Closing the loop: when job cost feeds the next estimate (without vanity scoreboards)

Many commercial roofing divisions run win, produce, and get paid as three separate conversations. The estimator bids. The field builds. The office collects. Then the estimator bids the next similar roof using the same assumptions—because nobody told them what actually happened. The money loop only closes when produce and get paid feed win.

Closing that loop means what you learned on the last job reaches the person pricing the next one. This piece closes Week 5, which started with job costing discipline and moved through the get-paid collection rhythm.

The loop is a process, not a report

Feedback to estimating is often treated as an occasional complaint: “that job killed us.” That is a feeling, not feedback. It does not tell the estimator which assumption was wrong, or whether the next bid should change at all.

A real feedback loop looks like:

• A closeout review on every commercial job, held soon after final costs land
• Variances explained in words, tied to specific estimate assumptions
• Lessons written down in a place the estimator actually uses
• A decision on each lesson: change the estimate, change the process, or leave it alone
• A check on the next similar bid that the lesson was applied

If lessons live only in someone’s memory, the next bid inherits the same blind spot.

Where the loop breaks

The tells are familiar:

• The closeout review never happens because everyone is on the next job
• Costs are too loosely coded to compare against the estimate
• The field blames the bid, the estimator blames the field, and nothing changes
• A single bad job triggers padding across every future bid
• Change-order work gets mixed into sold scope, so the lesson is wrong
• Good news never travels—jobs that ran clean teach nothing

These are design problems, not people problems. The loop needs an owner, an artifact, and a slot on the calendar.

The closeout review (short, factual, repeatable)

Keep it brief enough that it actually happens. Use the same questions every time:

1. What did we sell? Scope, exclusions, and key assumptions from the estimate and the award handoff.
2. What did it actually take? Labor, materials, equipment, and subs by the same codes the estimate used.
3. Where did it differ—and why? Reasons in plain words: access, weather, sequencing, missed scope, unwritten changes, productivity.
4. Was the cause the bid, the job, or the handoff? Each one has a different fix.
5. What changes next time? An estimate assumption, a checklist item, a qualification question—or nothing.

Question four matters most. Not every variance means the estimate was wrong. Some mean the handoff was dirty or a change went unwritten. Fix the right thing.

Turning lessons into estimating inputs

A lesson is only useful if it changes something the estimator touches. Useful destinations:

• Estimating assumptions. Update the standard assumptions for similar roof types, access conditions, or project sizes.
• Scope and exclusion language. Add clarity where a gap caused a dispute. This builds on estimating and scope discipline.
• Qualification questions. If a type of job repeatedly surprises you, ask about it earlier in pipeline qualification.
• Handoff checklist. If the variance came from missing information at award, fix the package.
• Billing terms. If collection stalled, capture those requirements at bid time.

Keep a short lessons log. One line per lesson: the job, the assumption, the change made, and who made it.

Watch for over-correction

A painful job creates pressure to overreact. The estimator adds cushion everywhere, the next bids get less competitive, and the division starts losing work it could have delivered well. That is not learning. That is flinching.

Before changing an assumption, ask whether the variance came from something that will repeat—a roof type, an access condition, a customer pattern—or from a one-time event. Repeating causes deserve estimate changes. One-time events deserve a note and a process fix. When you are unsure, flag the assumption and watch the next similar job before rewriting the standard.

Seats: who owns the loop

Use the seats from naming seats without a bloated org chart:

• Produce brings field truth: what happened and why.
• Get paid brings cost and billing truth: final costs, changes, and collection friction.
• Win owns the decision: what changes in estimating, qualification, or scope language.
• The owner owns the standard that the review happens—and settles disagreements once, then updates the rules. See what only the owner should own.

Put the closeout review inside your weekly operating rhythm instead of creating a new meeting.

What not to invent in the name of feedback

Under pressure, owners reach for scoreboards. Be careful.

Avoid:

• Win-rate or margin charts that nobody can connect to a specific decision
• Padding every bid because of one painful job
• Rankings that turn the review into blame
• Averages that hide the job that actually went sideways

Prefer:

• Variance reasons in plain language
• A lessons log with decisions and owners
• A check on the next similar bid
• Clean jobs reviewed too, so good assumptions are confirmed

Build better measures once the loop runs reliably. A scoreboard without a feedback process just decorates the problem.

Rules that keep the loop closed

Rule: every commercial job gets a closeout review. Short is fine. Skipped is not.

Rule: compare by the same codes. If the estimate and the job cost speak different languages, nobody learns anything.

Rule: diagnose before you adjust. Bid, job, or handoff—fix the one that caused the variance.

Rule: lessons have a home. If the estimator cannot find it on the next bid, it does not exist.

Soft next step

This week, pick one recently completed commercial job. Hold a short closeout review using the five questions above. Write down every lesson and decide for each one: change the estimate, change the process, or leave it alone. Then name the next similar bid where the lesson must be applied—and ask the estimator to confirm it was.

If you want help building a commercial division where produce and get paid feed win work—not vanity scoreboards—talk it through with Executive Roof Coach. The next bid should know what the last job taught you.