Intranet Improvement Proposal

In 2003 Chris Reed of BlueLightning Studios and I collaborated on a speculative business proposal to Solectron about why and how they needed to improve their callcenter intranet and knowledgebase (KB). Chris, having worked there between contracts, knew what condition the intranet was in. Working at HP had taught me that KBs have to be maintained, and it’s best if they’re maintained by the people using them,1 because nobody else has the same incentive or expertise to keep it up to date. Therefore we recommended turning the KB into a wiki with some editorial controls, and using an enterprise-strength wiki engine on a Debian Linux server for maximum value at minimum cost.

Intranet Improvement Proposal (PDF, 1.3 mb)

The Long Version

When Chris worked there, the Solectron callcenter in Belleville had a rudimentary intranet and KB, but the de facto way to get information was to ask your neighbour or a second level tech (henceforth a “second”) for advice. Having worked in callcenters before, Chris knew that this was an inefficient, unreliable and unscalable way to do research. When he and I had worked together at Compaq, we had a comprehensive, up-to-date KB, which made it much easier and faster to support our callers. Personal advice is a wonderful way to get answers, but often those people are busy, and you could look the answer up for yourself.

Chris smelled an opportunity to use his web programming skills. He drafted a proposal and sent it to me2 for review. While correcting his spelling and grammar, I saw that the structure needed some work, so I made an outline and took it apart. I had such a good time revising his ten page draft that I spent the next four months working on the proposal, expanding it in complexity and scope. (This should reveal how much challenge, excitement, and joy I was finding in my own job by that point.) The final version was over seventy pages including appendices. I handled all the writing and editing, the majority of the research,3 and designed many innovative features for the intranet.

There were some hurdles waiting for us. Except for Chris, I had no techs from Solectron to mine for information. Because I had a full-time job already, all work on the proposal had to be done in my spare time. Biggest was that neither Chris nor I had tackled a project of this size, type, or complexity before.

Chris and I decided almost immediately that a good KB would use a content management system (CMS). I pushed hard for a wiki because having the techs write material would keep the KB comprehensive and up-to-date, but cost less than the editorial staff it would otherwise take. So, more important than the perfect CMS would be to reform the callcenter’s culture so that the techs expected and contributed high-quality content. We believed that using an existing CMS would let us put the programmer effort into the stuff that had to be customized, like the database design, live reports of queue statistics, webforms, and a report designer application.4

Just writing the proposal for this was a big project, let alone the thing itself. We needed to understand the needs and wants of not only the techs, but also of management, human resources, and training, and accommodate them. We had to research content and knowledge management, information architecture, usability, operating systems, scripting languages, databases, and content management systems. It was the first thing that I’d really gotten my teeth into since finishing Humanities, and bigger than anything I did during my first degree. I’m sorry that Solectron wasn’t interested in making any changes, because it felt wonderful to do real work instead of just unlocking passwords, and I wanted more.

  1. This was before Wikipedia became a household word and people admitted that user-maintained content might be worthy of consulting. Only geeks knew about wikis at all.
  2. His friendly neighbourhood English expert.
  3. At least 75% of the research, planning, and writing. This wasn’t as unfair as it might sound, because Chris would have been doing at least 75% of the implementation. I understand usability, technical writing, content and knowledge management, and know more about *nix package management systems. Chris understands various web programming languages, databases and *nix system administration.
  4. Roughly a year later, during Excel XP training, I found out that my report designer was frighteningly like Excel’s pivot tables, which I’d never seen before.