Use Piloting when investing in digital solutions.
[ Technologies > Proof of Concept > Prototype > Pilot > Production ]
This Pilot project was undertaken as part of our Innovation Lab Services. The project was an important part a larger programme of work to digitise a corporate library containing thousands of historical newspaper articles, media clippings, and company annual & interim reports, and make them accessible and searchable via a corporate intranet.
With earlier Proof of Concept and Prototype projects completed to confirm technical feasibility of the WordPress platform and to deploy a full production-grade Pilot to help confirm technical and business use cases ‘in the field’ with a subset of users.
A Content-driven website
To recap, whilst WordPress itself is relatively straightforward for hosting simple websites, our client specifically wanted to use its content management capabilities to store an online digital library of thousands of PDF files and other media types (video, audio), make them searchable, and provide the user with the ability to filter documents in situ using keyword searching and being able to filter on several fields (e.g., country of origin, publisher, date range, author, document type etc).
The above image shows an early example of the UI where a number of filters were applied to the library to locate a specific newspaper article.
Data migration - processes and quality assurance
Whilst the earlier Prototyping allowed for unit testing using a subset of data, the Pilot allowed for considerably larger data volumes (from hundreds of megabytes, up to hundreds of gigabytes).
The Plan-do-check-act cycle - a four-step model used to ensure data quality
By conducting a full-scale migration of data (PDF files and associated metadata) we were able to better understand the nature of the data, and through this, provide greater detection of data errors (e.g., non-printable characters which were not present in the prototype test data). With this new knowledge, we were able to streamline our migration processes and build more stringent error detection capabilities within the migration tools used (predominantly python automation scripts). We also used this new knowledge to improve processes and tools/scripts used further upstream in the PDF document creation stages.
By adopting a highly automated approach to data management, multiple test cycles were possible during testing to initially simulate, and later actually realise, error rates within agreed tolerance levels of the business and its users.
End Results
This engagement allowed the business to simulate a sizeable part of the intended digital solution, and confirm that all success criteria had been met. With those milestones achieved, it now allows for progression to full-scale production with much greater confidence.
