Before the first sale, what is your most honest signal?
a sale is the cleanest signal, but it can take a long time to arrive and it can be too rare to explain what is happening.
I have been thinking about a smaller event ladder: the right person sees the message, visits, understands the promise, completes the core action, returns, and then asks how to pay.
The important part is not collecting every metric. It is deciding which event would change the next experiment. A click may only say the headline worked. A completed workflow may say the problem was understood.
What was the first pre-sale signal that made you believe you were learning something real?
The most honest signal for me has been catching when a signup isn't actually a real prospect. We had one from a security researcher and another from an app review demo account, so I stopped treating raw registrations as demand and only count people who fit the audience and complete the core workflow.
still waiting for the first sale ..
For me the honest signal is not a return visit — it is someone bringing a second, messier case after the first one actually helped.
First use can be curiosity. A second photo / second listing / second client file means they put you in the workflow. If they only come back after a nudge, I treat it as marketing, not product.
From 3 months of dogfooding my own product, the signal I trust most is time-to-second-use — not “returned once” but “came back days later and asked a new question unprompted”. A second session with a different question than the first means it earned a place in the workflow. First use can be politeness; a second question is need.
I don't have a first sale yet. The honest signal I've actually seen is people saying they'll try it later and then going quiet.
Clicks and signups still happen. They don't tell me much. The silence feels more real.
The earliest signal I trust is not return alone, it is a returning user asking about continue-use constraints: how much they can do, whether limits apply, or how payment works. That question only appears after someone has already evaluated value, so it is closer to purchase intent than an unforced return, which can also come from habit. I would put that question type at the top of the ladder and treat a plain second visit as a weaker but earlier event.
Everyone here is ranking signals by strength: click < completed workflow < unforced return < brings a second case < asks the price. But there's a second axis nobody has sorted on, and it decides whether any of those events mean anything: who produced it. Every signal on that ladder is contaminated the moment the person knows you or you recruited them. A friend, a follower, someone you DM'd, someone from your own audience will return, forward it, complete the core action, and ask how to pay out of goodwill and a little social debt, not need. That is not demand, it is politeness with more steps, and it is the most flattering data you will ever collect, because it does everything your ladder asks for and predicts nothing you can scale. You cannot scale your own network. The honest pre-sale signal is the same event performed by a cold stranger: someone who arrived from a channel you do not own, has no idea who you are, and owes you nothing. When that person comes back unprompted and does the core action again, it was the product that pulled them, because there was no relationship to pull them. So I would add one tag to your event ladder before you run it: source, warm (knows you / you invited them) vs cold (stranger from an open channel), and only count the cold column. It will gut your numbers, and what is left is the only pre-sale evidence that survives contact with a market you do not personally know. The concrete question back to you: of the pre-sale signals you have already seen, how many came from people in your orbit vs total strangers? Because if the ladder is mostly built on warm traffic, you have measured how much your network likes you, and one cold stranger completing the core action unprompted is worth more than ten warm returns. It is the first evidence the thing works without you in the room.
My honest signal ladder, from most honest to least:
Someone enters a credit card — not a sale yet, but they trusted you with payment info. That's a bigger deal than any click metric.
Someone does real work inside the product — not a tutorial, not a demo flow. They enter THEIR data. For my product (property management), that means adding their actual property address and lease terms. When they do that, they've mentally committed.
Zero return logins — this was my wake-up call. I had 6 users enter credit cards in a week but literally none of them came back a second time. That told me the promise was working (they signed up) but the product wasn't delivering fast enough on that promise.
The signal I ignored too long: where the user stops. I was tracking signups as success. When I finally mapped the exact step each user got stuck on, I found 6 different friction points in a 10-step flow. The "first sale" signal was noise — the "where they quit" signal was the one that changed what I built next.
Vanity metrics I stopped tracking: page views, time on site, feature requests from people who haven't signed up.
For me, I think the strongest signal would be someone actually asking for the product after understanding what it does.
Clicks and views are useful, but they can mostly tell you that the message caught someone's attention. If someone takes the time to look through the product, asks a specific question about what's included, or asks where they can get it, that's much stronger evidence that the problem is relevant to them.
I'm currently testing a digital HR product, so this is something I'm trying to figure out myself. I'm paying more attention to those conversations than just the number of people who visit the page.
The most honest pre-sale signal I have found is voluntary continuation.
A click can mean curiosity. Completing the first workflow can mean the promise was clear. But when someone returns without being chased, brings a second real case, shares data needed to continue, introduces a teammate, or asks what it would cost to keep using the result, the experiment has created commitment.
I would define the ladder before running the test:
The useful signal depends on the hypothesis. If you are testing the message, a qualified visit may be enough. If you are testing recurring value, anything below voluntary continuation is still weak evidence.
A day-7 return fails the test you set in the post. You want the event that changes the next experiment, and a return signal arrives a week after the experiment that produced it - by then the next thing is already chosen, on nothing. Four of the seven answers here name some version of it.
Where the signal gets measured matters as much. Every platform reports on its own performance, so its dashboard always flatters it. Anything read there is the least honest number available, however good the event itself is.
What survived for me was picking the result and the date before the thing went live.
The most honest pre-sale signal I've found is unprompted return. Not "came back because I emailed them" — came back on day 3 or day 7 with no nudge from me, and did the core action again. Anyone will try a thing once out of politeness or curiosity. Almost nobody re-opens a half-finished tool a week later unless it removed real friction from their week.
The second one, which is even more useful because it comes earlier: people describing your product to someone else in their own words. A screenshot in a team Slack, a forward, a "you should look at this" reply. That tells you the promise is transferable, which is the thing that actually decides whether marketing will ever work. A click only proves the headline was interesting to one person; a forward proves it survives being retold.
I'd add one caution about your event ladder: "completes the core action" is easy to fake with onboarding hand-holding. If you walked them through it on a call, that event tells you nothing. I ended up tagging every activation as assisted vs unassisted, and only the unassisted ones counted. Cut my numbers by two thirds and made them mean something for the first time.
Also worth noting the inverse signal — people who finish the workflow, say nice things, and never come back are the most misleading data you'll get. Warm feedback with zero return is a polite no.
The hidden part is that choosing a signal forces clarity about what you actually believe. By measuring "returned to dictate" instead of "downloaded," you're betting that repeat behavior reveals something real. The event ladder isn't just about finding a signal - it's about betting your next experiment on what you chose to measure. Different signals reveal different truths about whether people actually want what you're building. Pick the wrong one and you're optimizing something that doesn't predict real customer behavior.
The strongest part is the event ladder. Separating “attention” from evidence of problem understanding is useful; a completed core workflow is a much stronger pre-sale signal than a click.
That event ladder matches how I see it. A completed workflow tells you more than a click because the person went far enough to feel the friction you're trying to remove. For DictaFlow, the useful early signal wasn't downloads. It was people coming back to dictate after their first try. If you can name the behavior people should repeat without a prompt, you've found something worth watching.