We're validating two product ideas at once — trying to figure out which problem is worth solving first

We’re driving them through separate landing pages rather than putting both ideas on the same page.

That’s intentional: we want each problem to be evaluated on its own merits, without one headline or value proposition influencing the other. Then we can compare the signals across the two — signup conversion, survey responses, willingness to pay, and the current workaround.

And I agree that the “immediate bleed” of variable income is an interesting hypothesis. That’s exactly the kind of thing we want the data to confirm or challenge rather than assume upfront.

That’s an interesting hypothesis. We agree that Cadence has a more naturally shareable output — “here’s what I can safely spend today” is easier to pass along than a cancellation workflow.

But we don’t want to assume that translates into organic growth. The real question is whether people would actually share that number, and whether the people they share it with have the same problem.

We’ll definitely keep acquisition and organic potential in the validation mix alongside trust, willingness to pay, and retention.

This is probably the best criticism of our current process.

You’re right that without a predefined kill rule, it’s very easy to keep collecting evidence until the team finds a metric that supports the idea it already likes.

We haven’t locked the final numerical thresholds yet, and I’d rather be honest about that than invent a precise number just for the sake of having one. But your framework is very close to how we’re going to structure the next phase.

For each idea, we’ll define a fatal assumption first, then compare the survivors across the same evidence chain: observed workaround → real commitment → willingness to pay → acquisition → cost to serve.

If the fatal assumption fails, the idea dies regardless of how much the team likes it.

That’s a much better decision rule than “whichever waitlist converts better.” Thanks for pushing us on this.

That’s a fair take. The immediate, tangible value of “help me stop wasting money” is definitely one of Offboard’s strongest points.

The interesting question for us is whether that value is strong enough to overcome the more episodic nature of the problem. We’ll be watching that alongside Cadence’s potential daily habit loop rather than assuming either one wins on paper.

Appreciate the perspective!

Thanks! Cadence’s “one number” approach is exactly the part we’re most interested in testing.

The potential habit loop is compelling, but we want to make sure the signal is more than “this sounds useful” — especially around whether people would actually use it to make spending decisions every day.

This is a really important dimension we hadn’t put enough weight on.

You’re right that “can we serve a free user at near-zero marginal cost?” could materially change the economics of the two ideas, especially if Offboard requires human involvement for cancellations or Cadence requires recurring financial-data costs.

We’ll model the per-user cost for both before committing engineering time. The fact that we can answer that question now, without building anything, makes it exactly the kind of assumption we should be testing at this stage.

Appreciate the push — this is probably going straight into our validation checklist.

That’s definitely one of the risks of validating two at once — keeping both alive too long.

For us, the first thing we want to kill is the idea that shows weak evidence of an urgent problem, not necessarily the one with fewer waitlist signups.

Once we have the next batch of data, we’ll use the same criteria across both: current workaround, willingness to pay, trust/permission required, and cost to serve.

And congrats on getting to the first-customer stage — that’s a much better problem to have than building something nobody wants.

That’s a really useful perspective, especially from someone who has experienced variable income firsthand.

We agree that Cadence has a potentially much stronger daily retention loop, while Offboard may be more naturally episodic. At the same time, we don’t want to assume that daily relevance automatically translates into willingness to pay.

That’s one of the things we’re trying to separate through the research: frequency of pain, strength of the current workaround, and actual willingness to pay.

And we’re glad the “one number a day” framing stood out — that’s the core promise we’re most interested in testing.

Interesting — that’s exactly the kind of behaviour we’re trying to understand better with Offboard. Thanks for sharing.

This is probably the most useful pushback we’ve received so far.

I agree that raw waitlist conversion could favour Offboard simply because the value proposition is easier to communicate, so we don’t want to treat that number as a direct proxy for willingness to pay.

We’ll add the “what do you do about this today?” question, including the effort involved in their current workaround. That should help us distinguish between a problem people recognize and one they’re actively spending time or money trying to solve.

And yes, the two products have very different operational profiles. Offboard’s complexity is largely in reliably executing cancellations across merchants, while Cadence’s challenge is earning enough trust for someone to rely on that daily number.

Also appreciate the comment on “one number a day.” That’s actually the part of Cadence we’re most interested in validating — whether that promise resonates as strongly in practice as it does on paper.

Absolutely agree. A waitlist can tell us there’s interest in the problem, but not necessarily that our proposed solution is the right one.

We’re adding questions around how people currently handle each problem and what their existing workaround looks like. That should help us understand whether we’re solving an actively managed problem or simply describing an annoyance people recognize.

The quality of the workaround is a much more interesting signal than an email address alone.

This is a great point. We’ve been treating the waitlist as the first commitment signal, but the trust boundary is probably an even more meaningful test.

We’re going to look closely at what the next real permission would be for each product and where people drop off. If someone is willing to give us their email but not trust us with the actual action/data required, that distinction is important.

Appreciate the suggestion — this gives us another useful layer to the validation.

Thanks for the vote! We’ll see what the data says once we’ve got enough signal from both.

Thanks! 🙌 Really appreciate it. We’re excited to see where the validation takes us.

That’s a really good point. We’ve been treating the waitlist signup as the stronger signal than problem recognition, but you’re right that even a signup can mean different things depending on the product.

Offboard has a much more tangible value proposition, while Cadence’s value is partly emotional, so comparing raw signup rates alone could definitely mislead us.

We’ll add a real willingness-to-pay question to the Cadence survey, and probably to both surveys for consistency. That should give us a better signal than email capture alone.

The goal is to find the strongest evidence of intent to pay, not just the most people willing to click “join.”

Thanks — and it's interesting that you're doing market research on a payment-related problem at the same time.

The "one number a day" framing keeps coming up in conversations because it removes a lot of the mental calculation people currently have to do when income isn't predictable. But we're still treating that as a hypothesis, not proof that the product should exist.

And yes, we're finding the same thing with direct user conversations: it's slower than we expected, but the quality of what you learn is dramatically better than trying to infer demand from a survey alone.

At this point, what tips the scale for us won't be simply which idea gets more interest. We're looking for the problem where people can show us a painful existing workaround, have a clear reason to solve it now, and are willing to take some kind of real action toward solving it.

Curious about your payment-related research too — especially what you're seeing once you move from stated interest to actual behavior.

Thanks! That's exactly what we're trying to avoid — spending months proving that we can build something before proving that people actually need it.

And completely agree that finding the right audience can be just as difficult as building the product. In some cases, it probably matters even more.

We'll definitely share where we land once we have enough evidence to make the decision meaningful. Appreciate you following along, and good luck with your own products too!

This is a really useful way to think about the comparison.

We agree that raw waitlist conversion isn't an apples-to-apples metric here. The "yes-cost" is fundamentally different: Offboard is easier to imagine because the value proposition is outsourcing an annoying task, while Cadence is asking someone to adopt a new way of making spending decisions.

So a smaller Cadence waitlist wouldn't necessarily mean weaker demand. It could actually mean the opposite if the people signing up have a much stronger problem and clearer intent.

We also like the idea of normalizing the acquisition side: same promotional effort, then compare cost-per-signup by channel rather than just total signups. And the "how disappointed would you be if this didn't exist?" question is a much better measure of intensity than "would you use it?"

Your point about Offboard is especially important. We're increasingly thinking that cancellation tracking isn't really the product — it's the easy part. The potential wedge is absorbing the friction between "I want to cancel this" and actually getting the cancellation completed.

That distinction gives us another assumption to validate before we build anything.

Really appreciate the framework. This is exactly the kind of pushback that makes the experiment better.

Exactly — the gap between recognition and willingness to change behavior is the signal we're most interested in.

We'd add one caveat to the Offboard side, though: we're being more careful now about treating "click-to-cancel" as a universal external forcing function. The regulatory landscape is more fragmented than our original framing suggested, so we're verifying that assumption rather than building the thesis around it.

For Cadence, you're right that the behavioral hurdle is different. The product isn't just asking someone to track something; it's asking them to trust a number enough to make a decision with it.

That's why we're moving beyond the waitlist and looking at actual behavior: what people currently do, how often they do it, and whether they're willing to take a small action to solve it.

The interesting comparison for us isn't just which waitlist grows faster — it's which problem creates the strongest evidence of commitment.

This is really interesting, especially the distinction between an annoyance and an anxiety-driven problem.

The fact that you're seeing the same "how much can I actually spend right now?" question among small business owners is a useful signal for us. We initially framed Cadence more around freelancers and independent workers, but the underlying problem may be broader than the original audience.

We also like your point about trust. If someone is going to make a daily spending decision based on one number, that number has to earn credibility over time. That's a very different product relationship from simply tracking expenses.

We're definitely not calling Cadence the winner yet, but conversations like this are exactly why we're validating both before committing to a build.

Thanks for sharing the perspective from the other side of the problem.

Appreciate the comment — although I think this may have landed on the wrong post 😄

We're currently comparing two product ideas around subscription cancellation and variable-income budgeting, so bitrate control and OCR are a little outside this particular experiment.

But we'll happily take the video advice for another project.

We agree with this, and that's actually the reason we're doing the validation before building either product.

We're not planning to split engineering across two products. The goal of running both through this early stage is to avoid committing to one problem based purely on intuition.

Once we have enough evidence, the plan is to pick one problem, one audience, and focus hard on it.

Finding the right audience is probably going to be just as important as building the product itself, so we don't want to discover that lesson after spending months on the wrong one.

This is probably the strongest counterpoint to the way we initially framed the experiment.

You're right that two equal waitlists don't necessarily mean two equal opportunities. We're also comparing two different problems with different audiences, acquisition hooks, and levels of friction, so distribution can easily become a hidden variable.

The idea of making the commitment asymmetric is interesting. An email signup tells us "this sounds interesting"; a call, workflow share, or small refundable deposit tells us "I'm willing to spend something to solve this." That's a much cleaner signal.

And the month-four question is uncomfortable, but probably useful. We don't want validation to become an elaborate way of avoiding a decision we already know we want to make.

Our current thinking is to use the waitlists to identify where interest exists, then deliberately increase the cost of the signal through conversations and small commitments. If one problem still wins after that, we'll have much more confidence that we're measuring the problem rather than our own attention.

Really appreciate the perspective from actually having lived through the two-product version of this.

This is very close to where we're taking the next step.

The "show us the last time you dealt with it" approach is particularly useful because it gives us three things at once: whether the problem actually happened, how painful the current workaround is, and what the first product workflow might need to look like.

We also like the idea of testing a small commitment rather than asking another hypothetical question. A refundable $1 deposit or sharing the current workflow is much harder to fake than "I'd definitely use this."

At this point, the winner for us won't be the list with the most people. It'll be the problem where we repeatedly see people using ugly workarounds and being genuinely frustrated by them.

Great suggestion.

This is a really good framing, especially the distinction between a problem people recognize and a problem they're motivated enough to change behavior around.

We agree that the "animated conversation" is probably more valuable than the raw waitlist number. The waitlist is useful as a first filter, but it's not going to tell us whether someone has a painful enough workaround to actually change behavior.

The YNAB comparison is also exactly the concern we have with Cadence: making the number clearer doesn't automatically mean people will trust it or build a new habit around it. That's something we need to validate before touching the product.

We're still keeping the audience fairly broad at this stage rather than locking into one vertical too early. The next round of conversations should help us see whether a particular group has a much sharper version of the problem.

And yes — we're much more interested in "show us how you deal with this today" than "would you use this?"

This is a really good point, especially the “habit vs. one-time job” distinction.

I agree that Cadence has the stronger retention thesis on paper. The interesting question for us is whether Offboard can turn a naturally episodic problem into something recurring — monitoring upcoming renewals, finding forgotten subscriptions, and periodically handling cancellations rather than being a one-and-done tool.

I also really like your point about audience reach. “Everyone with subscriptions” is definitely a much weaker acquisition thesis than a clearly defined group like 1099 workers + accountants + contractor communities.

That’s actually why we’re running the waitlists before building: we want to test not just who converts, but who has a repeatable problem and a reachable audience.

This is probably the strongest argument we've heard so far for Cadence.

The "can I safely spend this week?" question is exactly the kind of recurring decision we're interested in — especially if people are currently answering it with a spreadsheet, multiple bank balances, or just gut feeling.

And we agree that 10 people showing us their actual Sunday-night workflow would tell us far more than another 1,000 waitlist signups.

That's going into our next validation round. Rather than asking freelancers whether they'd use Cadence, we'll ask them to show us how they currently answer that question and what they hate about the process.

Really appreciate the perspective from having worked with both segments.

This is a very useful correction, and you're right that our framing was too broad.

We were treating "click-to-cancel" as a cleaner external forcing function than it actually is today. The regulatory landscape is more fragmented, and we shouldn't build the Offboard thesis around an assumption that there is a single US-wide tailwind when there isn't.

We'll correct that framing and, more importantly, verify the underlying competitor claims around Rocket Money, cancellation assistance, pricing, and bank access before treating them as part of the opportunity.

That's actually a good reminder of why we're doing this research before writing product code: the interesting part isn't getting people to agree with an idea, it's finding the assumptions that don't survive scrutiny.

Thanks for taking the time to flag it.

This is a really good distinction. We agree that the workaround is probably a much stronger signal than someone simply saying "yeah, I'd use that."

The Sunday-night question is especially useful because it gets us away from hypothetical feedback and into what people actually do today. If the answer is a spreadsheet, a bunch of bank balances, or "honestly, nothing," that's much more interesting than a survey response saying they'd like a better budgeting tool.

We'll add this to the next round of conversations and ask people to actually show us how they currently figure it out, rather than asking whether they'd use Cadence.

The "ugly workaround they still hate" is probably exactly what we should be looking for in both directions. Thanks — this gives us a much better validation question.

That's a very fair point — and honestly, that's exactly the distinction we're trying to make.

We don't see waitlist signups as validation on their own. They're a first signal of interest, not commitment. The next step is to test whether that interest survives when there's an actual commitment attached to it.

Your distinction between the two is also interesting: Cadence has a natural daily loop, so there's an ongoing behavior we can validate. Offboard has a much more immediate "this annoys me enough to fix it" moment, which could make willingness to pay easier to test.

Our goal right now is basically to find out which problem people don't just say they want solved, but are actually willing to do something about.

Appreciate the push — this is exactly the kind of feedback we wanted from posting both ideas here.

This is a really good way of framing it. We hadn't thought about the three-week waitlist retention curve as a second validation layer, but we agree that it could reveal something the initial conversion can't.

The distinction between "I have this problem" and "I'm willing to change my behavior around it" is probably the most important thing we're testing.

For Offboard, the external forcing function may reduce that gap because people already have a reason to act. Cadence is more interesting in a different way — if people are willing to change a daily financial habit just to get that one number, that's a much stronger behavioral signal.

We'll definitely track this beyond the initial signup and see whether the people who join are still engaged after a few weeks. That's likely to tell us more than the raw waitlist count alone.

That's exactly the distinction we're trying to measure.

So far, we've deliberately avoided treating "I have this problem" as validation on its own. Someone agreeing with the problem is useful, but joining a waitlist is a much stronger signal because it requires at least a small commitment.

We're especially interested in seeing the conversion gap between the two: problem recognition → waitlist signup. If one gets a lot of agreement but very few signups, that's probably telling us something important about urgency (or willingness to change).

We'll share the numbers once we have enough data to make the comparison meaningful.