Monday, February 27, 2017

Finding the gems in your team

I'm away at a conference this week, so I wanted to share this brief link to a podcast interview from last year, at the CIO Talk Network.

The topic of the interview was "Discovering and Polishing the Real Gems In Your Team" and featured two leaders sharing advice about how to find and develop the next stars within your organization.

"A leader’s team may consist of different types of workers. For some of these workers, it’s just a means to an end, but there are others who believe that the work they do, matters. These workers demonstrate complete integrity and dedication towards their jobs and have a deep understanding of their organization’s goals. So how should a leader go about discovering and developing them?"

The interview features Helen Norris, Chief Information Officer, Chapman University and Tom Kolditz, Founding Executive Director, The Doerr Institute for New Leaders at Rice University.

I'm friend with Helen, and we peer-coach each other, so it was great to hear her advice in this interview.

Monday, February 20, 2017

The open leader recently posted a great about "open leaders" and "Why we need open leaders more than ever." In brief, the article discusses how leadership changed in the Twentieth Century with globalism and teams that were geographically displaced. And this new mode of working led to a new kind of leader: the "open leader."

The article makes clear that leadership can happen at all levels of the organization. And I agree! To be most effective, you need to develop your emerging leaders so people can "lead from the balcony" or wherever they are, to provide vision and drive strategic direction. From the article:
As organizations continue becoming more open, even individuals without "leadership" titles feel empowered to drive change. These organizations remove the chains of hierarchy and untether workers to do their jobs in the ways they best see fit. History has exposed 20th century leaders' tendencies to strangle agility through unilateral decision-making and unidirectional information flows. But the new century's leader best defines an organization by the number of individuals it empowers to get something done. There's power in numbers—and, frankly, one leader cannot be in all places at all times, making all the decisions.
The article provides a list of five attributes for the new "open leader":

1. Control
"Where the leaders of old are focused on command-and-control positional power, an open leader cedes organizational control to others via new forms of organizational governance, new technologies, and other means of reducing friction, thereby enabling collective action in a more efficient manner."
2. Communication
"The open leader seeks to engage an organization by sharing information and context (as well as authority) with members of a team."
3. Trust
"Open leaders embrace uncertainty and trust their followers to do the right thing at the right time."
4. Autonomy
"Where the powerful command-and-control 20th century leader is focused on some position of power, an open leader is more interested in the actual role an individual plays within the organization."
5. Empowerment
"Open leaders are focused on granting authority to members of an organization."
How do you compare to the new "open leader"?

Monday, February 13, 2017

How to write an email

Some people claim there's an art to writing an effective email. In my professional career, I follow these general rules for writing emails:
  1. Write email
  2. Delete most of it
  3. Click Send
I find brevity is often best. If I write too much, I may lose my audience. But there's a tipping point: you need to provide a certain amount of context. If you don't "set the stage" or don't provide enough background, your recipient may not understand your message. How little is "too little" in an email?

Let me draw on a few real-life examples. Outside of my regular job, I sometimes teach an online computer science class ("CSCI 4609") at a university. As an online class, we have a flexible class size, although we try to set the "reserve" limit at ten seats. After the first ten students sign up, you need a permission number. But I'll usually give a permission number to whoever asks for one. Usually.

The exception is for students who don't meet the pre-requisite. We require that students take another computer science course ("CSCI 2101") before taking mine. This provides a grounding for the material in "CSCI 4609." Such pre-requisites are not unusual at the university level. If you don't have the pre-requisite, you need to get a permission number from the instructor (me) so you can register for the class.

Let me use a few examples from this class to show how providing a bit of context can make your request more clear:

Too little information

You can imagine my confusion when I received an email like this one from a student who wanted to take my class: (not the actual email, but representative)
Hi. Please give me a permission number so I can register for your class.
Why did this student need a permission number? There were already ten students in the class, so the system might have prompted the student that the class was "at reserve" at the student needed permission to add the class. Or the student might not meet the pre-requisites. I need to know more before I can give a permission number.

In the first example, it's no problem to add the student; I don't mind more students in the class. But if the student didn't meet the pre-requisites, then I need to check that the student won't be in over their head. I want them to succeed in the class, but if they take the class before they understand certain basic concepts, they will find the class very challenging.

I responded to this student to learn more about their situation. Each email was equally brief, requiring us to exchange several more emails before I determined if the student should be added to the class.

Too much information

While it's challenging to respond to too-brief messages, it can be equally difficult to make sense of the small novels that some people send me. Here's another example from a student who wanted to take my class: (not the actual email, but representative)
Hello Professor Hall. I was looking through the catalog to find another class I wanted to take this semester and I saw your class, CSCI 4609. This is a really interesting topic and I'd like to take it. I talked to a few of my friends about it and they think this is something that will really help me after graduation, so now I really want to take your class. I was in the student center last night going through my course registration for Spring semester and I tried to add your class. The registration system is a little difficult to use, so I'm not sure what the message is trying to tell me. When I add your class I get a little red box that pops up and tells me I needed to contact you and get a permission number. But I'm a senior and this will be my last semester before I graduate. I have taken all my other CSCI classes already, and this semester I only have my capstone plus some classes for my minor. I have taken the interface class, last year with Dr. Smith, so I meet the requirements. When you have a moment, I need to get a permission number so I can finish registration in the class. I know we're out of queued registration but your prompt attention would help me to get the class for next semester. I would really appreciate it. Thanks so much!
Okay, wow. That's a lot of information. And the actual email I received was much longer than the sample I wrote for you.

This email is unclear and unintentionally buries the request in an avalanche of extraneous information. I don't need to know where you were when you tried to register for the class. I don't need most of the information here.

Reading long emails like this one seem like more work than reading other emails, because I need to parse the email to figure out what's really going on. Please keep them short for the sake of the person who will read it. Breaking up the email into paragraphs will also make the information more clear.

Just the right information

Let me edit the above email into a message that has just the right amount of information:
Hello Professor Hall,

I would like to take your class, CSCI 4609. I am a senior and have the pre-requisites, but the system says I need a permission number to take the class.

Can you send me a permission number?
That's much easier to read! Notice that the email provides a brief context to "set the stage" before making a simple request? The use of short paragraphs helps the reader to parse the email for the important points: The student has met the requirements, but needs a permission number from me. That's a clear request, and one that doesn't require any follow-up from me.

Monday, February 6, 2017

Above/Below The Line

Across my career, I have helped in several IT consolidations. We are also doing that now in my current position.

IT consolidations can be tricky to communicate. I find the biggest hurdle is explaining how an IT consolidation can help both parties. In most IT consolidation situations, you start with two groups that are managing up and down the "stack" of technology. A key goal is to leverage size and scale so each group takes on the role that is best suited to them, so you simplify the work and let people focus on their value to the organization.

For example, you might have a development team that creates web-based applications for the overall organization. To run their web-based applications, they need a web server, so they install and manage a traditional "LAMP" stack: their own Linux server with Apache, MySQL, and PHP. Except that in most instances, "manage" is an overstatement. The group needs to focus on maintaining the web-based applications they write, and the system administration of the Linux box, MySQL server and Apache instances tend to fall behind. Meanwhile, in another part of the organization, you have an IT infrastructure group that's dedicated to running and maintaining Linux servers, and MySQL databases, and Apache instances.

In examples such as these, IT consolidation makes a lot of sense. If you could peel off the "infrastructure" work from the development team, and move those to the IT infrastructure group, you gain efficiency. The development team can focus on writing new applications, and the infrastructure team can take on the system administration.

Unfortunately, not everyone sees IT consolidation in such plain terms. When you talk about moving part of someone's job into another group, emotions get involved and consolidation becomes more of a challenge than it is meant to be.

To help get around these issues, and communicate more clearly about the benefits of IT consolidation, I've developed a model that helps both sides understand how IT consolidation will work out to everyone's benefit. I call it "Above/Below The Line."

Like most of my examples, I prefer to communicate it visually. Applications come in different forms, so let's envision different examples of an application stack:

Local development

You might have the above example, where an in-house development team writes an application. This usually runs in some kind of framework, such as a web server or other container. Underneath that, you need an operating system, which runs on a physical or virtual server, with storage and network.

Business application

Many applications are instead delivered by a vendor that you just need to run somewhere. A classic example would be a client-server application, but you might have a vendor-delivered application running in a framework, similar to local development. Again, you also need an operating system, which runs on a server, with network and storage behind it.


Whether the application is locally-developed or vendor-delivered, many will require a database system to store structured data. In terms of the "stack," the data schema is served by a database engine, which runs on an operating system on a server. Network and storage support the overall system.
You should see some similarities here. Each system shares several components or "layers" in an overall "stack." You can draw these components in a simple diagram, such as this:
Local developmentBusiness rulesData schema
Container (web, …)Delivered applicationDatabase software
Operating systemOperating systemOperating system
As I mentioned, before IT consolidation, the development team might manage the entire "stack" in order to support their application. But post-consolidation, you need to separate what is "application" from what is "infrastructure" by defining roles. Having described the "stack," the next step is to define the separation of roles. Let me use color here: green to represent the application team and blue for the infrastructure team:
Local developmentBusiness rulesData schema
Container (web, …)Delivered applicationDatabase software
Operating systemOperating systemOperating system
I've added an obvious line to indicate where most organizations choose to separate roles. After IT consolidation, the application team operates "above" the line, managing the local application development, configuring containers, managing business rules for vendor-delivered applications, managing data and database software. In this model, the infrastructure team operates "below" the line, supporting the operating system, servers, storage, and network.

This is a great model, but it's also a simple model. In any realistic environment, I find you need to make certain exceptions. You'll always need some cooperation between the two teams. The purple line becomes more difficult to draw, because it's position will differ based on your environment. An obvious example is installing software; in most organizations, the infrastructure team may need to be involved to install applications. Although usually the application team can take things after that:
Local developmentBusiness rulesData schema
Container (web, …)Delivered applicationDatabase software
Operating systemOperating systemOperating system
And the opposite is sometimes true: the application team may need to do some things at the operating system level. For example, in a Linux system, the application team often is allowed to configure the web server software, and stop and start it using operating system rules (such as sudo). The line becomes less clear, but it is still visible:
Local developmentBusiness rulesData schema
Container (web, …)Delivered applicationDatabase software
Operating systemOperating systemOperating system
I created the "Above/Below The Line" model before Cloud was a thing. But Cloud applications are easy to fit into the diagram. Cloud applications are hosted externally, similar to vendor-hosted applications. So Cloud applications rely on the organization's network, but otherwise live completely "above" the line. An integration team (which might be the same as the application team) manages business rules in the Cloud application.
Local developmentBusiness rulesData schemaBusiness rules
Container (web, …)Delivered applicationDatabase softwareCloud application
Operating systemOperating systemOperating system
When I've shaped an IT consolidation discussion using this framework, I find we experience fewer miscommunications and less stress. The "Above/Below The Line" model helps both sides understand the roles involved after IT consolidation. The key takeaway is the infrastructure team does the "heavy lifting" of managing operating systems, servers, storage and networks so the application team can focus on what they do best: building and supporting applications that drive the business.