Pricing Page Autopsy: The A/B Test That 2x'd Close Rates

Derek Simmons3.1% 14 months. That's where we start today.

Elena ReyesOh man, 3.1 versus 8.9% median from ChartMogul. This founder was sitting at less than half the benchmark for over a year straight.

Derek SimmonsYeah, yeah, yeah. And not doing nothing about it, just doing the wrong things. Welcome to ARR Autopsy, everybody.

Elena ReyesI'm Elena Reyes.

Derek SimmonsDerek Simmons, and today we are cutting open a very specific kind of failure. year pricing that looked fine on the surface until it quietly killed expansion for 14 months straight. So here's what we found. 280K ARR founder, three-person team, 1.5% MoM growth, two price changes based on pure gut feel, zero actual signal from either one. Classic, right? And then month 15 hits, first MRR dip they've ever seen. That's when everything shifts. And then they actually start running experiments. A tier restructure that drove an 18% sign-up lift. A billing toggle flip that pushed annual plan mix from 15% to 41%. 41% on a five-week test.

Elena ReyesWow.

Derek SimmonsThin traffic, sufficient signal, big outcome.

Elena ReyesWe also dig into the 37% churn drop after restructuring and why that solo operator segment Smith finally stopped bleeding out.

Derek SimmonsAnd then the part I love. Okay, so get this. The billing system completely fell apart.

Elena ReyesThree concurrent plan versions live, sales team is still pitching the old structure, complete chaos.

Derek SimmonsThree weeks to fix all of it. The wins broke the infrastructure.

Elena ReyesWe kind of warned them, sort of. Okay, let's Autopsy this thing. Derek Simmons, you're up.

Derek Simmons3.1%. That's it. No setup. Just 3.1%. Sit with that for a second.

Elena ReyesOkay, I'm sitting here with coffee. Should I genuinely be alarmed?

Derek SimmonsProbably. So according to ChartMogul's January 2026 study of 200 SaaS products, the median opt-in trial-to-paid rate is 8.9%. And here's the thing about that number.

Elena ReyesWait, hold on. The distribution isn't even normal, right?

Derek SimmonsNo, PulseAhead broke this down recently. The spread is bimodal. total. You're either below 5% or above 15. Almost nobody clusters at the median.

Elena ReyesSo 3.1% puts you in that danger zone, actually below it.

Derek SimmonsDeep in it. And the founder we're talking about today sat there for 14 months. Fourteen months. No way.

Elena ReyesFourteen months watching 3.1% and the pricing page just untouched?

Derek SimmonsNobody touched the pricing page.

Elena ReyesOkay, but like, were they just in build mode? Tunnel vision on feature? Features and ignoring the signal?

Derek SimmonsThat's the question, right? And Kirro published a piece on this in March that kind of reframes it. They point out that a lot of founders treat pricing as a one-time decision. Set it once, move on, the number just sits there.

Elena ReyesWhich is honestly kind of wild because that number is a message every single month.

Derek SimmonsYeah, it's telling you 14 months worth of something.

Elena ReyesExactly. Exactly.

Derek SimmonsSo here's what I want to know, and this is where it gets good. The founder had to have seen the dashboard. Every month. 3.1% staring back.

Elena ReyesSo what's the move? What did they actually do with it?

Derek SimmonsThat's the whole question. What does a founder do or not do for over a year when the number is that bad? What did the rest of the business even look like at that point? So here's where the story actually starts. 14 months of that number and the business underneath it was small, three people total, the founder, one part-time engineer, one contractor doing content.

Elena ReyesSo what's the ARR sitting at during this stretch?

Derek Simmons280K, roughly 23K MRR. Not dead, not growing.

Elena ReyesAnd month over month, where are they actually growing?

Derek SimmonsThat's the thing. Around 1.5% month over month on a good month, some months flat.

Elena ReyesAh, so there's basically zero compounding happening here.

Derek SimmonsYeah. And the acquisition side, so the channels they were running, it was SEO, some light cold outreach, a small paid budget, nothing exotic.

Elena ReyesAnd what had they already tried on the pricing side before any of this?

Derek SimmonsOh, this is where it gets good. Two price changes, one in month four, one around month nine, both based on vibes.

Elena ReyesVibes? Like literally just vibes and screenshots from competitors?

Derek Simmonscompetitors? I mean, gut feel plus screenshots. They saw a competitor pricing at $99, jumped from $89 to $89, saw another at $149, bumped to $128. No hypothesis, no test window, no measurement. Classic move.

Elena ReyesAnd I'm guessing neither one of those changes actually moved the dial.

Derek SimmonsBarely budged. And here's the thing. They couldn't even tell if it moved because they changed price or because of something else that month. No clean read, zero data hygiene.

Elena ReyesRight, and this is the pattern we see constantly. Founders treat pricing like a one-time setup, ship it, and ghost it. Then when they finally change something, it's just opinion fighting opinion instead of actual signal. Exactly. And Directive Consulting proved this out. CAC up 60% since 2020. Growth rates compressed. So you're spending more to pull the same trial user and they're still not converting. That math doesn't recover on its own.

Derek Simmonsown.

Elena ReyesSo what actually snaps them out of it? What's the moment that forces a different approach? Month 15. The founder pulls up the dashboard and MRR has actually dipped. Small dip, but a dip. First one ever. And here's what separates them. Instead of blaming the market or the product or the team, they asked the actual question, what specifically is failing? The dashboard said it clearly. Traffic fine. Trials fine.

Derek SimmonsConversion step was the leak.

Elena ReyesThat's when they finally made it testable.

Derek SimmonsFirst time and the Kirro benchmark data from ChartMogul in January 2026 report actually gives you the frame why this matters. The 8% median conversion rate isn't one number. It breaks down by model type, ACV, whether you require credit card up front, which changes everything because now you're testing against real benchmarks from real market data. Not just your gut. Right. And that's exactly the switch that flips. Pricing stops being a one-time decision and becomes a system. So the obvious next question is, what was the first thing they actually tested? tested, and honestly, what they tested first, it's probably not what you'd expect either.

Elena ReyesNo, it really isn't. We'll get into it.

Derek SimmonsSo here's the first experiment. Walk me through it. What was the actual hypothesis, and what were they changing?

Speaker 3The tier structure. Specifically, one key feature locked behind the top tier got moved down to mid-tier. The hypothesis was straightforward: that top tier was acting as a wall, not an anchor. Visitors would hit the pricing page, see the feature they actually needed sitting on the most expensive plan, and just bail.

Derek SimmonsAnd what did they actually have in terms of traffic volume to run a... run a clean test.

Speaker 3About 300 to 350 unique visitors per week to the pricing page, which honestly is where it gets a little uncomfortable.

Derek SimmonsRight, that's pretty lean. At 3% baseline conversion, you need about 2,700 visitors per variation to hit 95% confidence on even a 10% improvement. Per Optimizely's 2026 benchmarks, you're looking at four to six weeks of clean data minimum.

Speaker 3They ran five weeks right at the statistical edge, which is exact. Exactly the point. You test with the traffic you actually have.

Derek SimmonsOkay, so what do they lock in as the primary success metric before running this thing?

Speaker 3Trial signups off the pricing page, not revenue, not paid conversions, just signups.

Derek SimmonsA nodding. That's actually the smart play. You can't use LTV as your primary metric when you don't have the volume to measure it cleanly. So what actually happened?

Speaker 3Signups went up 18% in that window. They moved from roughly 3.1%. percent page to trial to 3.7.

Derek SimmonsWait, hold on. Moving one feature to a lower tier outperformed 14 months of just guessing?

Speaker 3Yeah, yeah, and here's what matters about why, according to the resources guide. Most companies use their highest value feature as top-tier bait, but what it really does is muddy the signal on what problem you actually solve.

Derek SimmonsThat makes total sense. The influence flow data backs this up. Pricing pages without a clear recommended tier convert Convert 22 percent worse. It's the entire narrative of the page that matters.

Speaker 3Exactly. Now flip that over. Second experiment is where the real leverage shows up.

Derek SimmonsThe toggle?

Speaker 3The toggle. Their pricing page defaulted to monthly, standard setup, so they A/B tested switching the default to annual.

Derek SimmonsDeadpan. Let me guess. Most people just take whatever default the page gives them and move on. Dodo Payments actually has data on this. If monthly is is the default, fewer than 20% of visitors switched to annual on their own.

Speaker 4Wow.

Derek SimmonsFlip the default to annual and 40 to 60% stay on annual. So what actually shifted in their annual plan adoption? Before the test, about 15% of new signups were taking annual. After five weeks with annual as the default, it jumped to 41%. 41% from 15. That's not optimization, that's a fundamentally different revenue model.

Elena ReyesAnd the LTV math is straightforward. Annual customers renew once a year instead of monthly. You get 12 months to prove value instead of 30. That's a completely different churn dynamic.

Derek SimmonsDid they pair it with the savings callout too, like showing the annual price broken down monthly?

Elena ReyesYes, that was part of the variation. They displayed it as dollars per month billed annually with a two-month spree label on top. Pretty standard anchoring, but combined with the default switch, it clearly moved people.

Derek SimmonsSo one test moved the feature hierarchy, one test moved the default, together you get more trials and way more annual revenue. That's a real foundation shift. And here's where I want to hand it to you because this is where it gets good and uncomfortable. Conversion up, annual mix up, but something broke on the existing customer side. Yeah, that's the question because retention is where this story either holds up or falls apart. So this is where 3.1% actually becomes less about conversion math and more about what's bleeding out the bottom.

Speaker 3Right, because you're converting more people in the door and then, here's the thing, they turn right around and leave.

Derek SimmonsExactly. So what did the churn dashboard actually show after they moved that feature down? Month one, what did the founder see?

Speaker 3Month one honestly, pure noise. At this scale, you can't read anything in month one.

Derek SimmonsIn month one, but month two, month three, this really distinct pattern just materialized-one customer type stopped leaving entirely.

Elena ReyesWhich type?

Derek SimmonsSolo operators, single seat users locked into the old entry tier-they kept smashing into this feature wall, couldn't justify the cost, and were hemorrhaging out. The moment that feature moved to mid tier, the wall disappeared.

Elena ReyesSo they weren't churning because the product was bad; they were churning because the Because the packaging was wrong.

Derek SimmonsExactly. That's packaging churn, not product churn. Those are two completely different diagnoses.

Elena ReyesAnd a 37% churn on that cohort. Where does that land? Because here's the thing about early-stage SaaS below a million ARR. You're typically burning 5-7% monthly, that's the baseline. So cutting it by more than a third? That moves the unit economics.

Derek SimmonsYeah, and slashing 37% off a segment that was If that was already your worst leak, that's not just a win, that rewires your entire LTV math.

Elena ReyesQuickly, did it also open an expansion path? Because a tier restructure that only stops churn is only doing half the job.

Derek SimmonsGood catch. Actually, yes, the solo operators who stuck around, some of them organically moved up tiers. Six months out, expansion revenue from that cohort went from basically zero to real money. Funny, the founder didn't give exact numbers,

Elena ReyesWow.

Derek Simmonsbut the tier restructure essentially built upgrade paths that never existed before.

Elena ReyesWhich is the whole point. Stop the leak, build the ladder.

Derek SimmonsI love the way you framed that. Stop the leak, build the ladder. That's the whole game.

Elena ReyesOkay, so here's the thread I want to pull. Did they ever think about layering usage-based overages on top, like consumption overage on the mid or top tier?

Derek SimmonsAnd this is the part that gets spicy. They did look at it. The business case looked airtight on paper, high-usage mid-tier customers getting significant value Mm but

Elena Reyes-hmm.

Derek Simmonspaying the same as light users.

Elena ReyesClassic problem. Your best customers are your cheapest customers per unit of value.

Derek SimmonsBut the founder actually held firm. The reason? ARR predictability. Usage-based models mean your revenue gyrates month-to-month based on customer behavior. At 280K ARR, that volatility is legitimately... intimately terrifying dying

Elena ReyesYeah, when you're that early, variable revenue on top of a thin base is a cash flow nightmare.

Derek SimmonsSo they locked in the subscription floor, killed the packaging leak, and let the tier structure drive expansion. Usage-based layers might happen down the line, but not when you're building on this foundation.

Elena ReyesExactly. Fix the leak, prove the ladder works, then you add complexity on top of something stable. Six months in, they were finally reading the dashboard and seeing numbers that made sense. Made sense.

Derek SimmonsYeah, and then it all broke.

Elena ReyesSix months in, everything's finally clicking. And then, naturally, it all falls apart. Classic.

Derek SimmonsExactly. Their homegrown billing system just completely seized up trying to handle three different plan versions simultaneously. Grandfathered accounts, a new mid-tier, annual defaults. It had zero logic for it. Support tickets absolutely exploded.

Elena ReyesHow bad?

Derek SimmonsBad enough that the founders literally in the support queue answering billing questions for for two straight weeks.

Elena ReyesUgh. Okay, so what broke the logjam?

Derek SimmonsSo they basically slammed the door on new signups to the old plan. Clean segmentation, no mixing. Grandfathered customers get grandfathered, but new customers only see the new structure. Fix the support crisis in three weeks flat.

Elena ReyesThree weeks. And the sales team?

Derek SimmonsAnd meanwhile, the sales team's still out there selling the old tier structure to prospects. Weeks of this, completely disconnected.

Elena ReyesHow does that happen?

Derek SimmonsNobody ever updated the deck, and I've stopped being shocked by that anymore.

Elena ReyesIt is never unusual.

Derek SimmonsI know, I know. So they literally locked everyone in a room for one afternoon, got everyone on the exact same page with the new pricing, problem solved.

Elena ReyesAll right, so this is the part I actually want to nail down. What does the founder watch now every single week that they weren't tracking before any of this started?

Derek SimmonsAnnual plan mix, and specifically if it dips below 35. Thirty five percent.

Elena ReyesOkay, so 35% annual mix is the canary. Below that, you start digging.

Derek SimmonsRight. That's the first metric they check every Monday morning. Not ARR, not MRR. It's annual plan mix every single week.

Elena ReyesNoted. That is a specific honest answer, and honestly, that's the whole operating lesson here. You don't manage what you didn't think to measure. Now they measure it.

Derek SimmonsThirty-five percent. That's it. Remember that number.

Elena ReyesAll right, that's a wrap on this one. And honestly, that cold open still gets me.

Derek SimmonsRight? Just 3.1%. No preamble, no context. Derek dropped it cold and made us sit with it. That's pretty brilliant from a storytelling standpoint.

Elena ReyesI mean, that's the job. But the real gut punch was the 14 months. 14 months sitting below what ChartMogul's 2026 data puts it an 8.9% median for opt.

Derek Simmonsopt-in trials. And the core lesson people need to hear? Pricing by vibes is not a strategy. Two price changes based on competitor screenshots with zero hypothesis, zero measurement, zero control group? That's not experimentation, that's just guessing.

Elena ReyesNo hypothesis, no holdout group, no measurement window. None of it.

Derek SimmonsBut the second they actually got structured, toggled the default, moved one feature, ran real Real tests? Annual plan mix jumped from 15 to 41%. That's the whole story. That's what changes the math.

Elena ReyesNumbers don't lie.

Derek SimmonsIf this episode saved you from making a bad pricing bet, pass it along to a founder who needs it. Subscribe wherever you get your podcasts. Drop a review if you can. We'll see you next time.

Elena ReyesWe'll see you next time. Thanks for being here.