The offer that bleeds you

Why this matters to you right now

Bent Flyvbjerg studied more than 16,000 large projects in How Big Things Get Done (2023). Verdict: 91.5% overrun their budget, their schedule, or both. The less-known finding sits elsewhere. Most overruns are already baked into the project before launch, in scoping sentences that leave room for interpretation. His prescription fits in four words: "Think slow, act fast." Invest disproportionately in defining what "done" means before you start.

The consultant's reflex, when an engagement ends on degraded margin, is to look for the cause in delivery. A difficult client, an underestimated deliverable, a team that lacked rigour. Yet the bulk of the bleeding is decided between the writing of the offer and its signature, in sentences that seem clear to the consultant but leave the client room to interpret what counts as "done." That grey zone resolves systematically in one direction. The consultant delivers more, or the client pays less. Never the other way around.

Edition #22 on the phantom engagement described the drift during delivery and the signals that foreshadow it. This edition handles the upstream cause. The offer that makes the drift structurally possible.

The diagnostic test is simple. A valid Definition of Done passes one question: is it objectively done or not done, with no gradation. Not "done well," not "almost done," not "done to the client's satisfaction." Done or not done. If the phrase you are reading allows a nuance, it isn't a DoD, it is an intention.

If you cannot write a Definition of Done for the engagement you delivered last, retroactively, by rereading the offer you sent, your next engagement will bleed the same way.

Pricing: the DoD as a pricing instrument

The action: Price each phase of your offer against its Definition of Done, not against its hourly estimate. A phase has a binary state (done or not done), so its price is binary too. Three markers form an operational DoD. A firm time-box expressed in hours. A precise deliverable artefact with its minimum attributes: length, mandatory sections, format. A binary client acceptance criterion: signature, formal validation, or platform submission.

Why it works: The concept of Definition of Done is formalized in the Scrum Guide by Ken Schwaber and Jeff Sutherland as the objective threshold beyond which an increment is considered complete. Twenty years of agile discipline have shown that a deliverable without a DoD is a deliverable that doesn't end. Blair Enns, in The Win Without Pitching Manifesto, spells out the consequence for the consulting profession: a scope that isn't an act of firm diagnosis isn't a scope, it is a commercial intention. The two sources converge on the same mechanic. The DoD is the instrument that makes the fixed fee defensible, because it decouples billing from effort and ties it to outcome.

The trap: Time-boxing without an artefact, or stating an artefact without a time-box. A 40-hour phase that ends without a material deliverable leaves the client with no proof it is "done" and invites endless extension requests. The reverse: an artefact without a time-box turns the consultant into a hostage of the client's perfectionism. The three markers reinforce each other. Without one, the other two collapse. The DoD is also the angle that makes the fixed-fee structure of edition #10 on hourly billing truly operative. Without a DoD, the fixed fee is a gamble; with a DoD, it is a contract.

This week: Take your last proposal sent to a client. For each phase, write the three-marker DoD as if you had to sign today. Note every phase where you get stuck on one of the three markers. Those blockages are your structural bleeding zones, and their mapping is worth more than the proposal itself.

Sales and development: the DoD as prospect qualification test

The action: Present the three-marker DoD in the proposal conversation, not just in the written document. If the prospect explicitly resists the binary criterion ("Can we keep some flexibility?", "We'll see as we go," "Better to stay flexible"), it isn't a negotiation margin, it is a disqualification signal.

Why now: David Maister, in Managing the Professional Service Firm, insists on the proposal conversation as a moment of leadership. The consultant who concedes on the rigour of initial scoping also concedes on the authority they will hold during delivery. Bent Flyvbjerg, in the work cited above, applies the concept of reference class forecasting (originally proposed by Daniel Kahneman and Amos Tversky) to scope definition. A good DoD isn't built from an optimistic imagination of what the engagement should be. It is built by looking at what has actually been delivered on similar engagements, in volume, in lead time, in acceptance mechanics. The initial conversation is when that empirical discipline becomes visible to the client.

I personally spent years sending proposals by email and waiting for a written reply. The day I started presenting every proposal verbally before sending the document, I noticed something email had always hidden: the third marker (binary acceptance criterion) provokes a micro-hesitation in roughly one prospect in three. That hesitation, observed in the proposal conversation, foreshadows the phantom engagement that will surface weeks later.

The trap: Sending the DoD by email without an oral presentation. The prospect reads the document alone, at their own pace, and the consultant loses the chance to read the reaction. Resistance shows up in body language and in requests for clarification on a single phrase, usually the third marker. Without the conversation, those signals go unnoticed until the final invoice. The qualification grid of edition #7 on the engagement to decline gets naturally enriched by this test.

This week: For your next prospect, schedule 20 minutes to present the DoD verbally before sending the final document. Watch the reaction on the third marker. That is the one that separates the clients who want to work with you from the ones who want to occupy the grey zone.

Collaboration networks: the offer stress-test peer

The action: Designate a peer whose function is to reread your offers before they go out, with a single instruction, worded verbatim: "Point me to every sentence that can be interpreted in a way I didn't intend." Short format, 30 minutes maximum per offer, asynchronous if needed. This isn't a substantive review. It is a grey-zone detector.

The mechanism: Daniel Kahneman, in Thinking, Fast and Slow, documents the planning fallacy with two decades of behavioural research. We are systematically biased toward optimism about our own plans, and the only empirically validated antidote is the external view of a peer who doesn't share our cognitive load on the project. Mike Cohn, in User Stories Applied, shows the value of multiple perspectives on a story before its acceptance. The offer stress-test peer is the direct application of that principle to the commercial framing of consulting.

The format: A single, sharp instruction. Not "what do you think?", which produces a generalist review with no diagnostic value. Instead: "Point me to every ambiguous sentence. You don't have to propose a rewrite, just flag it." The detection-correction separation is essential. The correction stays your methodological responsibility. The peer doesn't correct. They reveal. This practice extends edition #2 on the network you neglect. Your best network isn't the one that feeds you engagements, it is the one that sees what you can no longer see in your own work.

The trap: Choosing a peer who shares exactly your practice domain. A peer who knows your craft perfectly filters ambiguities through the same grid you do, and will miss the same ones. The right peer is slightly foreign to your domain, just foreign enough to read the offer's language with a client's skepticism. A strategy consultant rereading an IT architect's offer will find grey zones that two IT architects wouldn't catch between them.

This week: Identify the candidate peer and send them your last delivered offer (with no current commercial stake) along with the single instruction. The first review can take a week or two to materialize; when it arrives, compare their inventory of grey zones with the frictions you actually lived through during delivery. The match rate will tell you whether this peer is the right structural choice.

Value creation: codify what bleeds to stop bleeding

The action: Every grey zone detected in an offer is a methodological gap to close upstream, not just a phrase to patch downstream. When you find a vague sentence that keeps coming back ("proof of concept," "ideation workshop," "audit"), the real repair is to write the canonical definition of that phase once, and to reuse the definition in every future offer.

Why it changes everything: Edition #6 on methodology as your product laid out the principle in the positive: codifying your approach builds defensible intellectual property. This edition flips the mirror. The absence of codification has a measurable cost, on every proposal. The DoD is the stress test of your methodology. If you can't write a three-marker DoD for a phase, the phase itself isn't yet a real methodology. It is an intention hiding behind familiar vocabulary. Kent Beck, in Extreme Programming Explained, states the logic in one line: "Make it work, make it right, make it fast, in that order." The codification of scope sits inside "make it right," not inside delivery.

I postponed codifying my own phases for years, telling myself every engagement was too unique to deserve a template. That was a rationalization. The day I wrote the canonical definition of my diagnostic phase, I realized I had been billing for ten years on a methodology I had never formalized.

The test: How many times has the same fuzzy phase appeared in your last three offers? If it shows up three times with no canonical definition between them, you aren't delivering a methodology. You are re-improvising every engagement under cover of a recurring vocabulary that looks like a methodology without being one. The logic connects to edition #13 on the systems that scale without you. What you codify frees you. What you don't codify taxes you on every engagement.

This week: Identify the phase that has bled most often in your last three engagements (proof of concept, audit, workshop, diagnostic, depending on your practice). Write its canonical definition in one page: typical time-box expressed in hours, standard artefact with its minimum attributes, binary acceptance criterion. That definition becomes the template for every future offer for that phase. The first draft takes two to five hours depending on prior clarity. The next twenty take fifteen minutes each.

AI: generate the marker in 45 minutes

DoD discipline is what makes AI useful for scoping. Without a codified methodology, AI invents; with a codified methodology, AI accelerates. Here is how to use it to close grey zones earlier.

The action: Use AI not to write offers faster, but to close grey zones earlier. Five use cases specific to offer scoping, accessible this week without complex infrastructure.

1. 1.Grey-zone detector in the offer draft. Submit your draft to Claude with this prompt: "Identify every sentence that can be interpreted differently by the client than by me. For each ambiguous sentence, propose the three most likely client-side interpretations." The AI objectifies what your familiarity with your own language makes invisible.

1. 1.Three-marker DoD generator. Feed the AI a vague phase description ("analytical proof of concept for the supply chain") and the client's sector context. Ask: "Generate a three-marker DoD: realistic time-box, precise deliverable artefact with page count and mandatory sections, binary client acceptance criterion. Propose three variants depending on scope." Gain: 45 minutes instead of four hours of careful drafting.

1. 1.Offer pre-mortem. Before sending, ask: "If this engagement were to drift toward a phantom engagement, where would the origin of the drift be? Identify the three contract zones most likely to be contested by the client at mid-point." The AI runs in five minutes a version of the pre-mortem Gary Klein popularized in the Harvard Business Review.

1. 1.Exploratory phase decomposer. For phases that are naturally ambiguous (audit, diagnostic, strategic exploration), ask the AI to break the phase into three or four shorter sub-phases, each with its own DoD. Result: strategic flexibility is preserved inside a firm operational frame, without giving up the frame itself.

1. 1.Translator of client expectations into binary criteria. After the discovery interview, feed the AI the transcript or your notes. Ask: "List every expectation expressed by the client, classified by level of implicitness (explicit, implied, assumed). For each one, propose a binary acceptance criterion." The DoD becomes the contractual crystallization of the discovery conversation, not a parallel construction.

Where most consultants get it wrong: AI generates DoDs that look rigorous without being so. A criterion like "the deliverable is of professional quality" checks the binary box on the surface but remains interpretable. Every AI-generated DoD has to be stress-tested against your real methodology. The practical rule: if you cannot explain why the binary criterion is binary in under 30 seconds to a skeptical peer, it isn't. AI speeds up the writing. It doesn't substitute for the methodological codification of the previous section.

The warning: The central trap is believing the AI will fill your methodological gap. It will only dress it in convincing language. As edition #14 on AI for senior consultants argued, AI amplifies an existing discipline, it doesn't create a new one. The consultant who hasn't codified their methodology will produce DoDs that look perfect and still bleed, because the AI invents what the consultant can't yet define.

This week: Take the last offer where you felt a grey zone during delivery. Run the original draft through the prompt of use case 1 (grey-zone detector). Compare the AI's list with the frictions you actually lived through. The recovery rate will tell you how many unbilled hours you could have avoided, and it will give you the exact measure of what the three-marker DoD will save you on the next offer.