Feedback Loops

One of the main challenges to the Theory of Evolution is Spontaneous Generation. Not the largely discredited folk theory that life pops up out of nothing, but the nearly undefinable moment when un-living matter becomes Life. It is that instant when some process becomes self-sustaining, self replicating and separate from its environment. Right now for chemical based life, the best guess hypothesis involves the idea of auto-catalytic loops. Auto-catalytic loops are the bio-chemical equivalents of placing a microphone too close to a speaker. If they don't have some sort of regulation, the process soon spins out of control consuming all the available resources, and dies from starvation. So at some point a collection of interconnected loops might have been separated out of the primordial goo, and could have started self replicating and had enough regulators to respond to external conditions. In essence you want feedback loops that create enough, not too much, or too little of something.

Only recently bio-chemists have been understanding the importance of regulators. Some cancers can be thought of as poorly regulated loops. People have long observed patches of DNA that were considered "junk" that produced small RNA strings, or microRNA. It turns out microRNA can act as a regulator for some processes. And some cancers have a genetic origin in these bits of microRNA, either because the microRNA is not restricting the process enough or the microRNA is over restricting the process.

In some ways this is analogous to the ways we communicate in organizations. If leaders under regulate communications, you may compromise your clients' or employees' privacy, or your organizations security. If leaders over regulate communications, leaders may not get important information and may unintentionally foster a culture of blame. I have observed the following unintentional bad communication policies:

Don't tell me you have a problem unless you have a solution. I remember a quality initiative that wanted to include employees. There were two important caveats. First, employees need to solve the problems and second, any problems that involved the provided software tools were off limits. I'm sure the corporate leaders thought they were sending the following message - we care about your work and want to empower you to make it better and easier. The message many employees received was - we won't invest time or money in you, fix your own problems. The end result was employees lost interest in the quality initiative, and stopped telling their managers about problems.

The Performance Improvement Plan is a tool managers use to salvage an employee who is not performing adequately. It is a way to clearly communicate expectations and required actions. It is a way to reform a "bad" employee into a "good" employee. Typically, when an employee is under a performance improvement plan, they are not permitted to apply for any other jobs in the company. The manager is hoping to communicate - we care about your contribution to our organization and would like to help you continue to make a contribution in your current role, and I'm going to be committed to working with you until that happens. The message the employee may receive is - it's my way or the highway. Generally, employees either leave the organization, or stop communicating honestly with their managers and become compliant yes men.

I've also observed some really good practices that regulated communications.

Periodically attend team meetings. Several years ago, I had a vice-president who would attend product review meetings. He didn't attend every one - there were over 20 departments with 3 teams each, but he attended often enough that everyone in the lab knew who he was. This sent two strong messages - I care about what you are doing and regardless of your positions you can talk to me.

Eating in the common Cafeteria can be another effective way to regulate communications. It makes a leader visible. The same vice-president ate at least two days a week in the lab cafeteria. At least two days a week, the vice-president was sending the message, I'm just like you and you can talk to me.

Get rid of the executive bathroom. My Father has a friend who is a plumber. He's a very down to earth guy who speaks in a very direct manner. This is the advice he gave to mutual acquaintance who was starting a machine shop with two partners. If you want to know what your people are doing, don't have an executive bathroom. The message sent when you don't have an executive bathroom, is I'm no better than you, you can talk to me.

What good practices have you seen? Leave a comment.

More information about microRNA:

High Hopes for a New Kind of Gene, Sylvia Pagán Westphal, Smithsonian Magazine, July 2009

Search or Derive

I was helping my son study for his Standards of Learning tests (known affectionately as SOLs - who thought that was a good idea?) He was a bit concerned because the study was around using Celsius instead of Fahrenheit. So, we went through different ways you could navigate the problems. Here are three of them:

• You could go with intuition - you could think in Celsius. Having lived in Europe - this is fairly easy for me - single digits is cool; 20° is comfortable; 30° is hot; 45° drink plenty of water and take steps to avoid heat prostration. If the temperature is below zero wear a coat.
• You could use derive a heuristic to get the temperature in Fahrenheit - double the temperature in Celsius and add 30° will get you close enough for temperatures in weather reports. You can derive the exact conversion formula if you know the freezing point and the boiling point of water in each scale. (The formula for the conversion is  F° = 9/5C° + 32°)
• You could look up an answer on the Internet.

We did not use the third method. So, thinking about Knowledge Centered Support:

• On every contact there is a decision point - do I search for an answer or do I derive an answer? 

When your body of knowledge is stable, your inquiries predictable, and your search tools are effective a decision is easy - search and only start deriving when you can't find the answer. When all three of these are present the decision becomes different - deriving may actually be the better approach. I know this first hand from supporting Macintosh products. 

At that time, Macintosh computers were losing market share. The Macintosh teams had a couple of teams of ten to twenty agents. There were dozens of windows teams. The knowledge base was predominantly windows solutions. (Apple snobs would point to the superiority of the Macintosh Operating System.) I didn't search much for solutions - I derived them. I had a mental check list of things I could do to isolate and diagnose a problem, and more than likely find a solution. In rare instances - mostly printing issues, I knew I could find a solution in the knowledge base and I would search. For the most part, I knew in 95% of my calls, I could derive a solution 100% of the time. At best, I could find a solution 30% of the time - the content was limited, the search engine was tricky, and I couldn't contribute to the knowledge base, and  there was no way to flag content gaps.¹ I really had only two choices: derive the answer or find a subject matter expert to ask.

Effective agents derived answers. The others struggled. Ironically, there was a woman who was struggling that I admired. She had derived her own rudimentary form of KCS - each time she encountered a call where she didn't know the answer, she would ask for help and carefully write the answer on an index card that she kept in a rolodex. She drove the subject matter experts nuts, because she was always asking for help - but she never asked the same question twice. If the answer was in her rolodex she used it.

If you are in an environment where agents or engineers need to derive answers because of the body of knowledge what should you do to encourage knowledge sharing? My best guesses at this time are:

1. Acknowledge the value of deriving answers - your new solutions are going to be derived. Reinforce the rewards for contributing creatively derived responses.
2. Remove obstacles to searching effectively, make sure your agents and engineers can easily give feedback, and communicate back the changes that are being made.

Is there anything I've missed?

¹ To be fair there was one feedback method - an annual survey sent out by the database administrator. I was very open with my critiques of the knowledge base. When I met the database administrator he recognized my name. Fortunately he didn't take the critique personally.

Manage Knowledge Socially

My colleague, Allan, pointed me to this blog by Anthony J. Bradley and Mark P. McDonald, Social Media versus Knowledge Management. It's generating some discussion in the office. The point of the authors is:

• Knowledge management is what company management tells me I need to know, based on what they think is important.
• Social media is how my peers show me what they think is important, based on their experience and in a way that I can judge for myself.

This is useful model, and it may apply very well to most organizations. The frustrated librarian/researcher/information explorer in me rebels because

• Social media doesn't guarantee information will be shared outside of a social network. Information sharing is a largely social activity. Most of what you know came from someone else. There is no benefit for a group of peers keeping information in a closely held tribe. We have a name for this in the corporate world - it's called a silo.
• Social media is often about the conversation instead of the information. I've been monitoring/moderating an online community. There is a constant call from the corporate knowledge management community to pull knowledge for reuse from the community. The problem is the majority of the value in the community is from the conversation not from an answer the community comes up with. The task of finding "reusable content" in a community like trying to write Wikipedia articles from the debates on what should be in each article.
• Social media only exposes the existing knowledge social network - i.e. who do people ask.
• Social media doesn't protect you from hierarchy, it just creates a different one base on reputation. Each user still has a responsibility to use due diligence assessing information. It is not enough that the author is a Vice President, or has a high on-line reputation.

I believe there is a role for making the shared information available to a wider knowledge. Knowledge should be managed socially - i.e. organizations should be engaged in encourage transparent information sharing, hierarchies be damned.  Systematic sharing is not necessarily anathema to peer review. Academic publishing has been very successful for decades with peer review. Similarly organizations can establish communities of practice and allow users to create and edit documents. Knowledge Centered Support (KCS) is one way organizations can manage information socially. One of the base principles of KCS is "Reward learning, collaboration, sharing, and improving" What is more social than that?