Section 1
Key takeaways
• A results guarantee is a pricing lever, not a confession of doubt: adding and optimizing one grew a business's sales by 49% over the guarantee-less version . • The refund stampede is a myth. A "Results or Refund" offer lifted conversion 31% while only 2.3% of buyers ever requested a refund , and stronger guarantees can actually lower refunds by 8-15% . • Paid pilots convert at a different altitude than free trials: roughly 60% to over 90% in B2B SaaS , and 40-60% versus under 10% for free trials of enterprise software . • The asymmetry is the entire game: you trade a 2-4% refund rate for conversion lifts in the 26-49% range. The math only works against you if your delivery doesn't. • A proof-of-value offer pre-selects buyers who will convert and pre-commits you to an outcome you already know you can hit. It is a qualification engine disguised as a guarantee.
Section 2
Why founders misprice their own guarantee
Start with the fear, because it drives everything. A founder imagines offering "100% money back if you don't get the result," and immediately pictures the worst client of last year demanding a refund out of spite. One bad memory becomes the whole risk model. So the guarantee gets watered down until it's safe, and useless. The data says that fear is mispriced by something close to an order of magnitude. Conversion Rate Experts, after running guarantee tests across many businesses, put it plainly: "In our experience, the invocation rate tends to be lower than companies expect, sometimes by an order of magnitude.", Conversion Rate Experts (Karl Blanks & Ben Jesson) That single observation should reset the internal model. You are budgeting for a refund rate of, say, 20%, and the actual rate lands near 2%. Every decision downstream of the inflated number, the soft language, the short window, the absence of a guarantee at all, is a decision made against phantom risk. Look at what happens when a business stops hedging. Conversion Rate Experts documented a sequence on a real sales page (the Geeks2U case): they didn't just add a guarantee, they tested and optimized it. The final, strongest version produced 24% more conversions, and the cumulative effect of the guarantee work was a 49% improvement in sales over the original "guarantee-less" page . The guarantee wasn't decoration on top of the offer. It was a large fraction of the offer's selling power. Now pair conversion with the cost most founders are actually afraid of. River's 2026 first-party A/B-test data is useful precisely because it reports both numbers together. Their "Results or Refund" template, use it for a defined window, and if you don't hit a specific measurable outcome, you get 100% back, produced a 31% conversion lift while only 2.3% of buyers requested refunds . Thirty-one points of lift; two-and-a-half points of cost. That is not a close call. If anything, the relationship runs the opposite way from the fear. River's data shows that strong, outcome-tied guarantees decrease refund rates by 8-15% rather than inflating them. The mechanism is intuitive once you see it: a guarantee that names a real, measurable outcome attracts buyers who actually want that outcome and believe they can use what you're selling to reach it. A vague "satisfaction" guarantee attracts tire-kickers who bail on a whim. Specificity is a filter.
Section 3
What the numbers say about extending the window
There's a second fear hiding behind the first: "If I make the guarantee longer or stronger, more people will exploit it." Conversion Fanatics tested exactly this with client products. When they extended a guarantee from 90 days to a full year, the product's conversion rate doubled, while the refund rate rose only about 3 percentage points . Sit with the asymmetry. A doubling of conversion is a structural change to the economics of the business. A three-point move in refunds is a rounding error against that gain. The longer window signals confidence, removes the buyer's last "but what if it takes a while to work" objection, and the people who would have refunded mostly don't, because by the time the window matters, they've either gotten the result or moved on. This is why a proof-of-value offer is not the same thing as a discount. A discount lowers the price to overcome doubt; it leaves the buyer's risk untouched and trains them to expect lower prices. A guarantee leaves the price intact and transfers the risk to the party who can actually control the outcome, you. That transfer is the product. You are, in effect, charging for the confidence to be measured. This is the same logic that runs underneath good outcome-based packaging of a service offer; the guarantee is the part of the package that does the persuading.
Section 4
Why does a paid pilot convert so much better than a free trial?
The strongest version of a proof-of-value offer in B2B isn't a guarantee bolted onto a full engagement, it's a paid pilot. A scoped, paid first step where the buyer commits real money to a small, bounded slice of the work, with a clear success metric and a refund if you miss it. The conversion numbers here are not subtle. SaaStr founder Jason Lemkin, describing paid-pilot-to-annual-contract conversion in B2B SaaS, reports that "the conversions vary fairly widely from about 60% to over 90%" . The highest rates, he notes, come when value lands fast and in production, when the pilot isn't a sandbox demo but real work producing real results inside the buyer's environment. Compare that to free trials. Monetizely, describing enterprise software, cites figures attributed on-page to McKinsey: free trials typically convert at less than 10%, while properly structured pilot programs achieve 40-60% conversion . (That McKinsey figure is a secondary citation, it appears on Monetizely's page rather than in a McKinsey document I read directly, so treat it as directional rather than gospel.) Even at the conservative end, a paid pilot converts at four to six times the rate of a free trial. Why the gulf? Three reasons, and each one informs how you should build the offer: 1. Money creates commitment. A buyer who paid, even a small amount, has skin in the game. They assign someone to the project, they make time, they actually use the thing. Free-trial users mostly don't. The paid pilot doesn't just predict conversion; it manufactures the conditions for it. 2. A pilot produces proof, not promises. A free trial asks the buyer to imagine the value. A paid pilot delivers a measured result against a pre-agreed metric. By the time the renewal conversation happens, there is nothing left to argue about, the evidence is sitting on the table. This is the same dynamic that makes a well-run discovery process that diagnoses rather than demos so much more powerful than a generic pitch: you're closing on facts you generated together, not claims you're asking the buyer to accept. 3. The success metric pre-handles the biggest objection. The objection in a closing conversation is almost always some version of "I'm not sure this will work for us." A pilot with a defined outcome and a refund answers that objection before it's spoken. You've already agreed on what "working" looks like, and you've already put your own money behind hitting it. That last point is where the proof-of-value offer connects to the broader work of preempting objections before they surface. The best objection handling isn't a clever rebuttal in the moment, it's an offer structured so the objection never reaches escape velocity.
Section 5
A worked example: a fractional ops consultant
Abstractions are easy to nod along with and hard to act on, so make it concrete. Take a fractional operations consultant selling a $6,000/month retainer to founder-led companies in the $2M-$10M range. Her close rate from qualified call to signed retainer sits around 25%. The drop-off is almost entirely "let me think about it", buyers who like her, believe the work matters, and can't justify a six-month, $36,000 commitment to someone they've spoken to twice. The guarantee-less version of her offer asks for a leap of faith proportional to the price. So she rebuilds the first step as a proof-of-value offer: • The pilot: a 30-day operations diagnostic plus the implementation of one specific system, priced at $4,500. • The metric: a pre-agreed, measurable outcome, for this segment, "reduce founder time spent in operations by 8+ hours per week, measured by a time audit at day 30." Specific. Checkable. Not "satisfaction." • The reversal: if the 8-hour reduction isn't hit by day 30, full refund. No negotiation. What changes? The buyer's decision is no longer "do I trust this person with $36,000?" It's "is it irrational to pay $4,500 for a measured 8-hour-a-week reduction that I get back entirely if it doesn't land?" That second question answers itself for most qualified buyers, which is the entire point of engineering the offer so that no is the harder position to defend. And the conversion math follows the published pattern. If the diagnostic-to-retainer step behaves like the paid pilots Lemkin describes, she's converting pilot buyers to retainers at 60%+ rather than wrestling 25% of cold-ish prospects into a big commitment. Her refund exposure, if it tracks River's and Conversion Fanatics' data, sits in the low single digits, and crucially, she only refunds when she genuinely missed the metric, which she controls. She has converted an unmeasurable promise ("I'll make your ops better") into a measurable, low-risk transaction ("8 hours back in 30 days or your money back"), and in doing so she's pre-qualified every buyer who says yes. Notice what she did not do: drop her price. The retainer is still $36,000. She added a measured, refundable on-ramp, and the on-ramp does the selling.
Section 6
The cost side: when a proof-of-value offer is a bad idea
Intellectual honesty requires naming the failure modes, because this structure is not free and it is not universal. It punishes weak delivery. The entire model assumes you can hit the metric most of the time. If your actual outcome rate is coin-flip, a results guarantee converts your conversion lift into a refund liability, and worse, into a reputation problem. The 2-4% refund rates in the data come from operators who can deliver. If you can't yet, the honest move is to fix delivery first and use the guarantee as a forcing function on yourself, not a marketing flourish. Vague metrics are worse than no guarantee. "100% satisfaction or your money back" invites disputes you'll lose and attracts the exact buyers you don't want. The refund-lowering effect in River's data comes from outcome-tied guarantees specifically. The metric has to be objective enough that both parties already know, on day 30, whether it was hit. If you and the buyer could argue about it, it's the wrong metric. It requires you to scope tightly. A pilot that quietly balloons into the full engagement at pilot pricing destroys your margin and trains buyers to expect the world for a fraction of the cost. The bounded window is load-bearing. Define what's in, what's out, and what "done" means before the money changes hands. Cash-flow timing shifts. A smaller paid first step means smaller upfront revenue per buyer, even if lifetime value rises. For a thinly capitalized service business, that's a real consideration, not a footnote. The model trades a larger, rarer "yes" for a smaller, far more frequent one, usually a good trade, but a trade nonetheless. If you want to pressure-test whether your delivery is reliable enough to stand behind a metric, the growth diagnostic is a faster way to find the gap than learning it through refunds.
Section 7
The BGA framework: The Irrational-No Offer
Here is the structure, built to make declining feel like the unreasonable choice. Three levers, each borrowed from the kill-shot's "if it doesn't work, we refund you" confidence. Skip any one and the offer leaks. 1. Measurable Outcome, anchor the refund to one specific, pre-agreed metric. Not "satisfaction," not "results," not "value." One number, checkable by both sides, on a known date. "8 hours per week back, measured by a day-30 time audit." "Three qualified sales calls booked in the first 21 days." "Page load under 2 seconds on the priority templates." This is the lever that makes it a proof-of-value offer rather than a discount or a soft refund policy. Rule of thumb: if you and the buyer could plausibly disagree about whether you hit it, it's not specific enough, rewrite it until disagreement is impossible. 2. Bounded Window, a short, defined pilot where value lands fast and in production. The window should be long enough to produce the outcome and short enough to force focus, typically 14 to 45 days for service work. "In production" matters: Lemkin's 90%+ conversions come from pilots doing real work in the buyer's real environment, not sandbox demos . Scope it explicitly. Name what's included, what's excluded, and what triggers the measurement. The bounded window is also what protects your margin and your sanity. 3. Full Reversal, 100% money back if the metric isn't hit, no friction. Not partial, not credit, not "we'll make it right." Full cash back, and make it frictionless, the friction you remove on the way in is the friction you must also remove on the way out, or the guarantee reads as a bluff. This is the lever the data rewards most: the strongest, clearest reversals produced the doubling in Conversion Fanatics' tests and the 31% lift in River's . You can afford it because you control the metric and your real refund rate lives at 2-4%. The operating principle, engineer the asymmetry. Stack the three levers and you've created a deliberate imbalance: a refund rate in the low single digits (2-4%) traded for conversion lifts of 26-49% . The offer pre-selects buyers who will convert (they wouldn't take a metric-tied pilot unless they wanted the metric) and pre-commits you to delivering an outcome you already know you can deliver. The loss leader doesn't lose. It qualifies, it closes, and it hands the renewal conversation a result instead of a promise. One more design note: once the pilot succeeds, the transition to the full engagement should be a default, not a fresh pitch. Pre-agree the path, "if we hit the metric, the pilot rolls into the retainer at $X, no new approval needed", so the win flows straight into the contract. That hand-off is where a deliberate pilot-to-retainer transition earns its keep, turning a proven result into recurring revenue without re-selling from scratch. The full ConvertOS playbook walks the offer, the objection map, and the renewal mechanics end to end; the template pack has the pilot scope and success-metric language you can adapt directly.
Section 8
You're running The Irrational-No Offer right when…
You're running it right when a prospect's most common objection has quietly disappeared, not because you got better at rebutting "let me think about it," but because your first step no longer requires a leap of faith worth thinking about. You're running it right when your refund rate sits in the low single digits and you treat each refund as signal about delivery rather than as theft. You're running it right when your success metric is specific enough that you and the buyer would never argue about whether you hit it, when your pilot window is bounded tightly enough to protect your margin, and when a pilot win rolls into the full engagement by default instead of by re-pitch. And you're running it right when you've stopped thinking of the guarantee as a concession you make and started thinking of it as the thing you charge for, being right.