Wednesday, June 11, 2008

Building a Publishing Platform One Product At A Time

Alex Humphreys, Director of Business Technology Services, Oxford University Press offered an interesting glimpse into the actual experience of a publisher transitioning products onto a new platform. Working with Mark Logic since 2004, this scholarly enterprise was working with journals, monographs, major reference works, etc. -- not necessarily straightforward with fairly stringent expectations for discoverability. Internally the hope was that the adoption of MarkLogic would enable OUP to deliver subscription based information services to the market rapidly and cost-effectively. The past two to four years has shown them that owning one's own platform so that you can deliver product using multiple vendors (on top of the Mark Logic server) is not without its challenges. Humphreys reviewed half a dozen high-quality information services in order to demonstrate how each experience in delivering a unique service drew on previous investment in time and staff resources. These diverse services included AASC, Oxford Language Dictionaries Online, Oxford Music Online, UK Who's Who, and similar high-profile library products.

The unique requirements for each product drove development of graphical templates and display options. However the fact that they were still being built across a single platform helped when Oxford chose to work with a variety of vendors (Semantico, IFactory and CDSI). That enhanced the speed of product development. The standardization saved them both time and money, spread staff resources more effectively across products. Development time frame has shrunk, allowing kickoff to launch in five months for some projects.

This held true even when they began work on non-reference information services, such as InvestmentClaims (, solving slightly different problems of discoverability and consumption of content. Investment Claims at present is a PDF-content-oriented site, although new content is being prepped for the service. (Alex noted that this was the first instance of using "bootstrap" code with a vendor. That resulted in shorted development time, measured in days or weeks rather than months.)

To his credit, Alex was honest about what they had seen in terms of costs in building and maximizing the benefits of the platform across multiple projects. His bullet points were these:
  • per-product development costs have not yet decreased due to the unique requirements of each product and its content
  • The iterative platform development might not have saved money to develop a platform but allowed them to spread those costs over time and multiple projects.
  • As the platform matures, we do predict development cost savings.
In closing, Alex noted these lessons learned in the Oxford University Press experience:
  • It takes multiple products before a feature becomes truly generic.
  • Easy to underestimate the soft support costs for a system.
  • It takes a strong stomach to own your own platform
Solid experience from a traditional scholarly publisher who seems to be thriving in the new publishing environment.

No comments: