Zopim Chat

Rodney Brim Report

The Rodney Brim Blog

An Unrecognized Solution to Estimating Project Time & Cost

May 12, 2015
manageprouser

Don’t you hate to be late on projects?  I do.  And it doesn’t help all that much to know that a significant percentage of projects aren’t completed within time or budget.  The Lousiana DODT even publishes their % of projects completed on time, which currently looks to be 53%.  Even though there’s a large volume of work written about how to manage projects better, and  an entire organization committed to this endeavor – PMIso as to solve this problem of getting stuff done on time, as well as an entire organization committed to this endeavor (PMI) it doesn’t seem to make the estimating problem go away.

Here’s a challenge:  if you want to get more accurate at estimating how long it’s going to take to complete a project, I’m suggesting you need to approach it with a significantly different mindset.  Let me explain.TwoModelsforTimeEstimates

First of all you’ll need to loosen up the grip on expecting that better project planning or more detailed hourly assessments to complete tasks or needed buffers will get you the accurate number.  As projects involve more and more innovation, it’s really difficult to predict total resource costs, even if you manage the project really tightly.   You see current approaches to estimating projects embodies a flawed assumption that effects most projects, especially when they involve greater levels of innovation for completion.

What’s the assumption?  In this short post I’d like to demonstrate that the challenge in completing projects on time is only partly about management practices, and just as importantly, it’s dependent on estimating the time to complete using the 80/20 rule.  The prevailing assumption is that all tasks can be equally assessed in terms of time and budget.  What I’ll try and show you below, is that in fact, there isn’t such a thing as average accuracy, which is how most projects are estimated with a buffer.  You can’t accurately estimate the time or budget to complete tasks without using the 80/20 rule… which most of us don’t invoke.

Let’s start with a premise:  As projects involve more innovation demands and less duplication of past efforts, the time to complete will be impacted by the 80/20 rule or Pareto’s curve, and without accounting for this phenomenon, project duration (and budgets) are likely to be significantly under-estimated.

Let me explain how this works given the 80/20 rule

1. Given the 80/20 rule, projects should have an unequal distribution of number of tasks that represent low value versus high value.  It would look roughly like this chart, with 80% of the tasks contributing necessary, but relatively low value, and a second group of tasks comprising the remaining 20% representing significantly higher value.  Some tasks will inevitably be much more significant in creating value than others.

80_20RelativeValue

Now let’s build on that with another point.

2. Given the 80/20 rule, projects should also have a similar distribution when it comes to difficulty of tasks.   That’s certainly true in the technology area where I do most of my consulting.  The same 80% of tasks that yield relatively low value, are also the lowest in difficulty to complete.  They have the least unexpected obstacles emerging, and are typically covered by current and (hopefully) documented processes.  On the other hand, when it comes to rating the difficulty of tasks, the most difficult 20% of the tasks, also tend to be the same top 20% that provide the most value going forward.  So let’s overlap those Pareto trend lines.

80_20DegreeofDifficulty

If we go one more step, it gets really interesting and begins to turn the world of project planning and estimation on its head as we know it.

3. Given the 80/20 rule, when it comes to estimating inaccurately the duration of project tasks, 80% of the tasks should account for only 20% of the estimate mistakes, and 20% should account for the remaining 80% of duration estimate inaccuracies or slipped due dates.  That means that the estimates for 20% of the tasks is going to be wildly off.  By the time it gets to your top 4% of tasks, they could be 16x off or more.  Your estimate of one week could be off by a factor of 16!  There goes your due date.

If in fact these three groupings of tasks overlap, which in my experience they do tightly, you quickly realize that the estimates on the most challenging, value producing tasks, are also the most inaccurately assessed when estimating time and cost to complete. 

Said another way, your current process for estimating the number of hours per task works for roughly 80% of the tasks.  Unfortunately the remaining 20% of your tasks in a project, which are going to be the most complex and consume the most time and cost, are going to account for 80% of your estimation errors… if the estimated hours to complete also don’t rise eponentially like a Pareto Curve.

80_20EstimateInaccuracy

Ok, so what’s a more accurate method of assessing time and costs to complete a project?  What if rolling up hours to complete from all the sub-tasks just won’t give you an accurate number, even if you do an excellent job of managing the project?  So then what?

Here’s the new formula for correctly estimating the duration or resource requirements for projects which better fits the Pareto curve, especially on projects which involve innovation.  

TR =  80E * TE + 20E * Pfactor * TE

  • Total Resources (Time) required =
  • 80E (ideal number of days estimated to complete the bottom 80% of tasks as ranked for difficulty) times
  • TE (your team effectiveness which can vary from .5 to 2.0; e.g. your fastest teams will beat your estimates by 50%, your slowest teams will exceed by 100%) plus
  • 20E (ideal number of days estimated to complete the top 20% of tasks as ranked for difficulty) times
  • Pfactor (a multiplyer based on the difficulty of the innovation inherent in the task; ranging from a simple multiplier to a square of the estimate when estimating in weeks), times
  • TE (your team effectiveness which can vary from .5 to 2.0; e.g. your fastest teams will beat your estimates by 50%, your slowest teams will exceed by 100%)

Does that look too complex?

Here’s a simpler formula.  Total Time = Time estimated for the most difficult 20% of tasks x 5x

You should be able to confirm if your modifier is really a “5x” or something different, if you look back on previous similar projects, and adjust accordingly.  But on highly innovative projects, just take the estimate for the 20% of the tasks that represent the highest value/complexity and difficulty and multiply by 5.  Your modified number may vary, but regardless, the time to complete the top 20% of your tasks will be the driver for time to complete, not the time estimates on the bottom 80%.

Bottom Line:
To accurately assess the time and dollar costs of projects you need to use the 80/20 rule, especially as projects increase in innovation and complexity.  That means you need to modify ideal time to complete by a significantly higher number for the top 20% of tasks in relative innovation and difficulty, then you will use for the bottom 80%… otherwise you’re going to miss your due date. 

 


1 Comment. Leave new

Spot on, as always. Thank you for the great thinking. Very useful and thought provoking. Should be used in all projects and in all industries. Great work.

Joanna

Reply

Leave a Reply

Your email address will not be published. Required fields are marked *