In part 3 of the series, the authors walked us through the key steps in planning complex projects.
With a plan of attack in place, it finally is time to start on development and implementation. The complexity of your project will depend on several factors; but, no matter whether you are implementing an out-of-the-box or highly customized solution, substantial KM technology projects usually involve an intensive period of technology development. Enterprise search – aggregating information from multiple sources and presenting it in a user friendly way – is inherently complex. So, we knew that implementing our project would not be easy.
KM Involvement in Development and Implementation
Even with an out-of-the box solution, you will have options for how different aspects of your project can be implemented. This becomes more complex if you wish to customize the solution. In our case, we increased the complexity by deciding to implement a new taxonomy tool and KM content management tool in parallel with our search implementation.
In the throes of a complex project, forgetting what you have decided and why is easy to do. Combined with developers bombarding you with questions regarding the functionality you want, this can quickly overwhelm you. As observed in an earlier post, having documented detailed requirements was our saving grace in these situations again and again.
A difficult reality in development and implementation is making compromises. Even with the most careful planning, nothing goes perfectly and being told that the design you have envisioned cannot be achieved is frustrating. What do you do? You go back to your design principles. You may recall that in our search project our overarching theme was a relentless focus on user experience. So, we always chose options that would best support the user experience, even if those options were not perfect.
Because very few KM projects involve stand-alone technologies, a heightened demand for tighter integration across multiple platforms and data integration become key aspects of most projects. In our implementation, we integrated as many data sources as possible into search, which proved critical to improving relevancy. Those data sources included our InterAction contact management platform, Aderant financial system, matters and experience database, website, blogs, document management system, and KM content management system.
But, integrating data from multiple sources brings its own challenges. In our case, data was sometimes populated from the wrong system or from the wrong field in the right system, which required a bit of detective work to sort out. Another challenge was working with the owners of some of the data to address inconsistencies that the search project revealed. Enterprise search and other technology projects cast a giant magnifying glass over your data (and its warts). You need to be prepared to address issues that might arise.
Manage expectations: Highly visible, significant technology projects often lead to high expectations. Throughout our development stage, we managed stakeholders’ expectations with regular updates and early demonstrations.
Tune and tweak: Do not be afraid to spend time making small tweaks that have a big impact. One of the most valuable efforts we put into our project was fine-tuning relevancy. Although taking the time mid-implementation to focus on relevancy was difficult, doing so significantly improved the user experience.
Master your metadata: One reason we were able to identify incorrect data was that we had prepared a detailed data map listing all of the metadata appearing in the search application, including what it should look like and where it was expected to come from. We revisited this (and other) documentation throughout the development stage.
Coming up in part 5: In our next post, we look at the role KM can and should play in testing complex KM technology projects.
By Andrea Alliston, Partner, Knowledge Management, and Nicola Shaver, Director of Knowledge Management, Stikeman Elliott LLP