-as-you-are, wedding in the park, and inviting everyone you both have ever known, the idea being that what the individual gifts lack in value will be offset by their sheer number.

But there is a problem with each of these approaches:

One: there are many guests—immediate family, relatives, close friends--that will have to be invited, regardless of their affluence; the expensive reception will be wasted on many who must be included, but who will nonetheless give meager gifts.

Two: Many people will give a non-negotiable gift, as opposed to cash, and one cannot combine a hoard of insignificant junk into a single valuable item, the way money can be aggregated in a bank account. No matter how many gifts you receive, just how many wall clocks do you need?

Or, in terms of editorial content: A workflow that infuses every photo caption and product shot with priceless gems of rich metadata is great in theory, but too expensive in practice; and populating just the core metadata fields on every asset is cheap, but has a limited value in return.

The solution to the wedding dilema comes from the out-of-the-box realization that a marriage ceremony is composed of multiple components, and the arrangements for one part can differ from those of another:

Have an informal exchange of vows in the park, inviting everyone you can think of. You will be given all the unexciting basics this way--the toasters, blenders, and fondue sets. Then, put on a lavish reception in the Bahamas and invite only the super-rich of the wedding ceremony guests. These few will bestow on you all the luxuries of fine china, silver, and bonds.

PRISM, an extensible metadata standard for the publishing industry, will make possible this type of creative-compromise solution. Multiple workflow schemes will be employed for multiple classes of content, so that assets can be routed to various annotation-workflow streams. All incoming and outgoing assets will be given the basic, “toaster” information for efficient B2B and B2C transactions; and based on repurposing value, advanced, “sterling” metadata can be captured and maintained for select assets where maximum research, repurposing, and resale capabilities are needed.

This is the under-explored territory where the publishing industry needs to venture; and PRISM--the publishing requirements for industry standard metadata--can clear a path for us through the content syndication jungle.

Previous

Table of Contents