The Engineering Deck That Gets Approved Shows the Board Only One Layer

In 2019 I prepared a head of engineering for a capital-expenditure approval at an industrial manufacturer — a seven-figure plant-automation programme he had spent four months designing. His deck was thirty-one slides and it was, in engineering terms, beautiful. It opened with the current line layout, moved through throughput analysis, then the automation architecture, then the reliability modelling, then a phased commissioning plan, and arrived at the recommendation on slide twenty-six. I watched him present it to a rehearsal panel I had assembled, and within three minutes the stand-in finance director had flipped to the back of the pack, found the cost summary, and stopped following the build entirely — reading ahead, circling two figures, waiting for the engineer to catch up to where her attention already was. When he finally reached the recommendation, she had been sitting on her question for twenty minutes and asked it as a challenge rather than a query. The real board did something similar the following week. The programme was eventually approved, but only after a second meeting, because the first one never got the board to the decision while they were still fresh enough to make it.

(This article was created with AI assistance; all stories and insights are based on 35 years of real client work.)

That rehearsal crystallised the rule I now teach every engineer who presents capital projects to a board: build the deck the opposite way to how you built the project. You built it from the foundations up, because that is how engineering works — you cannot pour the structure before the footings. But a board reads from the top down, because that is how decisions work — they need the conclusion first so they know what the evidence is for. Present an engineering deck in build order and you force the board to hold four months of reasoning in their heads before they learn what it is reasoning towards. This piece walks through the structure that fixes it: the Three-Layer Structure, which puts the decision on the spine and the engineering in the appendix; the which-layer test that keeps you from collapsing a won approval by over-answering; and how to build a single deck that satisfies both the non-technical majority and the one director who can read your calculations.

Why engineers build the deck from the foundations up

The build-order instinct is not laziness; it is competence misapplied. An engineer is trained to show their working, because in engineering the working is the proof, and a conclusion without the derivation behind it is worthless — even dangerous. Carried into the boardroom, that training produces a deck that earns trust the way an engineering review earns trust: by laying out every step so the reader can verify the logic. The trouble is that a board is not verifying your logic. It does not have the expertise to, it does not have the time to, and crucially it does not want to. It is deciding whether to commit the company’s money and risk to your recommendation, and that decision needs the recommendation first.

This is the mechanism behind the deferral. When the recommendation arrives on slide twenty-six, everything before it was, from the board’s point of view, evidence for a claim they had not yet heard. They cannot weigh evidence for an unknown conclusion, so they do one of two things, both bad. The patient directors disengage and wait, which means your most careful analysis is delivered to people who have stopped listening. The impatient ones flip to the back, find the number, and form a view before you have framed it — so your recommendation lands into a judgement that has already half-formed without your help. The general principle of presenting to a board of directors is that the board sets the order of attention, not the presenter, and the board’s order is always conclusion-first.

The reframe is uncomfortable because it feels like hiding your work. It is not. You are not removing the derivation; you are moving it to where a board reaches for it — after the decision is on the table, not before. The four months of reasoning still exist, still defensible, still rigorous. They simply stop being the path the board has to walk to reach the recommendation, and become instead the ground the board stands on once the recommendation has their attention. Same content, inverted order, and the inversion is the entire difference between an approval in one meeting and a deferral into two.

The Three-Layer Structure

The Three-Layer Structure organises an engineering presentation into three distinct layers, each written for a different moment in the board’s attention and each reached only when the previous one has done its job. The layers are not three sections the board reads in sequence; they are three depths, and the board descends only as far as it needs to. Most approvals are won entirely in the top layer. The lower two exist so that when a director wants to go deeper, the depth is there — structured, reachable, and clearly separated from the decision it supports.

The three layers are these. Layer 1, the Decision Layer, is the spine: the recommendation stated as a recommendation, the specific ask, the options you rejected and why in one line each, and what happens to the business if the board approves and if it does not. This is three to five slides, and it is all the board sees by default. Layer 2, the Case Layer, is the business reasoning behind the recommendation — the cost breakdown, the risk assessment, the comparison of options, the payback or continuity argument. The board reaches Layer 2 when it engages with the decision and wants to discuss it. Layer 3, the Proof Layer, is the engineering itself: the throughput analysis, the reliability modelling, the architecture, the calculations. It lives in a clearly tabbed appendix, and the board reaches it only when a director challenges a specific technical point and you turn to it on demand.

What makes the structure work is that each layer is complete in itself and written in the language of its moment. Layer 1 is in the language of business consequence, so a non-technical director can decide on it alone. Layer 2 is in the language of commercial reasoning, so it answers “why this and why now” without descending into specs. Layer 3 is in the language of engineering, so it satisfies the specialist completely — precisely because it is not diluted by trying to also serve the non-specialist. Separating the layers lets each be fully itself, which is impossible when all three are interleaved on the same slides, where the engineering dilutes the decision and the decision oversimplifies the engineering.

Layer 1: the only layer the board sees by default

Layer 1 carries the whole approval, and its discipline is severe: nothing enters it that the board does not need to make the decision. The recommendation goes first, stated as a position you are taking, not a topic you are introducing — “we should commit to the automation programme this quarter to remove the line-stoppage risk that cost us eleven days of output last year,” not “Plant Automation Programme Review.” Then the ask, exact and unhedged: the amount, the timeframe, the decision you want today. Then the options you considered and rejected, one line each, because a board trusts a recommendation more when it can see what it beat. Then the consequence both ways: what the business gains by approving and what it risks by waiting. That is the layer. Five slides at most.

The hardest discipline in Layer 1 is keeping the engineering out of it, because every instinct you have says the engineering is what justifies the recommendation. It does — but the justification belongs in Layers 2 and 3, available the instant the board wants it, not pre-emptively on the decision slides. A single reliability chart on the recommendation slide feels like support; to the board it is the first step down into a layer they had not asked to enter, and it invites the build-order disengagement all over again. Hold the line. The recommendation slide makes the case in the language of business consequence and points to where the proof lives. The proof does not climb up to meet it.

There is a credibility dividend in this restraint that engineers consistently underestimate. A presenter who can state a complex recommendation in five clean slides and then produce four months of rigour the moment it is requested reads as someone in complete command of their material — because holding the detail in reserve, ready but not deployed, is a harder skill than showing all of it. The board reads the restraint as confidence and the confidence as competence. The thirty-one-slide build, by contrast, reads as a presenter who could not decide what mattered and so showed everything, which paradoxically reads as less sure of itself, not more. The same conclusion-first discipline carries a multi-year technology roadmap to a board that would never sit through the underlying architecture.

The which-layer test for questions

The Three-Layer Structure has a second use beyond organising the deck: it tells you how to answer questions without collapsing the approval you have just won. When a director asks a question, the engineer’s reflex is to answer as completely as possible, which usually means dropping straight to Layer 3 — explaining the mechanism in full because the question brushed against it. That reflex is how a won approval comes undone. A board that was ready to approve hears two minutes of reliability modelling in answer to a simple question, loses the thread, and starts to wonder what else it does not understand. The which-layer test prevents this. Before answering, name silently which layer the question lives in, and answer from that layer first.

Most questions live in Layer 1 or Layer 2, even when they sound technical. “Are you confident this will work?” is a Layer 1 question about decision risk; it wants “yes, and here is the one thing that makes me confident,” not a tour of the failure analysis. “Why not the cheaper option?” is a Layer 2 question about the commercial case; it wants the one-line trade-off, not the spec sheet. Only a question that explicitly asks how the engineering works — “walk me through how the failover actually happens” — lives in Layer 3, and even then you turn to the appendix slide and answer in translated terms, offering the full depth only if they press. Answer at the layer the question was asked, and stop. You can always go deeper if they ask again; you cannot un-say two minutes of detail that lost the room.

The test also keeps you calm under challenge, because it converts a vague feeling of being interrogated into a simple act of classification. A hostile-sounding question is far less threatening once you have located which layer it belongs to, because the layer tells you the answer’s shape and length before you have to find its content. Engineers who learn to classify before they answer stop over-explaining, stop sounding defensive, and stop volunteering the very detail that gives a sceptical director something new to probe. The discipline is not to withhold; it is to match the depth of the answer to the depth of the question, and to trust the board to descend further only if it wants to.

Where the calculations belong

The fear that stops engineers moving the calculations to the back is that the work will look absent — that a five-slide decision layer reads as though no real analysis was done. The opposite is true, but only if the proof is visibly available rather than genuinely missing. The thirty-one slides from the automation rehearsal were not the problem; their position was. As the spine of the deck they caused a deferral. As a clearly tabbed Layer 3 appendix — throughput on tab one, reliability on tab two, commissioning on tab three — reached the instant a director asks, they would have signalled exactly the rigour the engineer wanted to convey, without forcing the board to wade through it to find the recommendation.

A well-built Proof Layer is not a dumping ground; it is engineered as carefully as the project. Each appendix slide answers one anticipated question and is labelled so you can find it in two seconds while a director waits. The test for what belongs in Layer 3 is simple: would a director plausibly ask about this, and if they did, would I want a slide rather than a verbal answer? If yes, it earns an appendix tab. If no director would ever ask, it does not belong in the deck at all — it belongs in your project file. The discipline of building the Proof Layer is the discipline of anticipating the board’s questions and having a slide ready for each, which is also, not coincidentally, the best preparation you can do for the meeting.

Reaching into the Proof Layer confidently is itself a trust-building act. A presenter who answers “what’s the failure rate on the existing line?” by turning straight to the reliability tab and reading the number has demonstrated more command than one who put the whole reliability analysis on the recommendation slide and hoped no one would ask. The first looks prepared; the second looks anxious. The board does not measure your rigour by how much of it you show unprompted — it measures your rigour by how readily you can produce the right piece of it on demand. A deep, well-organised appendix that you navigate fluently is the strongest evidence of competence you can offer a board, precisely because it stays out of the way until it is needed.

Building for the two readers on every board

In 2021 I coached a different engineer — a head of reliability at an energy company — who had absorbed the structure and then hit the objection every engineer raises: “but there’s a former chief engineer on our board who will want the detail, and if I bury it he’ll think I’m hiding something.” This is the real political problem, and it is what drives the over-showing in the first place. We solved it with the layers, not against them. He built Layer 1 for the non-technical majority who held the decision, and a deliberately deep Layer 3 for the one technical reader — then, in his opening, he said one sentence: “The recommendation and the business case are in the first five slides; the full reliability analysis is in the appendix from tab four, and I’ll go there as deep as anyone wants.” The former chief engineer relaxed visibly. He had been told the detail existed and where to find it. He asked for it twice during the meeting, got it instantly, and became the recommendation’s strongest advocate.

Every board has these two readers, and the structure serves both better than a blended deck ever could. The non-technical majority gets a decision they can make without wading through specs. The one technical reader gets depth that is not diluted by the attempt to also serve everyone else — and, told up front where it lives, feels respected rather than managed. The blended deck serves neither: it is too deep for the majority and too shallow for the specialist, so it irritates both. The Three-Layer Structure works because it stops pretending one set of slides can do two opposite jobs, and instead gives each reader the layer built for them. Non-executive directors in particular read for decision and governance, not for method, which is exactly what Layer 1 is built to give them.

The opening sentence the reliability engineer used is worth stealing wholesale. Signposting the layers at the start — here is the decision, here is where the case lives, here is where the proof lives, and I will go to any of them on request — does three things at once. It tells the non-technical majority they will not be drowned. It tells the specialist the depth is there. And it tells the whole board that you have organised your material around their decision rather than your build, which is a competence signal before you have presented a single slide. One sentence of signposting buys you the latitude to keep Layer 1 clean, because everyone now knows the rest is coming if they want it.

One thing to do before your next board deck

Take your next engineering board deck and split it physically into three labelled sections before you touch the content: Decision, Case, Proof. Move every slide into the section where it belongs, and be ruthless — if a slide explains how something works, it is Proof, even if you are proud of it, even if it currently sits near the front. Then rebuild Layer 1 from scratch to five slides: recommendation, ask, rejected options, consequence both ways, in the language of business, with no engineering on them. Write the one opening sentence that signposts all three layers. Finally, take your five most likely board questions and, for each, write down which layer it lives in and the one-line answer you will give from that layer. Walk into the meeting able to descend on demand and disciplined enough not to descend uninvited.

Frequently asked questions

Won’t the board think I haven’t done the engineering if the recommendation comes first and the analysis is in an appendix?

This is the fear that keeps the thirty-one-slide build alive, and signposting resolves it. In your opening, say one sentence: the recommendation and case are in the first five slides, the full analysis is in the appendix, and you will go there as deep as anyone wants. The board now knows the depth exists and where to find it, so a clean Layer 1 reads as command rather than concealment. The presenters who get accused of skating over the engineering are the ones who strip it out entirely or never mention it. Keep all of it, structure it into a navigable Proof Layer, point to it at the start, and turn to it instantly when asked. A board reads a presenter who can produce four months of rigour on demand as more rigorous, not less, than one who shows everything unprompted.

What if the board genuinely needs to understand a technical point to approve the spend?

Then that point gets translated into Layer 1, not left in Layer 3. The which-layer test cuts both ways: if a piece of engineering genuinely changes the decision, it has earned a place on the decision layer — but in translated form, stated as a business consequence with the proof still in the appendix. A board deciding between two designs may need to grasp one real difference between them; put that one difference on a Layer 1 slide in plain language, and leave the full comparison in Layer 2 and the calculations in Layer 3. The mistake is assuming that because one technical point matters to the decision, all of them do. Almost always, one or two load-bearing facts change the choice; translate those up and leave the rest where the board can reach it.

How is this different from just putting an executive summary at the front of my existing deck?

An executive summary bolted onto a build-order deck still leaves the build order underneath, so the moment you move past the summary you are back to presenting from the foundations up, and the board is back to waiting. The Three-Layer Structure restructures the whole presentation, not just its first slide. Layer 1 is not a summary of what follows; it is the entire presentation the board sees by default, complete in itself, with Layers 2 and 3 sitting beneath it as reachable depth rather than as the real deck the summary introduces. The difference shows the moment a question comes: with a summary-plus-build deck you answer by walking forward through the build; with the layers you answer by descending to the right depth and returning. One keeps the board waiting; the other keeps the board in control of how deep it goes.

How long does it take to restructure an engineering deck this way?

Re-sorting an existing deck into three labelled layers takes an afternoon the first time, because most of the work is moving slides you have already built into the right section and writing a clean five-slide Layer 1 on top. The slower part is rebuilding the decision layer in the language of business consequence rather than engineering, which can take a couple of hours of genuine thought — but that thought is the same thinking the board will force you to do anyway, so doing it at your desk is cheaper than doing it under challenge in the room. Once you have built two or three decks this way the structure becomes automatic, and new decks come out in layers from the start. The investment is front-loaded; the return is every approval meeting that finishes in one sitting instead of two.

The Winning Edge — weekly newsletter

The Winning Edge is a weekly (Thursday) newsletter for senior professionals who present to boards and executive committees. One short email a week on the structural moves that separate the technical proposals boards approve from the ones they defer. Subscribe to The Winning Edge →

For the full set of skills behind a board-ready engineering case — slides, storytelling, confidence, and delivery — the Complete Presenter bundle of seven products brings them together as a single resource — £99 for everything, lifetime access.

About the author

Mary Beth Hazeldine is Owner & Managing Director of Winning Presentations Ltd. With 24 years of corporate banking experience at JPMorgan Chase, PwC, Royal Bank of Scotland, and Commerzbank, she advises senior professionals across financial services, insurance, consulting, and technology on the structural moves that turn complex technical cases into decisions boards approve.