Archives for October 2013

Estimation Frustration!!!

The world of project estimates is a bit mystifying and to some it can be down-right frustrating! Depending on what the organizational culture is like, a conversation around task assignment and the estimated time it will take to complete the task are often very interesting conversations. There is even a mindset floating around out there that it is best to just give in and call it SWAG – a Silly Wild Ass Guess – but, how does anyone arrive at a good sound estimate if everyone  thinks they are guessing?

When management asks for an estimate, they are not looking for a silly wild ass guess. They want good estimates. However, sometimes there is confusion regarding the meaning of the word “estimate” and the word “budget.” Estimating is defined as a technique for forecasting the cost and effort associated with a task. Estimates are basically created to provide an informed assessment related to an event. Project Managers are used to doing estimates all the time for construction projects and software projects, as they need to know those tangibles associated to time, resources, and materials when it comes to their budget. Budgets are determined based on the cost estimates. Once a budget is set, it is best leave it be.

The very word estimate implies that there is a thoughtful process behind the numbers that are based on available facts. The dictionary definition states that estimates are used “to calculate the approximate amount or extent of something,” or “to evaluate.”

Estimates are interesting because the estimate arrived at depends on the depth of understanding and experience of the estimator. Likewise, good estimates also rely on managements “buy-in” and confidence regarding the approach. Once the approach is decided upon all the fear melts away and the basic groundwork is in place to achieve success.

However, there are many different ways to do an estimate. Without knowing the variety of ways to achieve a decent estimate can sometimes lead project teams down a very slippery slope. So, if the task is to get a two million dollar project delivered in time to satisfy federal regulators , and there is a task to provide development estimates to complete the project, what is the best way to make those estimates?

Even the language of estimates is slippery. When there are terms like “top-down”, “bottom-up”, “analogous”, “parametric”, “three-point”, “what-if”, “apportionment”, and “Delphi” one has to be able to navigate the ambiguity of it all. Figuring out which approach your organization is most comfortable with is only half the equation. The rest is all about those interesting conversations.

According to the CDC (Centers for Disease Control and Prevention) Unified Process Practices Guide regarding project estimates there are a variety of techniques that can be used.

  • Top- Down  – very high-level, provides more of a “ball-park” estimate
  • Bottom-up  – very detailed pieces of work which are then aggregated into the total
  • Analogous  – expert judgment, based on historical information from similar projects
  • Parametric  – manage independent variables such as design specifications to describe the scope of the work
  • Three-Point  – uses a weighted average of three types of estimates to create  a formula to gain a weighted average, also known as the PERT (Program Evaluation and Review Technique) chart
  • What-if Analysis –  uses scenarios to evaluate the effects of changing various factors depending on project schedule, resources, scope, and quality
  • Dephi method – principle is centered around the “collective intelligence” to make decisions and determine forecasting aspects also known as Estimate-Talk- Estimate (ETE) or mini-Delphi used for a structured forecasting approach

Then there are some additional methods listed as best practices in the BA Body of Knowledge Guide.

  • Rolling Wave – uses a refinement of estimation tasks as one iteration leads into another
  • Historic Analysis –uses history as a basis for estimating
  • Expert Judgment – uses the expertise of those who perform the work

The aforementioned SWAG  is not listed, as the SWAG approach only tends to defeat the purpose. A guess is a guess… it is not an estimate, as a guess is not based on any sort of informed assessment. A true estimate takes in to consideration the experts understanding of the overall effort. To get to a decent estimate it helps to lay the groundwork, know the culture, understand the acceptable techniques, and define the overall approach. All of those aspects will bring meaning to any estimate.

As stated earlier, when trying to figure out an approach toward making estimates it is a good idea to consult management. To add to that, some organizations like to talk about “percent of confidence” or a “contingency factor.” Sometimes the discussion regarding an approach can happen as part of a project team and sometimes the estimates are elicited in a silo, it totally depends on the culture of the organization.

Plus, many large companies have something called a “capability model,” these organizations are pretty mature and they know how to arrive at decent estimates. But if a capability model is not part of the organization’s vocabulary it is best to follow the direction provided by management and the Project Manager.

Where “Time” is all about the various resources needed and how long it takes to get something done, “Materials” are all about the tools, equipment, and supplies needed to deliver on the customer’s expectations. There is the thoughtful approach to each task, as the granite countertops would not be applied to the counters prior to installing the counter cabinets.  Just like the roof of a house would not be added without the main walls and framework being built first.  The idea is that by doing estimations there are expectations being set, dependencies foreseen, and risks identified regarding the timeline for completion!

Good estimates are attainable, but only when there is a level of comfort and confidence in the estimate itself. Recognizing a good thoughtful approach will lead all good analysts to gain traction on that slippery slope called “estimation frustration.”