Basic Project Planning and Estimation 2/5/2007 Keith Rome ([email protected])[email protected]

Basic Project Planning and Basic Project Planning and EstimationEstimation2/5/2007Keith Rome ([email protected])http://www.mindfusioncorp.com/weblog/

Have you ever been on a Have you ever been on a project that was "over project that was "over

budget" or "late"?budget" or "late"?

Why?Why?Bad planningBad estimatesChanging requirementsScope creepAll of the above

How could you tell?How could you tell?Constantly missing milestonesRising headcountCutting planned featuresLow quality (lack of sufficient QA)Poor architecture (not enough

time allotted)

Typical “Project Typical “Project Planning”…Planning”…Boss: "How long will that project

take?"Dev: "Two weeks I think."(two weeks later)Boss: "How's it coming? You done

yet?"Dev: "Just two more weeks..."(3 months later, the project is

terminated)

How to remedyHow to remedyUse objective estimates, not

subjective swags Document the level of "certainty" Factor in the expected planning

defects (requirements flaws, design gaps, etc)

Apply some basic math Identify Critical Paths and

slippage limits (slack) Refine as you go

PERT to the Rescue!PERT to the Rescue!Program Evaluation and Review Technique

First created in the 50’s to support R&D projects related to submarine missile programs

Objective estimatingObjective estimatingCan base on functional

componentsUsually correlated with a Work

Breakdown StructureTry to target blocks of work

roughly 1 day in durationUpdate requirements as

discrepencies are found (estimating is a good way to "proof" your requirements)

How "good" do you feel How "good" do you feel about the estimates?about the estimates?Rate each work item for how

accurate you think the estimate is

Give a reasonable Low and High estimate as well as the Likely

Good idea to also rate each item for subjective risk to project

Factor in global modifiersFactor in global modifiersPercentage of requirements

completionPercentage of requirements

defective / flawedPercentage of anticipated design gaps

(the things we know we don't know)Variable estimate accuracy

percentage (start at 100%)These will apply as global multipliers

to all estimates

Projecting the project Projecting the project timelinetimelineExpected = (Minimum + (4 *

Likely) + Maximum) / 6Plot out dependenciesAdd zero-duration work items for

start and finish eventsDetermine Early and Late Start

and Finish timesProject duration is equal to the

Finish time of the last task

Start and Finish timesStart and Finish timesEarly Start = maximum Early

Finish of all predecessorsEarly Finish = Early Start +

Expected DurationLate Finish = minimum Late Start

of all successorsLate Start = Late Finish -

Expected DurationSlack = Late Start - Early Start

Critical Path(s)Critical Path(s)Work Items on the critical path

always have Slack = 0Can be multiple critical pathsCritical paths can change when

work items slip or compactIn most situations, the critical

paths warrant the closest attention by management

Slippage limitsSlippage limitsSlack determines how far a given

task can be delayedDelay beyond the Slack estimate

will likely impact the critical pathSlack can be used to find

resourcing opportunities◦Move resources from high-slack

tasks to low-slack (critical path)◦Can help condense the overall

project delivery date

Anyone know what this is?

Projected path of your project…

Measure and RefineMeasure and RefineTrack Actual times for each task

as they are completedUpdate the Variable estimate

accuracy percentage factorWatch out for shifts in critical

path

This process is methodology-This process is methodology-neutralneutralIncorporate into existing

management methodsEspecially suited for Iterative

WaterfallAlso can be useful with modified

RUP, MSF and Agile processes

Questions / Comments?Questions / Comments?