Friday, April 10, 2009

First impression of Harvard Business Review

When I was in the undergrad and grad programs in the business school, the profs kept having us read articles from Harvard Business Review. I remember b&w copied they distributed, and back then, it was the most boring and tedious reading I'd ever done.

Now wind forward 10 years, to last week. After paying $80 of my hard-earned money for a 12-month subscription, I finally received the first issue.

Why did I subscribe? Well, I've been reading a lot of business books lately, mostly on Sales and Marketing, and I've found a lot of value in them. Working with a lot of sales and marketing clients, I got a better understanding of their world through my reading. The human mind works by association, and now I go beyond listening to my client talk about their challenges and opportunities and nodding my head - I relate their experiences to what I learn in the books and can ask intelligent questions about their strategy and business. Subscribing to the Harvard Business Review was a logical next step in the learning process.

The copy of the magazine I received had little resemblence to the b&w smudgy copies from the college days. Colors and high-quality glossy paper were a nice surprise (okay, my expectations were really low). The paper is thicker than the paper used by the Economist; I am pretty sure that the publishers did this on purpose to send a subliminal message of the lasting value of the magazine and its articles. It also seems that the style of the articles got a little lighter - but maybe it is just me being smarter now (my girlfriend would disagree at times!).

I am now going through the issue so the next post will have my thoughts on specific articles. Come back on April 17!

Thursday, April 2, 2009

Top 10 Success Factors for a Bug Bash

Now that the bug bash I announced in the previous post is over, I wanted to share some experiences about organizing, conducting, and concluding a successful bug bash.  But first, let me tell you why a bug bash is so helpful for a software development project.  Theory aside, here's how my team has benefitted from our bug bash:
  1. Effectiveness.  We discovered over 50 issues in 1.5 hours.  Compared to regular testing, it would take a few days to get to this number.
  2. Usability Verification.  We received reassurance that the application is intuitive enough that people can use it without reading a user manual.
  3. Stability Verification. We received reassurance that the application is stable; out of 50 issues, only 1 was a critical error, and people were able to complete all tasks we asked them to complete.
  4. Install Verification.  We verified that the install process ran successfully on different configurations.
With these results in mind, let me now segway to the list of Best Practices we've leveraged to ensure the success of the bug bash.  

Before the Bug Bash

Best Practice 1 - Send an Advanced Invitation

People have busy schedules so make sure you give them enough notice.  In my case, I sent an invite 3 business days in advance.

Best Practice 2 - Set Specific Start and End Times

When scheduling a bug bash, we clearly identified the duration of the bug bash to remove any ambiguity around the time commitment we were asking for.

Best Practice 3 - Invite More People Than You Need

I wanted to have 8 people for the bug bash and ended up sending the invitation to 30 or 40 to get my 8.  Also, keep in mind that even you have enough people signed up, their schedules may change, and you'll be left out in the cold.  So try to book more people than you need.

Best Practice 4 - Invite People from Different Business Groups

Include multiple business groups into the Bug Bash - anybody who is willing to participate should be welcomed (unless your application is highly specialized and requires a deep understanding of a particular domain).  Different groups will bring fresh perspectives to the table, and you'll get bugs from behaviors developers or testers wouldn't think of.

Best Practice 5 - Provide Written Instructions With An Appropriate Level of Detail

Our Bug Bash was for a sales tool.  Sales people don't like reading manuals, and the tool needed to be intuitive enough for a first time user.  To keep the situation as real as possible, we provided our bug bashers with a description of a business situation in which the tool would be used and listed 8 high-level tasks our bug bashers had to accomplish.  We also included a list of known issues in this version of the tool so people wouldn't waste time logging them.

Best Practice 6 - Include A Feedback Form or Spreadsheet

To ensure that the bug bashers would submit feedback in a consistent manner and format, we provided them with a simple Excel spreadsheet with pre-defined columns.  The standard entry format made it extremely easy for our testers to consolidate the results.

During the Bug Bash

Best Practice 7 - Send Out a Reminder To Start and To Finish

Our bug bash ran from 1 pm to 2:30 pm.  At 12:58, I sent out an announcement to everybody who accepted the invitation and reminded them that the bug bash was starting.  Shortly after 2:30 pm, I sent out another announcement letting people know that the bug bash was completed.  Why should you do this?  Your bug bash is a high priority for you.  For your bug bashers, this is an add-on activity that they graciously agreed to do. Sending out the announcements helps them remember their commitment and releases them back to their regular work lives.  

Best Practice 8 - Be Available To Answer Questions

Our bug bash instructions included the team's Distribution List and a request to email us should any questions or issues arise.  We monitored the alias throughout the bug bash and were able to answer a few questions that came in. This ensured that people wouldn't get blocked from completing the tasks we asked them to complete.

After the Bug Bash

Best Practice 9 - Thank Your Bug Bashers

After the bug bash was over, we sent out a Thank You note to all the participants and  listed their names.  That's the least any team can do to express their appreciation of the effort put in by the bug bashers.
 
Best Practice 10 - Triage the Issues and Log Bugs

After our Tester consolidated the feedback of the participants into a single list, the team sat down and went through every single issue together.  We ended up dismissing some issues, raised a few questions to the client and a user experience architect as a result of others, and logged the rest as bugs into our bug tracking program.

Armed with this checklist of best practices, you too should be able to complete a successful bug bash and achieve the following goals:
  1. Effectiveness (high ratio of bugs to time spent)
  2. Usability Verification
  3. Stability Verification
  4. Install Verification

Friday, March 27, 2009

Bug Bash coming up

One of my projects is going live in mid April to a world-wide audience of thousands of users, and we decided to conduct a bug bash.

According to Wikipedia, "bug bash is where all the developers, testers, program managers, usability researchers, designers, documentation folks, and even sometimes marketing people, put aside their regular day-to-day duties and pound on the product to get as many eyes on the product as possible."

We are planning to have 8-10 people participate in the bash, and I am sure I'll have some lessons learned to share. Please come back on April 3 for an update.

Friday, March 20, 2009

Earned Value as a Survival Technique for T&M and Maintenance Projects

The Earned Value calculations rely on the assumption that you have created a Work Breakdown Structure and estimated the work for each work package. Works great when you know what you are going to be doing, but what about T&M or maintenance projects with requirements and tasks to be defined as the project progresses? Is Earned Value dead for those? Not so fast. Keep on reading, and you will learn how you can proactively identify budget issues when the WBS doesn't exist.

If you've watched one of my favorite TV shows, Survivorman, you know how it starts for Les Stroud. He looks at scarce things he has available and talks about what he can use them for in a survival situation. In our less dramatic situation, we normally have the following things available to us:
  1. Budget. Even though the project may be T&M, more often than not you have the target in front of you.

  2. Timeline. Even though you don't know what exactly you'll be doing, you should at least know how long the project will last.

This gives us plenty to work with. Below I'll show you two examples of how I survive with only these 2 numbers handy.

Example I - The client told you how much money they have to spend

For the sake of this example, let's say that your budget is $40,000, and the duration of the contract is 20 weeks. You will be providing status updates every other week so your reporting period is 2 weeks. This is a site maintenance contract so you have no clue as to what might come your way.

Step 1 - Calculate the average projected burn per reporting period.

Average Project Burn = Total Budget / # of reporting periods

In our example, Average Project Burn = $4,000

Step 2 - Put these numbers into a spreadsheet.



Step 3 - Build a chart based on this data.



If you need assistance with building this chart, you can refer to one of my earlier posts.

Step 4 - Start tracking.

At the end of every reporting period, I update the spreadsheet with the actual cost incurred during that period. After a few updates, your chart might look as follows:



This is important to you, because the chart provides you the visibility into a potential budget overrun way before the overrun becomes a problem. In this example, you see that for the last 3 periods, the project's actual burn consistenly exceeded the proejcted spend, and if all things stay equal, the project will end up costing more money than originally expected. If you include this chart into your status report, the client will become aware the issue now, with enough time for you and them to figure out an approach to either curtail the scope, improve the team's productivity, or increase the budget. Happy client = No suprises

Example II - Your estimates are resource- and time-based

In this example, you don't know the scope either, but you estimate the project spend based on the expected utilization, rates, and number of resources. So instead of dividing your total budget by the number of reporting periods, you estimate the spend using the following formula:

Spend for Period 1 = # of hours for Resource 1 * Resource 1's rate + # of hours for Resource 2 * Resource 2's rate + ....

To create a chart for this example, you can follow the technique I described in the earlier post and just replace the WBS work packages with reporting periods:

Wednesday, March 18, 2009

Stay Late, Earn Respect, Build the Team

I recently joined a PMI Certified PMPs group on LinkedIn and couldn't help jumping on a discussion regarding team building exercises. Are they useful or not?

My response was:

"Folks, the best team exercise is the one when you have a deadline, and you stick around in the office with the team even though they are doing all the work, and you are just sticking around to instill the team spirit. A good Project Manager is a leader who leads by example; you have the power to make an impact, and you don't need artificial exercises for it.

On a side note, I play Halo with my developers, I go to lunch with my testers, and I fight for my team in front of the clients. That's what helps me turn people I work with into cohesive teams."

I'd like to elaborate on this a little bit and give you a specific example of earning respect through this.

I recently delivered a project with an absolutely insane timeline - a public-facing web site with a potential audiences of hundreds of thousands. We did it on time and on budget, in 4 weeks. As you can guess, there were some late nights for the team, and that was expected (but not planned for!). As the project started, my Dev Lead asked me if I was the kind of a Project Manager who stays late with the troops until the job is done or goes home. My answer was, dude, of course I'll stick around (I learned the importance of staying up with the team at my previous job from Peter Hammond, the owner and President of CyberSavvy. Thanks Peter!) So a week before the launch, the push actually came to shove, and I stayed in the office until 9 pm with the team. The next day my Dev Lead came by my desk in the morning and said that he appreciated me sticking around and added that if I hadn't, "he and I would've had to have a little chat".

Earn respect and build the team by being there with them. There are few things more discouraging to the team morale than a PM who goes home when the team is burning midnight oil.

Monday, March 16, 2009

Tell-a-friend is one small step for a developer, a giant leap for a company

In early March Ascentium launched a public web site for one of our clients. To my surprise, one of the key requirements for the new site was a tell-a-friend feature aimed at promoting the company's product.

That seemed like a pretty insignificant piece of the functionality to have, and in my humble opinion, there were other features that should've been completed instead. Was I wrong!

Last week I was reading a book called Word of Mouth Marketing by Andy Sernovitz. On page 123, Andy lists three things that the reader should remember even if they choose to ignore the rest of the book:
  1. Ask people to spread the word.
  2. Put everything in an email
  3. Put a tell-a-friend link on every page of your website.

The idea behind the tell-a-friend link is that it is one of the best ways to spread the word. The book also clarifies that tell-a-friend should be as compelling and easy to use as possible, as it gets the company word-of-mouth marketing for free. It often surprises me how little things can prove to be so important, and this is definitely one of them. Lesson learned! After reading the book, I wrote the content for my girlfriend's web site last week (she just opened a family medicine practice in Kirkland, WA), and you'll be sure to find a tell-a-friend link on every single page of the site.

The morale of the story for my fellow Project Managers: understanding the business of your clients is a key success factor for the project. I would argue that it is as important if not more important than the understanding of the technology.

Friday, March 13, 2009

Thoughts on project overages in the context of a consulting company

I was chatting with my PM peers over a brown bag lunch this afternoon, and project overages became a topic of a heated discussion.

In the PMI/PMP world, "on budget" is one of the cornerstones in the definition of a successful project. PMs are expected to bring the project on budget, because:
  1. An overage eats into the company's bottom line.
  2. It is an easy metric to use; there is nothing fuzzy about it; you are either on budget or you are not.
  3. It is a "feel good" psychological statement (didn't you have it on the latest copy of your resume?)

I would like to make an argument that in a consulting company, small overages may be OK.

Here's why.

Most consulting companies have a bench. For folks outside of the consulting world, a bench is a term to describe available resources who are collecting a paycheck but don't have an active project to bill against at the moment. In the ideal world, a consultant on the bench stays productive by taking training, assisting with internal projects or sales and marketing efforts (if my manager happens to read this, blogging is a marketing effort too!), or actually taking care of themselves by enjoying a well-deserved vacation. In the real world, people come to work and surf the Internet, play video games, gel by the water cooler, show up late and leave early.

Since the resources on the bench get paid anyway, utilizing them on a project will have no impact on the profitability of the company, as long as their involvement doesn't interfere with the productive activities listed above or prevents them from joining another project. The potential benefits of using benched resources, on the other hand, are numerous:

  1. Rescued projects
  2. Happy clients and potential future revenue from referrals
  3. Improved morale (sitting on the bench is no fun! there are only so many online training courses one can take at a time)

That said, overages are a re-active, not a proactive measure. Similarly to how overtime is a safety net and should never be planned for (according to PMBOK), overages are a second line of defense and should never be part of the project plan.

So, my conclusion on overages in a consulting company is:

  1. When managed responsibly and staffed by benched resources, they cost the company $0
  2. They have a number of long-term benefits
  3. They are a safety net and should never be planned for

What do you think? I would love to get some comments on this post.