COCOMO Intermediate Model: Master 15 Cost Drivers with Downloadable Checklist

 

---

In the world of software project estimation, accuracy can make or break your timeline and budget. Enter the COCOMO Intermediate Model—a powerhouse tool that refines basic estimates by factoring in 15 key cost drivers. If you're diving into the software engineering cocomo model, this model stands out for its balance of simplicity and precision, helping teams predict effort, schedule, and costs with real-world nuance.

Developed by Barry Boehm in the late 1970s and refined in his 1981 book Software Engineering Economics, COCOMO (Constructive Cost Model) evolved from a basic formula into something more robust. The Intermediate version builds on the Basic COCOMO by multiplying the initial effort estimate by these 15 cost drivers, each rated on a scale from Very Low to Extra High. This post breaks it down step-by-step, explains how to apply it, and shares a free downloadable checklist to master them all. Let's get started.

What Makes the Intermediate COCOMO Model Tick?

The intermediate COCOMO model starts with the Basic COCOMO effort equation:

$\text{Effort}=a\times (\text{KLOC}{)}^b$

Where KLOC is thousands of lines of code, and a and b vary by project type (Organic, Semi-Detached, Embedded). For example, Organic projects use $a=2.4$ and $b=1.05$.

The magic happens in the Intermediate step: multiply this by an Effort Adjustment Factor (EAF) derived from the 15 cost drivers. The full COCOMO model formula becomes:

$\text{Person-Months}=a\times (\text{KLOC}{)}^b\times {\prod }_{i=1}^{15}{E}_i$

Here, each $E_i$ is a multiplier (e.g., 0.86 for "Reliable" rated Low). Nominal values are 1.0; deviations adjust up or down. Schedule follows as $TDEV=3.0\times (\text{Person-Months}{)}^{0.32}$. This setup shines for mid-sized projects, offering 15-20% better accuracy than Basic COCOMO.

Curious about variations? Check types of COCOMO model or compare COCOMO model 1 and 2 in Software Engineering for context—COCOMO II adds more drivers but feels heavier for beginners.

The 15 Cost Drivers: Your Estimation Superpowers

These drivers fall into four groups: Product, Hardware, Personnel, and Project Attributes. Rate each honestly based on your project's realities—grab our free checklist below for quick reference. Here's the breakdown with typical multipliers (from Boehm's tables):

Product Attributes (Focus on Software Itself)

• RELY (Required Reliability): How critical is failure? Extra High (e.g., avionics) = 1.40; Low (simple app) = 0.86.

• DATA (Database Size): Large databases inflate effort. High = 1.15; Low = 0.94.

• CPLEX (Complexity of Product): Algorithm-heavy? Very High = 1.42; Nominal = 1.00.

• RUSE (Developed for Reuse): Reusable code demands more upfront work. High = 1.07; Very Low = 0.95.

Hardware Attributes (Platform Constraints)

• TIME (Execution Time Constraint): Tight runtime? High = 1.11; Low = 1.00? No, it's 0.87 for Low.

• STOR (Main Storage Constraint): Memory limits? High = 1.06; Low = 0.87.

• VIRT (Virtual Machine Volatility): Frequent platform changes? High = 1.15; Low = 0.87.

• TURN (Turnaround Time): Slow compiles hurt productivity. High = 1.11; Low = 0.87.

Personnel Attributes (Team Factors)

• ACAP (Analyst Capability): Top analysts speed things up. Very High = 0.71; Low = 1.29.

• AEXP (Applications Experience): Seasoned teams? Very High = 0.82; Low = 1.22.

• PCAP (Programmer Capability): Elite coders = Very High 0.70; Low = 1.42.

• VEXP (Virtual Machine Experience): Platform pros? Very High = 0.82; Low = 1.21.

• LEXP (Language Experience): Fluent in the stack? Very High = 0.95; Low = 1.14.

Project Attributes (Management Choices)

• MODP (Modern Programming Practices): Agile, TDD? Very High = 0.76; Low = 1.24.

• TOOL (Use of Software Tools): Automation lovers? Very High = 0.71; Low = 1.24.

• SCED (Required Development Schedule): Rushed? High = 1.23; Low = 1.00? Actually, Low (more time) = 1.23, wait no—compressed schedules hike it: Golden = 1.43.

Pro tip: Multiply all E_i values for your EAF. A 100 KLOC Organic project at nominal (EAF=1) takes ~252 person-months. Bump RELY to High (1.15) and SCED to High (1.23)? EAF jumps to 1.41, effort to ~356 months. See software engineering cocomo model example for worked samples or software engineering cocomo model notes and software engineering cocomo model pdf for deeper dives.

Step-by-Step: Applying the Model in Practice

1. Estimate Size: Use function points or historical data for KLOC.

2. Classify Project: Organic (small teams), Semi-Detached (medium), Embedded (hardware-tied).

3. Rate Drivers: Score each of the 15 using the checklist—be brutally honest.

4. Calculate EAF: Product of multipliers.

5. Compute Effort/Schedule: Plug into formulas.

6. Refine: Cross-check with tools like COCOMO II calculators.

Real-world win: A fintech app team used this to spot high SCED and low ACAP risks, adding two analysts and extending by 20%, avoiding crunch-time disasters.

Common Pitfalls and Pro Tips

Don't overrate "Nominal"—most projects skew higher. Ignore hardware drivers for cloud-native apps? Big mistake. Always validate KLOC with proxies like story points.

For modern twists, blend with Agile: Use velocity for size, then apply drivers to sprints.

Grab Your Free Checklist Now

Mastery comes with practice. Download our COCOMO Intermediate 15 Cost Drivers Checklist (PDF)—a one-pager with ratings, multipliers, and examples. Download here (link placeholder—replace with your asset).

There you have it: the Intermediate COCOMO Model demystified. Whether you're a student prepping exams or a PM quoting clients, these 15 drivers turn guesswork into science.