Workflow failure · Operator field guide

Rate-card drift is a governance failure hiding inside old bids

A stale rate rarely announces itself. It enters through copied history, survives because the total feels plausible, and becomes the company's unofficial price one bid at a time.

Trace 01 · Signal → cause → consequence

Diagnose the operating failure before buying a tool

Visible signal

Recent estimates contain several values for the same role or package, and producers cannot distinguish a negotiated exception from an outdated default.

Underlying cause

Historical bids are easier to copy than the official rate card is to use. Exceptions are not tagged, reviewed, or fed back into governance.

Business consequence

The shop cannot tell whether it is deliberately pricing a relationship or accidentally inheriting old economics. Margin analysis becomes noisy and corrective action arrives late.

Control 02 · Operating principles

Three controls that survive the software

Drift is measurable

Compare each sold rate with the active card at issue time. The difference, reason, operator, and value at risk should be visible without opening the old estimate.

Exceptions are not defaults

Client accommodations, bundled packages, and strategic pricing can be valid. Keep them scoped to the decision instead of letting the next copied file promote them into policy.

Evidence changes the card

Update rates from repeatable vendor quotes, crew expectations, market evidence, and closed-job performance. Anecdote should trigger review, not automatic repricing.

Runbook 03 · Smallest useful workflow

Run this on one real job

Do not begin with a company-wide migration. Prove the control on representative work, record the exceptions, and expand only when the operator can trust the new state.

01

Run a 90-day drift audit

Extract the highest-frequency roles, packages, and percentage lines from recent bids. Compare them with the card that should have applied when each estimate was issued.

  • Use issue date, not current card, for historical comparison
  • Separate client-specific agreements
  • Quantify revenue and margin exposure
02

Classify every material override

Use a small reason set: market evidence, relationship concession, package logic, scope exception, correction, or unauthorized drift. Free-text notes can add context but should not replace classification.

  • Threshold for materiality is defined
  • Operator and approver are recorded
  • Unauthorized drift has a corrective owner
03

Fix the access path

Make the active card easier to use than an old estimate. Surface matching rates inside estimating, flag fallbacks, and explain why a recommended value applies.

  • Card is available inside the estimate workflow
  • Missing matches are visible
  • Copy-forward does not silently retain old values
04

Review patterns monthly

Look for repeated overrides by role, client, market, and producer. Update the card, create a client rule, coach the operator, or accept the pattern deliberately.

  • Repeated exceptions have a decision
  • Change notes are published
  • New defaults have an effective date

Instrument 04 · Evidence

Measure whether the control is earning its place

  1. Dollar-weighted deviation from active card
  2. Overrides without a classified reason
  3. Repeated exceptions promoted into reviewed rules
  4. Estimates using baseline fallback despite a matching company rate

Boundary 05 · Product truth

Where Production Engine fits today

The current build is strongest from company rate card through estimate, option approval, and initial budget creation. The design-partner program exists to test the next control on live work without pretending the whole production stack is finished.

Present in the current repo

  • Stores a tenant's active rate card independently from historical estimates
  • Places matching company rates in the estimator's primary context
  • Preserves estimate payloads and normalized line records for audit

Design-partner scope

  • Load the high-impact current card
  • Define override reasons and review thresholds
  • Use live bid corrections to specify the drift-reporting surface

Honest boundary: Production Engine does not yet claim a finished override-audit dashboard. The current foundation makes the records separable; the operating report must be validated with design partners.

Paid design-partner program

Put one live workflow under control in 90 days.

Implementation, rate-card and workflow mapping, access for five operators, and direct product-team collaboration.

$2,500 implementation + $499/month for five operators · 90-day commitment

FAQ 06 · Buying questions

Questions to resolve before implementation

How much variance from the rate card is acceptable?

Set thresholds by category and risk. The key is not zero variance; it is that material variance is intentional, attributable, and reviewed.

Should negotiated client rates live in the main card?

Usually as a scoped client rule or documented exception, not as the market default. Otherwise one relationship can distort every later bid.

Can actual costs set sell rates automatically?

They should inform the decision, not make it alone. Sell rates also reflect overhead, utilization, risk, market position, and the value of the production approach.

Index 07 · Internal route

Continue the operating system

Paid design-partner program

If this failure costs real producer time or margin, test it on a live job.

Implementation, rate-card and workflow mapping, access for five operators, and direct product-team collaboration.

$2,500 implementation + $499/month for five operators · 90-day commitment