Bottom Up Budgeting for Architecture and Engineering Firms

Most project budgets in A&E firms are built top down. Total fee divided by phases, phases divided by roles, roles divided by months. The math works. The result rarely reflects the actual work.

Bottom up budgeting flips the order. Start with tasks. Sum to phases. Sum to total. The number you get is harder to produce but far more honest.

Top down vs bottom up: the real difference

Top down

Start with the total fee. Allocate to phases by traditional percentages (15-20 percent SD, 20-25 percent DD, 35-40 percent CD, etc.). Allocate phases to roles. Done in 90 minutes.

Strength: fast, gives you a defensible budget for proposal stage when you do not have time for granular detail.

Weakness: the budget often does not survive contact with the actual work. The CD phase looks reasonable on paper and runs 30 percent over because the actual deliverables were never tallied.

Bottom up

Start with the task list. Estimate hours for each task. Roll up to deliverables. Roll up to phases. Roll up to total. Done in two to four hours.

Strength: the budget reflects the work. The PM can defend every line. Tracking actual against budget at the task level catches drift early.

Weakness: takes time and requires real estimating discipline. Garbage hours estimates produce a garbage budget faster than top down does.

When to use which

The choice is not all or nothing. Most strong firms use both, in sequence.

• Proposal stage: top down for speed.
• Contract signed: bottom up for the executable budget.
• Quarterly reforecast: bottom up to reflect actual progress.

The top down number on the proposal and the bottom up number internal should land within 10 percent of each other. If they diverge by more, the proposal needs revisiting before signing.

The five step bottom up process

Step 1: Build the task list

Walk each phase and list every deliverable. For each deliverable, list the tasks needed to produce it. A typical SD phase for a 50,000 sq ft commercial project produces 30 to 80 tasks.

Examples of tasks: floor plate study at three options, code analysis memo, structural concept narrative, schematic plumbing riser, exterior facade study at two options, client review meeting and revision cycle.

Step 2: Estimate hours per task

Two columns: task and estimated hours. Use historical data from comparable past projects if you have it. If not, ask the senior team member who would do the work to estimate. Junior staff under estimate. Senior staff with experience estimate within 15 percent of actual.

Step 3: Assign role mix

For each task, what mix of roles will do the work. A schematic plumbing riser might be 70 percent senior engineer, 30 percent technician. A code analysis memo might be 50 percent project architect, 50 percent senior architect.

Step 4: Apply rates

Hours by role multiplied by billing rate per role. Sum to get fee per task, fee per deliverable, fee per phase.

Step 5: Sanity check against top down

Compare the bottom up total against the top down number from the proposal. Within 10 percent: proceed. Between 10 and 20 percent: investigate the gap before locking the budget. Over 20 percent: something is wrong with one of the two numbers, find which.

The contingency line

Bottom up budgets need a contingency line at the end. Even with disciplined estimating, the team will encounter scope ambiguity, client driven rework, and unforeseen coordination cost.

For projects with stable scope, 10 to 12 percent contingency on labor is normal. For projects with unclear scope, 15 to 20 percent. Skip contingency entirely and the bottom up budget produces overruns that the top down version would have absorbed in slack.

How to track against the bottom up budget

The point of a bottom up budget is that you can see drift early. The PM compares actual hours to estimated hours weekly, at the task level for the active phase.

Three traffic light states.

• Green: actual hours within 10 percent of estimate.
• Yellow: 10 to 20 percent over. Investigate this week.
• Red: over 20 percent. Conversation with the principal this week.

Without the bottom up structure, drift is invisible until the end of the phase. With it, drift is visible in the second week.

The team behaviors that make it work

A bottom up budget that nobody updates is worse than a top down budget. Three behaviors keep the model alive.

1. The PM updates progress weekly, the same 15 minutes as WIP review.
2. The senior team member who built the estimate is the same person who reviews the actual against estimate.
3. The lessons feed forward. After project close, the estimating team reviews variance by task type and updates the firm's estimating reference.

The compounding return

Each bottom up budget produces estimating data. After 10 projects, the firm has real intelligence on how long specific task types actually take. After 30 projects, that intelligence is sharper than any vendor benchmark or industry average.

Fee estimating at proposal time gets faster and more accurate when fed by bottom up history. Project budgeting gets sharper when the discipline becomes institutional.

The firms that commit to bottom up budgeting for two years end up with an estimating asset that is genuinely worth money. The firms that stay with top down keep guessing.