Section 1
Key takeaways
• Free work doesn't devalue you. Untargeted free work does, the spec lottery where you're one of nineteen-plus competitors carries roughly a 5% win rate, versus about 33% against two . • Buyers reward demonstrated value over a pitch: 96% rate a seller's focus on the value they can deliver as influential to the purchase decision, and value-driving sales organizations win 52% of deals versus 45% for those that don't lead with value . • What separates the deal-winner from the runner-up is bringing a new idea, sales winners are 3x more likely to do this . A built deliverable wins only when it carries insight, not just free execution of the buyer's existing brief. • The fix is a three-gate test before you build anything for free: Named, not Crowd; Insight, not Labor; Slice, not Whole, plus a hard cap on how much of your time the free-sample close is allowed to consume. • A free-sample close is a sales tactic, not a business model. The day it becomes the business model, you've mortgaged your firm on a maybe.
Section 2
Why "never do spec work" is bad advice for high-ticket services
Spec work, short for speculative work, meaning a deliverable you build before you've been paid or contracted, in the hope it wins the business, gets a bad name for good reasons. The design industry spent two decades fighting "free pitches," where agencies produce real creative for a shootout and the client keeps the ideas whether or not anyone gets hired. The objection is sound. The conclusion drawn from it is too broad. Here's the problem with the blanket rule. The same RAIN Group research that found value-focus to be the top buyer priority also found that value-driving sales organizations, teams that lead with demonstrated value rather than a capabilities deck, win 52% of the deals they pursue, against 45% for organizations that don't . That seven-point gap is the entire upside of the free-sample close, expressed as a win rate. If you adopt "never do free work" as a rule, you forfeit the mechanism that produces that gap. And the gap compounds at the company level. Across high-growth versus negative-growth firms, 87% of high-growth companies run a value-based sales approach, compared with just 45% of companies in decline . Treat that figure as established rather than breaking, the underlying study dates to 2021, but the direction is consistent with everything else: the firms that grow are the firms that show up demonstrating value, not describing it. So the buyers want proof. The winners supply proof. The question was never whether to demonstrate value, it's how to do it without handing your work away. That's a positioning problem before it's a sales problem, which is why how you frame what you build matters as much as whether you build it; the StoryOS work on narrative positioning covers the upstream half of this.
Section 3
The Kill Shot is elite spec work, the data behind the angle
In the BGA vocabulary, the Kill Shot is the moment you stop selling and start showing: you hand a prospect something finished that solves a real, named piece of their problem. A teardown of their funnel. Ten qualified leads pulled to their exact criteria. A three-email sequence written in their voice, ready to send. Mechanically, that is spec work. The reason it isn't the cautionary kind comes down to a single finding. What most separates a sales winner from the runner-up isn't price, isn't relationship, isn't responsiveness. It's bringing new ideas and perspectives to the buyer. RAIN Group found sales winners are 3x more likely to bring new ideas and perspectives than second-place finishers, and that this was the number-one factor separating the two groups . A good audit is a delivery mechanism for a new idea. So is a sample lead list that reveals a segment the buyer hadn't been targeting. So is a sequence that reframes their offer. This is the whole hinge. A free deliverable that contains a new idea is the most predictive thing a deal-winner does. A free deliverable that merely executes the buyer's existing instructions contains no new idea, it's labor, and labor given away trains the buyer to expect labor for free. The exact same artifact (an "audit") can be either, depending entirely on what's inside it. That's why "should I do a free audit?" is unanswerable as asked. The artifact isn't the variable. The insight is. It also explains why interactive, personalized free deliverables outperform static ones so dramatically. Generic gated PDFs convert lukewarmly; an assessment or quiz that returns a result specific to the person, the digital cousin of a tailored audit, posts a 40.1% start-to-lead rate with a 65% start-to-finish rate on the Interact platform . The personalization is the new idea. The same logic scales down to a one-to-one Kill Shot: the more the deliverable could only have been built for this one buyer, the harder it closes.
Section 4
When the free-sample close devalues you instead
Now the other edge. The case against spec work isn't a myth, it's just specific, and the specificity is the lesson. The clearest devaluation pattern is the crowd bid. When you build free work as one of a large field of competitors, you're not running a sales tactic, you're buying a lottery ticket. The Interaction Design Foundation's analysis of on-spec work puts numbers on it: pitching against two competitors gives you roughly a one-in-three shot, but pitching against nineteen drops your odds to about 5% . At that field size you will lose nineteen times out of twenty, and you'll have built real work each time. The expected value is deeply negative, and it gets worse the better your free work is, because strong free work attracts more bidders to the same shootout. There's a second, slower failure mode that doesn't show up in win rates: the balance-sheet drain. As Joe Rinaldi, VP of Business Development at Happy Cog, puts it, "Each time agencies pre-invest in spec work, they mortgage their organizations on winning the work and the payment that will follow" . Every hour of unpaid build is borrowed against a deal that may not close. One mortgage is a calculated bet. A pipeline built on mortgages is a business that's solvent only as long as its close rate holds, and close rates don't hold forever. The third failure mode is the quiet one: anchoring. When you give away the whole deliverable, you don't just lose that project's revenue, you reset the buyer's reference price for the category. They now know an audit takes you a day and costs them nothing. The next time you quote a paid audit, you're negotiating against your own free one. This is the mechanism behind "training buyers to expect free labor", it isn't that buyers are exploitative by nature, it's that you taught them the price, and the price you taught was zero. Notice that all three failure modes share a root cause: the free work wasn't bounded, wasn't targeted, or wasn't insight. Which means they're all preventable with a test applied before you build.
Section 5
How do you tell elite spec work from free labor before you build it?
You apply three gates. If a prospect and a proposed deliverable pass all three, build it, the math is in your favor. If it fails any one, don't, or restructure until it passes. This is the framework.
Section 6
The BGA framework: The Proof-of-Value Line
A three-gate test you run before you build anything for free, plus one discipline cap so the tactic never becomes the business model. Gate 1, Named, not Crowd. Build only for one identified buyer who has confirmed two things: that the problem is real, and that there is a budget to solve it. Never build as a public bid against an open field. This single gate is the difference between the ~33% odds of a two-competitor situation and the ~5% odds of the spec lottery . The qualifying conversation that earns you "named" status is non-negotiable: if you haven't confirmed the problem and the budget out loud, you don't have a named buyer, you have a hopeful guess. Concretely, for a fractional-CMO offer: you do not build a free strategy deck for every inbound. You build one only after a discovery call where the prospect has said, in some form, "yes, lead-gen is broken, and yes, we have spend allocated to fix it." If you can't get that confirmation, the gate is closed, and the upstream discipline that produces clean confirmations lives in the LeadOS work on qualification and disqualification. Gate 2, Insight, not Labor. The deliverable must contain a new idea the buyer did not already have. This is the 3x-winner behavior, made operational . Before you build, write down the one thing this deliverable will tell them that they don't currently know. If you can't name it, you're about to do free labor, not a Kill Shot, stop. A worked example: a paid-ads agency offered two versions of the same "free audit." Version A listed what was wrong with the account, high CPA, weak creative, broken tracking. That's labor: the client already half-knew all of it, and handing it over for free just gave them a punch list to take to a cheaper shop. Version B contained one finding the client couldn't have produced themselves, "your best-converting segment is being suppressed by your own broad-match terms, and here's the reallocation that recovers it." Same format, same hours. Version A devalued the agency. Version B closed a retainer, because it proved a way of seeing the account the buyer couldn't get anywhere else. Gate 3, Slice, not Whole. Ship a complete-but-bounded proof: one audit, ten leads, one sequence, never the full engagement. The slice must be finished enough to be credible on its own and small enough that it implies the engine behind it without being the deliverable itself. Ten qualified leads prove you can build a list; they are not the list. One sequence proves you can write the funnel; it is not the funnel. The boundary is what protects your pricing, the buyer experiences a complete result, then realizes the complete result is one-twentieth of what the engagement produces. Get this wrong in the generous direction and you've given away the job. Get it wrong in the stingy direction, a half-finished slice, and you've proven nothing. The craft is in the cut. The discipline cap. Bolt one rule on top of all three gates: cap the total time the free-sample close is allowed to consume. The Interaction Design Foundation's working heuristic is to keep speculative work under roughly 20% of your time . The cap is what keeps the tactic from metastasizing into the business model, the exact failure Rinaldi describes, where the firm runs on mortgages . When you're inside the cap, every free build is a sales investment you can afford to lose. When you blow past it, you've quietly become a shop that works for free and occasionally gets paid. A useful way to hold the cap is to track free-build hours the way you track ad spend: a fixed monthly budget, with a target return. If you're spending 20% of capacity on proof-of-value builds and closing better than one in three named opportunities, the motion is paying for itself. If you're past 20% and closing less than that, either your Gate 1 qualification is too loose or your Gate 2 insight is too thin, and the systems for tracking and automating that follow-up are where the AutomateOS work on pipeline systems earns its keep. If you want to pressure-test your own motion against these gates, the ConvertOS playbook walks the full closing system the Proof-of-Value Line sits inside, and the free growth diagnostic will tell you in about ten minutes whether your free work is currently an investment or a leak.
Section 7
A worked scenario: the same audit, two outcomes
Take a mid-market B2B service firm, call it a $2M-a-year demand-gen agency, deciding whether to "do a free audit" for two prospects in the same week. Prospect One came in through a public RFP. Six agencies were invited; all six were asked to submit a sample strategy. The agency's instinct was to build their best work and win on quality. Run it through the gates: Gate 1 fails immediately, this is a crowd, not a named buyer, and at six-plus competitors the odds sit far below the two-competitor 33% line and head toward the spec-lottery 5% . The correct move isn't a better deck. It's to decline the spec portion and offer instead a single, sharp insight on a call, proof of thinking without proof-of-build. They did. They didn't win the RFP. They also didn't spend forty hours losing it, which is the point. Prospect Two came from a referral and a discovery call where the founder confirmed the problem (pipeline had flatlined) and the budget (a retainer was already approved, pending vendor choice). Gate 1 passes: named, problem confirmed, budget confirmed. The agency then asked the Gate 2 question, what's the one thing we can show them they don't know? They found it: the prospect's inbound was strong but their lead-to-meeting handoff was leaking most of it before sales ever saw it. That's a new idea, not a punch list. Gate 3: they built a bounded slice, a teardown of one funnel stage plus a sample rebuilt handoff sequence, finished and usable, but explicitly one stage of a five-stage system. The slice closed a retainer larger than the one originally scoped, because the insight reframed the size of the problem. Same artifact category, "a free audit", in the same week. One was free labor avoided. One was an elite Kill Shot that closed. The variable was never the audit. It was the three gates.
Section 8
You're running The Proof-of-Value Line right when…
You're running it right when you can look at any free deliverable on your calendar and answer three questions without hesitating: which one named buyer is this for, and have they confirmed the problem and the budget? What is the single new idea inside it that they couldn't have produced themselves? And what is the boundary that makes this a complete slice rather than the whole job? If all three answers are crisp, build it, the win-rate math is on your side. You're also running it right when your total free-build time sits inside a hard cap you actually track, when declining a crowd bid feels like discipline rather than fear, and when no prospect has ever received the full deliverable for free. The day you can't name the buyer, can't name the idea, or can't find the boundary, you're not closing, you're training someone to expect your labor for nothing, and you'll feel it in your next quote.