Reed Stuedemann of Caterpillar's CatU (Caterpillar University) spoke at Tuesday evening's KM Chicago meeting, which was attended by a wide variety of people from the Chicagoland KM community. It was nice to "see" so many after lower attendances at the end of 2004. (I attended the meeting virtually via WebEx conferencing, along with several other participants and the speaker.) He discussed the communities of practice that have developed over the past five+ years as Caterpillar's Knowledge Networks.
The money quote of the evening: "If it doesn't support the business, then you are probably wasting your money." In other words, Caterpillar doesn't bother setting specific metrics around collaboration or usage of the knowledge network. They rely on solving business problems and having the Knowledge Networks system support that activity. Fantastic. This should be happy news to James Robertson, who has said 'Knowledge sharing' should be avoided.
Reed talks about this system as addressing business needs within Caterpillar. It grew out of a grass-roots effort in the late 1990's from the research organization when Caterpillar shifted from functional silos (engineering, development, sales...) to business units (machines, engines, financials...). They realized the need to focus on business process and problem/solution sharing.
The project moved into the CatU organization (part of HR) once the system moved to more global usage. The goal for CatU is "build people," and knowledge sharing has become a huge aspect of how Caterpillar helps build people. In relation to this, Reed showed the learning model as a triangle with sides of Leadership, Knowledge Sharing, and Learning Culture with "build people" in the center. Until 2005, the system was "free" to use by the business units. They now have to provide support based on how their people use the system. Based on evidence so far, this has not changed how these organizations are using the system.
There are now several thousand communities that are managed through the system. When asked how deep the usage of knowledge networks spans, Reed said that the CEO is a member of at least three communities (strategic planning), giving some idea that everyone in the organization is using the tool to help conduct business. He highlighted the value that new employees get from using the knowledge networks to come up to speed on topics and issues that directly affect their jobs. Reed also admitted that about 1000 communities are dormant and need to be disbanded. In fact, he stated that building a community lifecycle process is a challenge for him in the coming year. This would be very much in parallel to the content lifecycle process that is handled by each community manager.
In terms of value that the system provides, they did an extensive study that estimates the average value of every discussion is $600 with a wide variability. Two discussions saved more than the cost of the system for the whole corporation. They also discovered that those discussions which include their trusted partners resulted in a higher per-discussion savings than those that were wholly internal to Caterpillar.
Technically, the system is a web-based discussion and content interface, which is heavily integrated into email. Active members of any community (manager, experts, members) receive email from anything posted into the community. Many of the processing activities are enabled through email as well, so a community manager needn't login to the system to approve a new person's membership, for example. Other participants can login to the website and view the discussions, but can't participate until they become members. People are not given access to all communities. They must request the community manager, helping keep community participation to people who want or need to participate. Search is provided via Verity and was quite fast in the demonstration. Search looks across discussion titles and body; community titles; and expert bio's.
Some of the additional lessons that Reed highlighted in the conversation
- Usability is critical. They did a system redesign in 2002 with significant usability input, and the system usage went through a major inflection point (growth), based on a graph of discussions over time.
- Assume that nothing is self-evident. In relation to usability and functionality, once the system was spread to a wider and wider community, the "little things" became much more troubling.
- Listen to the customer. Understand and explain "what's in it for me" for the customers.
- People won't write unless they feel comfortable knowing who will get their question. The Caterpillar system lets users easily ask "who will get this message," based on who is in the community.
- Measure and monitor usage, but don't tie usage to personal goals. Tie personal goals to business needs.
- Don't put this kind of project under IT, if possible. It is not a technology project, it is a business project that makes use of technology.
- Caterpillar looked for commercial software, but none met their specific needs around how the technology fits into the business processes.
This work is also described in the following articles, books and conferences.
- Knowledge System that Captures and Leverages and Organization's Expertise, yet2.com TechPak on the project.
- 3,000 CoPs at Caterpillar Worldwide, summary of the T+D article below with some graphics and additional opinion.
- Virtual Communities at Caterpillar Foster Knowledge Sharing, by Vicki Powers in T+D Magazine, v58, n6, p40-45 (purchase).
- Collaboration and beyond, KM Magazine, Jan 2004, v7, n5 (subscribers only).
- Motivation and Barriers to Participation in Virtual Knowledge-Sharing Communities of Practice (pdf), by Alexander Ardichvili, Vaughn Page and Tim Wentling of UIUC, presented at the OKLC 2002 Conference.
- Reed Stuedemann has a chapter in Knowledge Management Lessons Learned: What Works and What Doesn't (ASIS Monograph Series), Michael Koenig, et. al., 2004.
- This work has been presented at a number of conferences, based on the number of conference proceedings that turn up on a Google search (Braintrust, KMWorld, etc). It also appears in a 2004 report on CoP's from the Ark Group, Communities of Practice: Lessons from Leading Collaborative Enterprises.