Monthly Archives: June 2008

Build vs. Buy

Every company needs specialty tools to make their company run – little things to help make things easier.  For example, a cabinet shop might need custom-built jigs.  An auto mechanic may build a few of his own tools for jobs he performs frequently.  High-tech companies certainly build plenty of in-house programs to manage information technology – databases, data entry, customer records, etc.

The question I’m posing is this: when is it right to build those custom tools in-house, or when should you go out and buy a suitable tool instead?  For instance, should an auto mechanic build his own 500 piece wrench set, or go down to Sears and buy them?  Should a tech company build a time tracking product from scratch, or license something like Standard Time®?

Since I’ve personally worked in consulting for a number of years (in another life), I’ve seen it all.  I’ve seen companies build products that were available off-the-shelf for 1/100th of the cost.  And, I’ve seen people shoehorn freeware code and products into commercial use.  Both were mistakes.

I tend to say that companies should only build products that are much more expensive to buy.  In other words, don’t try to write your own compiler – Microsoft does that just fine.  Another reason to build is when you have exotic business needs.  One last reason might be if you have time on your hands.  If you’re not making money serving customers, maybe you’ll have time to write and maintain a few necessary tools.  But that’s a dangerous position to be in.

In IT, we ignore such advice because we know we can build everything.  But I suppose the mechanic with a little metal-working and welding experience does the same thing.  🙂

What tools have you built, and regretted it?  Post a comment and let me know!

 

–ray

Fear or Fearless?

We’ve all heard the adage, “Afraid of change”.  It’s preached throughought the business world and even in our personal lives.  We are constantly told to step outside of our comfort zones, embrace change…don’t be afraid!  It is great advice that everyone has faced at some point. 

So why is it that as soon as the media starts telling us how terrible the economy is, we freeze up with fear?  Even in uncertain times we must press forward and get past our fear.  Yet it seems totally acceptable to use the economy as a reason to remain static.  This is bogus.  I agree that we must scrutinize our expenditures and evaluate new ideas, but we should do that in good and bad times.  What frustrates and kills businesses is doing nothing!

If you have an idea to save your company money and improve revenue, why let fear stand in the way?  During good and bad times we must always be prepared to step out and embrace change.  I can honestly say that I’ve had my wins and losses with fear.  However, it always comes down to a simple choice that we make…be controlled by fear or choose to be fearless.  I’d like to hear what you think? Are you driven by fear, or are you fearless?

–Warren

Our Customers Do Our QA

This is not as bad as it sounds.  A lot of companies are in this boat.  They don’t have the revenue to hire full time QA engineers and testers.  So, they do what they can and rely on customers to report bugs, oversights, and possible enhancements.

Let’s briefly discuss how a ‘real’ QA department operates, and then contrast that to the poor-man’s solution.  Normally, a QA manager and team of testers ensure that products delivered straight from the engineer’s drawing boards ship with minimal bugs.  Each tester is assigned one or more areas of the product.  They follow a QA plan and checklist.  Every feature is scrutinized, often put through the paces as real customers would use it.  Bugs are sent up the chain, through the QA manager, and back to the original engineers for fixing.  Once resolved, QA engineers verify the fixes.

A QA department is nice to have.  They find hundreds of issues, and save the company a lot of money and embarrassment.  Product defects are resolved before they hit the shelves.  In the end, the QA department is usually worth their pay.  After all, that’s why the department exists in the first place – to save the company money.

But if you can’t afford the salaries, and you have a small number of customers, the development and manufacturing engineers will need to perform the dreaded duty.  (It is monotonous work.)  Problem is, engineers bristle at repetitive tasks like product testing.  They won’t do a thorough job, and you still end up getting bug reports from customers.  Plus, you’ll have to pay the engineers for their extra testing work – and they aren’t cheap.  Bugs that reach the outside still cost you money in customer dissatisfaction and lost sales, but perhaps not as much as the salaries for a full testing crew.  That’s the real gamble.  And, there are so many intangibles that a true cost analysis of each method is difficult.  But, you’ll know when you need a QA team – when customer complaints begin to kill your product momentum.  When that happens, put together a team quick!

 

–newshirt

Bill Gates’ Exit Interview

Have you seen the Channel 9 interview with Bill Gates?  The link is below.  The interview is about a half-hour long, and focuses on Gate’s plans for the future as he transitions out of Microsoft and into his philanthropist foundation to help the poor.

http://www.microsoft-watch.com/content/corporate/a_month_of_gates_7.html

 

Here are some quotes I found interesting.  These are the kinds of lines that drive my own thinking.  They drive me to fight for our brand position in the marketplace.

You’ve got to have product plans for every product you expect to release in the next few years.  And those plans have to be very realistic.

 Making sure we take some risky bets.

We are a very self-critical culture.

Software seems to be so complicated…  Software has some composability today…  But we have to make it easier to write big software.  We’ll have a lot more progress in the next decade.

 It’s always surpring to me there’s little attention paid to what we are doing for business users.

We care about the information worker.

We’re always sharing about where were going so people can make plans.

 

Best of luck, Bill.  Care to comment on the Channel 9 interview or this humble post?  🙂

–newshirt

Define: Actual Work

Actual Work: Hours collected in a timesheet or with a timer indicating the total project hours an employee has actually worked.

 

Collecting actual work in a product like Standard Time® in important.  It allows you to compare your original forecasts with the actual hours worked by employees.  This can lead to some valuable metrics, including percent complete and estimation variance.

Assuming your original estimates are close, percent complete will tell you how far along the project is.  This is obviously more valuable to project managers and executives who are slightly distanced from the day to day tasks.  In some cases this can be their only indicator of the project status.

Refining your estimations can also be valuable, especially if your company performs the same projects over and over again.  Only actual work can give you exact numbers, and that can lead to more profits.

–ray

The Long Product Pipeline

Ever wonder why there are so few product-based companies, and so many service-based ones?  It’s all about cash flow.  The images below illustrate this, but they need some explanation.

Product-based companies
If you’ve got bright ideas of building a product and making lots of money, think again…  There’s a long pipeline from product development to big piles of sweaty cash.  First, you have to build the product.  Nobody’s paying you to do that.  It’s your nickel, pal.  And you can spend as much time as you like.  But then comes marketing and promotion.  Get it out there for the world to see – but still more time without pay.  And then, when they find your product, you’ve got to sell it and collect your reward.  It can take months or years.  And it can be exhausting.  See the stages below.

 

 

Service-based companies
Want cash quick?  That’s why there are many more service-based companies in the world.  In that model, you can find a client and begin invoicing them after the first thirty days.  The path to that stack of sweaty cash is much shorter.  But, the stacks are smaller.  And, you tend to live on them so they never accumulate.  Lose a client, and you’ll have to start over.  Even this model has it’s risks.

 

 

 

So which model suits you?  Got the time of invest and wait for the “big” payoff?  Or, do you prefer earning as you go along?  There are plenty of advantages to both, and only you can decide what’s right for you.  Hopefully, this short post will help in your decision-making process.

 

–newshirt

Battles in the Conference Room

Having varying points of view is healthy for company growth.  However, sometimes the varying opinions have underlying meanings that aren’t always what they seem.

For instance, how often do IT teams resist implementing new programs out of fear, laziness or just the extra work it creates?  I mean they are tasked with everything from support desk tickets, computer set-ups for new employees and even firewall security.  So it’s easy to get bogged down with the extra work a new program rollout will cost instead of focusing on the big picture and long term benefits.  Other times people resist out of simple personality conflicts.

When this happens you will hear legitimate-sounding excuses of how the installation may conflict with another program or of how it may take up too many resources.  This happens because one manager doesn’t want to help another manager or due to many other normal and very human reasons.

So the next time you have someone objecting to what most everyone else is pushing for, ask yourself, “Why?” It could be a very real objection worthy of consideration.  Or, as I found out last week, it could be a person impeding progress because they simply do not want to put more work on their plate.  The tough part is determining which is true. After all, we are all human!

 

–Warren

Define: Technical Debt

Technical debt: Engineering changes you owe a project before it can be signed off as completed.  E.g. software, electrical, mechanical, ergonomic, and other engineering issues to be resolved.   If you’re tracking bugs, enhancements, defects, and other project issues, then you no doubt have some technical debt.  In other words, your team owes the project something before it can be said to be complete.  Every one of those issues should be documented, fixed, and verified by an objective party before project completion. Engineers, and in fact everyone, tend to sweep little things under the rug.  Especially when they don’t know how to resolve them.  Can’t fix that annoying little bug?  Easy answer: don’t tell anyone.  That’s just human nature.  Hide your deficiencies.  Make yourself look good.  Ignore the hard things. A good project team culture faces up to difficult little deficiencies.  A simple line item in a bug tracking tool like Standard Issue® makes them public, and that resolves most of the “facing up to them” issues.  Now everyone knows they exist, and must eventually be dealt with.   –ray

90 Days is Never Enough

If I give you 90 days to complete a project, but don’t hand you a checklist, chances are the project won’t be completed.  A project without clear deliverables is a green light to surf the web all day.  It doesn’t have the teeth to get anything concrete done, and leaves you without direction.  Given such a project, most people will simply say, “I’ll do what I can,” and then do almost nothing.  I’ve seen it happen.

 

Measurable goals
Consider making your goals measurable in some way.  Find a way to boil them down to simple numbers.  Example: this new feature will let us process 1,500 new transactions per day.  Or, the new system will let 1,000 new customers to submit service requests.  Don’t just say, “make it better.”  Make it measurable instead.

Make it quick
Giving arbitrary timeframes like 90 or 180 days without milestones is foolish.  Some people will believe they have all the time in the world, and wait until the night before to start.  Remember cramming?  That’s the mode most many people work in.  Instead, tighten in the deliverable dates, and let the project go late, if necessary.  That tends to keep people on task.

Make it simple
Don’t try to boil the ocean.  It never works.  Simple, foundational projects work best.  If they are completed to your satisfaction, you can build upon them later.

Now you can roll your eyes (like I do) when you hear, “it’ll be finished in 90 days.”

–ray

How To: Add Progress Lines in MS Project

Progress lines in Microsoft Project help see where tasks are behind schedule.  They’re hideous to look at, but serve a useful purpose.  This post shows how to add progress lines to a Microsoft Project file.  Buckle up; this may get rough.  🙂

 

Start by adding a few tasks to a new project:

  1. Add Task 1, with 10 hours duration
  2. Add Task 2, with 20 hours
  3. Add Task 3, with 30 hours

The tasks and Gantt bar should look like this.  At this point, we have no progress lines, just simple task bars in the Gantt chart.

 

 

Add a Project Status date:

  1. Choose Project, Project Information
  2. Enter a ‘Status date’ for when you would like to check task status (the status date progress line will be red)
  3. Click OK

 

Add progress lines:

  1. Choose Tools, Tracking, Progress Lines
  2. Click ‘Always display current progress line’
  3. Click ‘At project status date’
  4. Click ‘Display selected progress lines’
  5. Click in the list and choose the dropdown arrow
  6. Select a date for a progress line (these lines will be black)
  7. Click OK

You should now have two progress lines on your Gantt chart, and things may have gotten a little ugly.  As you move the task bars, the progress lines will update.  Tasks before the progress line will cause the line to go leftward (that’s the ugly part).  What good are they?  Backward facing lines are those tasks you need to move forward.  They need to be rearranged to meet your current project plan.  The image below is an example.  Notice how the lines go backwards to tasks that are behind schedule.

 

 

–ray