The technical office as a margin guardian before you sell

# The technical office as a margin guardian before you sell

A simple filter so your best deal doesn’t become your worst project.

Closing a deal isn’t the same as winning a project. Winning—at least the way I define it—means delivering what you promised without bleeding internally: no permanent urgency, no invisible hours, no internal blame, no “surprise costs” discovered when you can’t change anything anymore.

When margin collapses, teams usually fight it late: calls, meetings, pressure, heroic patching. But margin rarely breaks in the final week. It breaks at the start, when the project enters the system with uncertainty, variability, and ambiguous promises.

This is where a role makes a huge difference: the technical office as a margin guardian.

I’m not talking about bureaucracy. I’m talking about a system that does one job: prevent selling what cannot be executed coherently (or at least make the real price of that promise visible).

## The mechanism: how profitability evaporates quietly

I see it as a chain:

1. Promise: what was said and what the client believes they bought.
2. Real specification: what will actually be built/installed.
3. Variability: what changes because it wasn’t decided (or measured).
4. Friction: the cost of coordinating the uncertain (rework, doubts, back-and-forth).
5. Hidden cost: untracked hours, rush purchases, extra site visits, avoidable errors.

If your technical office doesn’t validate this chain before the sale becomes execution, you operate on “faith margin”: you hope it fits.

## Signs you’re selling “not-executable” projects

If these feel normal, your margin is exposed:

• • Quotes full of assumptions: “to be confirmed”, “subject to site conditions”, “we’ll adapt on site”.
• • Last-minute changes as routine: when change is the default, it’s not flexibility—it’s lack of closure.
• • Strong visuals, weak decisions: beautiful renderings but missing what drives cost (interfaces, tolerances, constraints).
• • Real cost appears late: you discover it at purchasing or production.
• • The technical office is firefighting: solving contradictions created upstream.
• • “Star” projects that later choke operations: the promise sells, but the system can’t sustain it.

I’m not debating whether you should sell. I’m debating how you sell without mortgaging execution.

## What this article is NOT

This is not a full implementation playbook. I focus on:

• • symptoms
• • risks
• • decision criteria
• • how to detect where margin breaks before signature.

Implementation (roles, cadences, templates) belongs to a tailored diagnosis.

## The most expensive blind spot: pre-sale → execution

Custom projects have a conversion point: sales becomes operational commitments.

That’s where you decide:

• • what is closed vs open
• • who owns each risk
• • what variability is acceptable
• • what depends on third parties
• • what options are truly equivalent (and what explodes cost/lead time).

Without a gate, the project enters with noise. Noise gets paid.

## My definition to avoid blame: “technical margin”

When things go wrong, the conversation becomes emotional. I prefer a cold definition:

Technical margin = expected margin – uncertainty cost – variability cost – friction cost.

The technical office can reduce uncertainty, variability, and friction before they turn into hours and errors.

## The minimum system that works: 4 gates before you “truly” sell

You don’t need a monster. You need four repeatable decisions.

### Gate 1: Feasibility (can we execute with our real standard?)

• • Is it buildable/installable with your real standard (not the ideal)?
• • Does it fit capacity and promised lead time?
• • Are there critical external dependencies?

Typical alarm: “Yes, but…”

### Gate 2: Sufficient specification (is what drives cost actually closed?)

You don’t need everything closed, but you do need:

• • key dimensions and tolerances
• • critical materials/finishes
• • interfaces and constraints
• • responsibilities (who does what, when, under what criteria).

Typical alarm: “We’ll decide later.”

### Gate 3: Risks and assumptions have owners (is the open part governed?)

Assumptions are fine. Invisible assumptions are not.

Every assumption needs:

• • validation event
• • decision owner
• • estimated impact
• • deadline to close.

Typical alarm: assumptions with no owner.

### Gate 4: Economic coherence (does the price reflect complexity and coordination?)

Criteria—not opinions:

• • minimum margin by project type
• • complexity score (variety, singularities)
• • coordination cost (subs, site windows, extra visits)
• • change probability.

Typical alarm: an “entry price” hoping to fix later.

## Value engineering: saying “no” without saying “no”

Healthy value engineering preserves perceived value while protecting cost.

It’s healthy when decisions happen early and impacts are explicit. It’s toxic when substitutions happen late, without criteria, and future impact is hidden.

## The board I want: digital, simple, alive

One board that answers:

• • which gate is each project in?
• • what risks are open and who owns them?
• • what assumptions exist and when must they close?
• • what is the complexity score?
• • what is the likely margin impact (even as ranges)?

## Where it really breaks (and why it hurts the same way)

In audits, I usually find the same leak points:

• • environment interfaces are underestimated
• • “small” decisions repeated many times aren’t small anymore
• • uncontrolled variation
• • blurry responsibilities
• • no operational limits for commercial promises.

## Diagnostic questions I ask

• • What part of margin is decided before signature vs after?
• • Which assumptions repeat in most quotes—and why?
• • Which project types consistently become painful—and what do they share?
• • Where does rework emerge: design, production, installation, service?
• • Who can stop a sale on feasibility—and by what criteria?
• • What “surprises” repeat? (repeating surprises are patterns)

## Fast checklist

• • Can the technical office say “NO” with criteria and an alternative?
• • Is there a formal gate between sale and execution?
• • Do assumptions have owners and deadlines?
• • Is there a minimum spec standard?
• • Do you measure rework linked to pre-sale ambiguity?
• • Do you review complexity before signature?
• • Do you treat high-risk types differently?
• • Do key decisions live in one accessible place?

Fail 3+ and margin isn’t protected. It’s gambled.

## Closing

If you want, I can help you diagnose where margin breaks before you sell, and design minimal gates that fit your reality without bureaucracy.

Diagnóstico express