How to Estimate a Freelance Project Accurately (PERT, Bottom-Up, and the Buffer Math)

Most freelancers estimate by gut feel and pay for it on every project. The fix is not "be more careful." It's a method: bottom-up breakdown, PERT for uncertainty, historical data as the calibrator, plus a 10-25% buffer based on project type. This guide is the working playbook with a worked PERT example you can run today.

Why Most Freelance Estimates Miss

The freelancer estimation problem is structural, not a discipline issue. Three patterns produce nearly every bad estimate:

1. Estimating in big chunks. "The whole landing page is 30 hours." Big numbers hide work. Break it into 1-4 hour tasks and you'll find 8-12 you forgot.
2. No buffer. Optimistic estimate becomes the number you quote, then revisions, calls, file handoff, and unexpected tooling issues eat the margin.
3. No memory. Without tracking actual hours on past projects, every estimate is a fresh guess. After 30 projects you should know your numbers; most freelancers don't.

The good news: each of these has a fix that takes minutes per project to install.

The Three Methods (Use All Three)

The strongest estimates combine all three: bottom-up for the structure, PERT for uncertain pieces, historical average as the sanity check.

Method 1: Bottom-Up Breakdown

Bottom-up is the foundation. Per Motion's deep dive on bottom-up estimation, it is the most reliable starting point for any project, no matter how familiar.

The rules:

• Every task is 1-4 hours. If it's bigger, you're hiding something; break it down further.
• Include non-deliverable tasks. Discovery calls, revision rounds, file prep, handover, project management.
• Be specific. "Design hero section" beats "design page." Specificity surfaces forgotten work.

Example: landing page bottom-up

This is the bottom-up estimate. Notice how "design landing page" hid 25 hours behind one phrase.

Method 2: PERT (When Uncertainty Is Real)

PERT (Program Evaluation and Review Technique) gives you a weighted average of three scenarios. Per Project Management Academy's PERT guide, use it for tasks where you can't bottom-up estimate confidently because you've never done quite this work before.

The formulas:

PERT estimate = (Optimistic + 4 × Most Likely + Pessimistic) / 6
Standard Deviation = (Pessimistic - Optimistic) / 6
95% Confidence Range = PERT ± 2 × Standard Deviation

The 4× weighting on the most likely outcome makes it heavier than a naive average, which matches how project work usually plays out (most projects land near the most likely estimate, occasionally explode toward pessimistic).

Worked Example: Custom Landing Page

PERT = (12 + 4 × 20 + 40) / 6 = (12 + 80 + 40) / 6 = 22 hours

Standard Deviation = (40 - 12) / 6 = 4.67 hours

95% Confidence Range = 22 ± 9.3 = roughly 13 to 31 hours

The PERT number (22) is higher than your most-likely guess (20) because it accounts for the realistic probability that the project skews toward pessimistic. That's by design.

What the Standard Deviation Tells You

Per PM Study Circle's three-point guide, the standard deviation reveals how much risk your estimate carries. A high SD (in this case, 4.67 hours on a 22-hour estimate = 21% volatility) means the range is wide and you should price for the upper end. A low SD means you can confidently quote near the PERT number.

Method 3: Historical Average (The Most Accurate One)

If you've done 5+ similar projects, your historical actual hours are the most accurate estimator you have. Beats PERT, beats bottom-up.

The process:

1. Open your time tracking log (you have one, right? See freelance time tracking systems if not).
2. Pull actual hours from your last 5-10 similar projects.
3. Calculate the mean.
4. Add a 15-20% buffer for the new project's unknowns.

Example: average of last 5 landing pages = 31 hours. Add 18% buffer = 36.6 hours. That's your number.

Compare that to your PERT estimate (22 hours) and bottom-up estimate (25 hours). The historical number is highest because reality includes things you forgot to bottom-up estimate. Trust the historical data.

If the three numbers disagree by more than 30%, run the bottom-up again to find what you missed.

The Buffer: Why 15-25% Isn't Optional

Per Monday.com's 2026 estimation guide and AACE International standards, the buffer is the single highest-leverage line in your estimate. Industry standards:

• Work-level (individual task): 10-20%
• Project-level (overall): 15-25%

Why buffer matters:

• Client illness or vacation that delays decisions
• Scope clarifications that add small tasks
• Tool failures (Figma export bug, hosting issue)
• Your own learning time on new techniques
• Email and Slack overhead
• Delivery and handover details

Buffer percentages by project type:

Per AACE International guidelines, early-stage Class 5 (concept screening) estimates may carry up to 40% contingency, dropping to 15% for Class 2 (definitive) estimates as scope clarifies.

Skipping the buffer is the #1 cause of underbilling. The freelancers who say "I always lose money on projects" almost universally don't add a buffer.

Presenting the Estimate to the Client

A single number (e.g., "30 hours, $4,500") gives the client false precision and gives you no room to renegotiate when reality shifts. Present a three-tier estimate instead:

You can tighten this for the formal quote (e.g., "Project investment: $4,500 for the most likely scope; additional time billed at $150/hour if scope expands beyond [list]"), but the conversation about variability has already happened.

The handoff to a clean quote is where the proposal builder and quote writing post take over. The estimation work feeds the document; the document protects the estimate.

Hourly vs Fixed Price: How Estimation Changes

Estimation feeds two different pricing models:

Hourly billing: estimate is for client communication only. They want a cap or expected range. Track actuals, bill what you worked. Buffer is less critical because you bill all hours.

Fixed price: estimate is the number you live with. Buffer is mission-critical. Underestimate by 30% on a fixed price and your effective hourly rate craters. Overestimate by 30% and you might lose the project to a competitor.

For the deeper tradeoff, see project quote vs hourly estimate and freelance pricing models.

The Calibration Loop (This Is Where Freelancers Stop)

Per Scopestack's research on estimation accuracy and Planning Poker for freelancers, the single biggest accuracy improvement comes from looking back. After every project, log:

Review quarterly. Look for patterns:

• "I underestimate revision rounds by 50% on every project."
• "Series A clients add scope 80% of the time."
• "Custom illustration tasks always take 1.5x my estimate."

These patterns become your new buffer rules and your new bottom-up assumptions. Per Planning Poker's research, after 10-15 projects of deliberate calibration, estimation becomes near-automatic. Within 12 months your estimates land within 15-20% of actual on familiar work, which is roughly the upper limit of what any freelancer hits.

Common Estimation Mistakes (Avoid These)

The patterns that consistently cause underbilling:

• Optimism bias on familiar work. "I've done this before, it'll be quick." Then the client adds three revision rounds.
• Skipping the bottom-up step. "I'll just round up the number." Round-ups always under-include hidden tasks.
• Treating one estimate as the budget. Without best/likely/worst framing, the client expects "likely" and you absorb every overage.
• Estimating only deliverable work. Calls, emails, project management, handover, file prep can be 20% of the project. Estimate them.
• No buffer. This is the dominant pattern. See above.
• Ignoring standard deviation in PERT. Knowing the SD tells you whether to quote near the mean or the upper range.
• Never reviewing. Without a calibration loop, you'll be wrong by the same amount on project 100 as on project 5.

What to Do Next

If you're in the middle of estimating a project right now:

1. Bottom-up the tasks at 1-4 hours each. Include non-deliverable work.
2. Run a PERT calc if any task is uncertain. Calculate the SD.
3. Pull historical hours from past similar projects (start tracking if you haven't).
4. Add 10-25% buffer based on project type (use the table above).
5. Build a three-tier estimate. Present the middle to the client.
6. After the project closes, log actual vs estimate. Review the gap.

For the quote document itself, see freelance quote writing and how much to quote. For protecting the estimate from scope creep, see handle scope creep and scope of work. For the historical-data source that powers your calibration, see freelance time tracking systems.

Estimation is a skill that improves only with reps and feedback. The freelancers who stay bad at it are the ones who never look back. Track the gap, and your future-self will quote with a confidence the present version doesn't have yet.

References

1. Bottom-Up Estimating Deep Dive, Motion
2. Project Estimation Methods 2026 Guide, Monday.com
3. Project Contingency Budget, Monday.com
4. PERT Three-Point Estimating, Project Management Academy
5. Three-Point Estimation Formula, PM Study Circle
6. Three-Point Estimating, Galorath
7. Project Estimation Techniques and Challenges, Scopestack
8. Planning Poker for Freelancers and Consultants, Planning Poker
9. 6 Project Estimation Techniques, Scoro
10. Bottom-Up Estimating in Project Management, Wrike