Growth has a paper trail -- capture it!
A couple of weeks ago we posted here about Powsight — a system that builds a real profile of your SaaS and tells you where the potential leverage is, based on hard data like Stripe and Google Analytics.
The post got great responses, but one idea stuck with us longer than the rest.
A founder pointed out something we hadn't fully sat with: everything we'd built assumed you already had customers and traffic to point the system at. His situation was earlier and messier than that — pre-revenue, a handful of channels (organic content, community, affiliates), and no clean signal on which one was actually working versus which one just felt productive.
What he said he wanted was simple: the same discipline our feedback loop gives people with data, just done by hand. Write down "I tried X, here's what happened, here's whether I'd do it again." Not because a model needs the data yet — just because that habit of reflection is valuable on its own, long before there's anything to measure.
We talked about it back and forth in the thread. He mentioned his actual workflow was a plain text file — one dated entry per channel, what he tried, what happened, whether he'd repeat it. It worked, but there was no way to cross-reference it against anything. If he wanted to know whether the weeks he posted more actually moved downloads, he had to do that comparison by hand.
That framing made the shape of the feature obvious. So we told him: we'd build a version of it and call it Journaling.
We shipped it.
⭐ Why this matters at every stage, not just the ones with data
Journaling exists because a huge part of what happens in a SaaS business never shows up in a metrics dashboard. You try a positioning angle that flops. You have a conversation with a user that changes your roadmap. You decide not to build something, and that decision matters just as much as the ones you ship. None of that lives in Stripe or GA — but all of it shapes what happens next.
So now, whatever stage your product is in, your journal entries feed directly into your Insights and Digital Profile — the living dataset we build about your business. The system doesn't just know what happened to your metrics; it starts to understand more about potential whys.
That's especially powerful for pre-revenue founders. Before there's a funnel to instrument, journaling lets you start generating meaningful datapoints from day one — just by documenting the journey as you live it. No customers required to start receiving powerful insights.
🚀 It compounds
This is the part we're genuinely excited about, because it doesn't stop at journaling itself:
• Digital Profile gets richer, faster — not just from metrics, but from your own account of what you tried and why.
• Insights get sharper context. We also built a new insight type, Progress, specifically to read your weekly journal entries against your financials and analytics — so it can tell you when to push harder, and when to stop and focus on other things with higher potential impact.
• Metrics stop being numbers in a vacuum. You can finally see data in the context of your actions.
None of these pieces are new on their own. What's new is that Journaling gives all of them something to build on, even before a product has a single paying customer.
❓What's still missing?
We built Journaling because one conversation showed us exactly what was needed — that's honestly one of the best parts of building in public. So we want to keep doing that.
If you're a founder — pre-revenue, early traction, further along, doesn't matter — what are the things/tools you need to accelerate your SaaS growth?
We'd genuinely like to know what we can do to help.
Also, lots of thanks to @Exemplar_Frameworks for this idea 🫡
Glad the conversation turned into something shipped rather than just a nice idea in a thread — that's rarer than it should be. Curious how you'll handle the point where a founder's journal entries start contradicting what the metrics say once there's actual data to compare against.
and a pretty useful one, thanks for your inputs 👍
to answer your question: we'll flag it to the user as that's the main goal -- find out what works and what doesn't
hard data is king, user inputs are the context -- so surfacing contradictions in order to learn faster is the most useful outcome of this feature
That's the right call — treating the mismatch as signal rather than noise. Be interesting to see whether the contradictions cluster around specific channels or specific founders' blind spots once you've got a few months of data to look back over.
will keep you posted :)
This is exactly what I look for in a pitch and rarely find: not the metrics, but the reasoning trail behind them. Founders who can show me "I tried X, it didn't work, here's why I killed it" get funded faster than founders who only show a clean growth chart, because the trail proves they're running real experiments and not just getting lucky once. Make sure those journal entries are exportable, that's exactly the kind of thing that belongs in a data room later.
maybe there's also some value in aggregating and analyzing entries across the founders multiple products (because we do allow multiple products at once for the same founder), so we can identify patterns across all of your saas apps and thus identify bias or really successful approaches (of course by that time you'd already be aware of them)
but definitely not a focus for now, just a nice to have maybe?!
thanks for the reply and suggestion 👍
Journaling is a smart ship, and the founder who asked for it was pointing at something real. But there's a tension in how you're framing the value that's worth naming, because the part you're most excited about is the weakest part, and the part you're underselling is the strongest.
You're selling it as "datapoints from day one that feed Insights and compound." Careful there. Journal entries aren't the same species of data as Stripe or GA. Those are behavior, what actually happened. A journal entry is narrative, a founder's own account of why something worked, and that account is exactly the signal most likely to be wrong, memory encodes what you intended as what occurred. Feed confident-but-wrong self-reports into an automated correlation engine and you don't get sharper insights, you get a model fluently agreeing with a founder's bias. Garbage-in compounds too.
The genuinely strong thing is the part you mention almost in passing: the discipline. "I tried X, here's what happened, would I repeat it" is valuable to the human writing it whether or not a model ever reads it, because the reflection itself catches the self-deception. That's the sell. Not "it makes our insights smarter," but "it makes you honest with yourself before there's data to keep you honest."
So the question I'd sit with: does Progress read the journal as ground truth, or as a claim to test against the metrics? Because one amplifies bias and the other corrects it, and that design choice is the whole ballgame.
great point 👍
this has been risen in various forms in the other comments as well so it seems I didn't communicate the mechanism/value well enough
there's one distinction that has to be made here -- there's:
that's why
so to answer your question:
100% claim, no truth -- the journal acts as additional context around which outcomes are being judged against (ex: you're saying you've posted 100 times on reddit but we see no traffic from it, then it doesn't work)
what's probably key here as you said is the accelerated learning loop -> you believe something (do 1 indiehacker post/day), 1 week later data proves otherwise so we flag it -> you change the approach, angle or drop the action altogether as it had no results
I'm going to create a post to explain this loop better
does this make more sense? I mean, in terms of communication (because you've already got the idea)
also, thank you for the thorough analysis -- very valuable feedback
Yes, the mechanism is now clearly right, "100% claim, tested against data" is exactly the trustworthy version, and your Reddit example (you say 100 posts, we see no traffic, so it didn't work) is the whole product in one sentence. So this is purely a communication gap now, not a design one, and it's worth fixing because the same three of us misread it the same way for the same reason.
The trigger is your verb choice. "Datapoints that feed Insights and compound" makes the journal sound like input, data being poured into the model, which is exactly what sets off the garbage-in alarm. But you just told me the journal isn't fed in, it's put on trial. Those are opposite motions, and your copy uses the word for the wrong one. Switch the verb and the worry evaporates: not "your notes enrich our dataset," but "tell us what you did, we'll tell you whether the data backs it up."
The deeper tell: your writeup describes the plumbing (journal → Insights → Digital Profile → compounds) when the exciting thing is the verdict (you believed X, the data says no, change course). Nobody gets excited about a richer dataset. Everybody gets excited about the tool that catches them fooling themselves. Lead with the accountability, not the architecture, that accelerated-learning loop you mentioned almost in passing is the headline, not a footnote.
For the explainer post you're planning: what's the sharpest real example where the data contradicted a founder's journal entry? Open with that, the mechanism explains itself the moment someone sees a belief get falsified.
this is great, thanks 🙇
I'll wait for data on this and then I'll write a post with this exact framing
this has been very helpful, thanks a lot -- I'm super curious how this angle will affect traction so I'll note it down in the journal.. haha
if the journal flags this thread as the thing that moved traction, that's the loop working exactly as designed. Good luck with it, looking forward to the writeup.
The self-report vs. hard-data tension in this thread is the real question, and I don't think it's fully resolved by either side. Pure self-report drifts toward "I did X and got lucky, so I'll call X causal" — MananShah's point. Pure attribution-via-onboarding ("where did you hear about us") has the opposite problem: it only catches the founder as the last touch someone remembers, which quietly erases every low-traffic, high-context channel (a good reddit reply, a thoughtful IH comment) that planted the idea three weeks before someone actually signed up.
The thing that might actually resolve it: don't ask "did this work" at logging time — ask it retroactively, in batches. Once a founder has 15-20 entries and a few actual outcomes (signups, replies that went somewhere, a DM), have them look back and mark which prior entries they'd now credit, rather than guessing in the moment. Hindsight-tagging is messier data than a clean funnel, but it's honest about the fact that most of what works for pre-revenue founders doesn't have a clean single-touch attribution path — and it avoids the "I posted and got 5 users, therefore posting works" trap MananShah flagged, because the founder is judging with more information, not less.
you're right, hindsight would be a much better fit here
but lm ask you this: why have the founder guess and share instead of analyzing the data ourselves and pointing out "this is what worked based on your data and activity -- do you agree?"
it's much cleaner, less spammy and it's obviously rooted in data not feelings -- I believe this is the way to accelerate a saas growth, by accelerating the founder's understanding of what worked or didn't via insights so the founder can make better decisions, reorient quicker if the case or double down on what worked
the point is not to spend time on generalities, but learn from specific insights about your business context, your audience, your customer behaviours, churn, etc -- if you accelerate this, you increase the rate of success simply because of the founder's better understanding of his own business context
The dated plain-text log is probably more useful than a dashboard right now because it preserves the reasoning behind each decision. Make every entry answer one question: What did I expect to happen? Later, you won't just compare activity with results. You'll see which assumptions were wrong and avoid repeating work that only felt productive.
this is awesome -- we can extend what you're saying to pro-actively flagging these assumptions in time (let's say 1 month of data) so we can stop you from investing too much time on something that was invalidated by data
ex: "you expected X but it didn't happen and there's no signal it would happen with the current approach, so rethink your hypothesis or your tactics"
That connection between weekly actions and output numbers is huge — tracking data without knowing the contextual story behind it usually leads to misinterpreting metrics.
As a video editor and visual communicator working with founders, here’s what I see missing that could really accelerate growth:
Visual Context & Automated Micro-Stories: Metrics tell us what happened, and the journal captures why. But for public updates, social proof, or investor pulse checks, founders need a way to turn those weekly "Progress" insights into digestible visual assets (e.g., auto-generated progress cards, mini data-story visual snippets, or timeline graphic exports).
Pre-Revenue Momentum Indicators: Pre-revenue founders often measure progress purely by finished tasks, which can get demoralizing. A system that tracks "Narrative Velocity" — mapping strategic outputs (like content released, outreach done, or features shipped) directly against audience engagement shifts — would help founders stay focused on high-potential leverage points before money comes in.
Brilliant integration overall — connecting human context to hard financial data is a game-changer! 👏
thanks :)
good points -- we're working on it 👍
the exact founder situation you're describing is where I am right now, pre-revenue, splitting time across reddit, indie hackers, and x, with genuinely no clean way to know which one's actually contributing versus which one just feels productive because I enjoy it more. ran into this literally yesterday trying to run my own account through an analytics tool and getting "sparse profile, can't read this" back because most of my activity is replies, not posts, which the tool didn't weight the same way
answering your actual question: what I need most isn't more data collection, it's something that helps me compare across channels that don't share a common unit. a reddit comment, an ih thread, and an x reply don't produce the same kind of signal, so "which channel is working" is a genuinely hard question to answer by feel alone. if Journaling eventually let me log "spent today on ih, here's what came of it" next to "spent today on reddit, here's what came of it" and surfaced which one correlates with actual downstream outcomes (waitlist signups, real conversations), that's the comparison I don't currently have any way to make
the "no customers required to start receiving insights" framing is the right instinct for pre-revenue people specifically, most tools assume you already have the thing they're trying to help you optimize
hmm, very good point
we're actually doing that (analytics do show referrals) but not at the granular level you suggest (this post brought X visitors, this reddit comment brought Y visitors)
what if I'd let you add URLs (posts, reddit comments, etc) when logging to your Journal?
that means we can at least track impressions/views over time and cross-reference with traffic
that would help, but I think there's a gap between "this URL got X views" and "this URL correlates with actual downstream outcomes." views on a reddit comment or an ih thread are usually pretty low regardless of how good the exchange was, the value isn't really in traffic, it's in whether that specific interaction turned into something, a reply that led somewhere, a profile click, eventually a signup. URL logging tells you reach, not whether the channel is actually working
maybe the honest answer is you can't fully solve this with passive tracking alone, since a lot of what happens after a good reddit reply (someone reads it, checks your profile, quietly signs up days later) doesn't leave a clean attributable trail no matter what you log. might be worth pairing URL logging with a simple self-report at the time of logging, "did anything come of this, yes/no/not yet," even if it's subjective, since for low-traffic high-context channels like community replies, the founder's own gut read might be more honest data than any traffic number.
yeah, I don't think so -- if no traffic comes from reddit and you're logging "I posted on reddit and now I have 5 new users" then that's just coincidental and you're not learning what really worked -- you're going to repeat that over and over hoping for the same result when that's not going to be there
the causality can be established in the onboarding flow -- ask the users where they heard about you, and if we find Journal entries with reddit posts, we can definitely say they had an impact and drove sales 👍
I'm going to think more about this so I find a solution because it's quite a nice problem to solve
fair, and honestly a better critique of my own idea than I gave it credit for while writing it. self-reported gut-read data really is just confirmation bias with extra steps if there's no ground truth to check it against
the onboarding-attribution approach is the actual fix, cross-referencing "user says they heard about us via reddit" against "founder logged a reddit post around that time" gives you real correlation instead of either passive traffic numbers or unverified gut feeling. that's a genuinely different, better mechanism than either of us proposed
good problem to be stuck on, most attribution tools solve this by assuming you already have paid ad click IDs to work with, solving it for organic/community channels specifically is the harder and more interesting version
I think so too :)
didn't know that, thanks for sharing
glad it was useful, genuinely one of the better threads I've been part of this week
The version I would actually keep using has three enforced fields per entry: what I tried, what I observed afterward, and whether I would repeat it. That is the loop from the original founder's request, and it stays useful because the third field forces a decision instead of a diary. The risk I would watch for is turning journaling into a general notes box; once entries stop carrying that repeat-or-not verdict, the insights layer loses the connection between actions and outcomes.
That makes sense — I was treating “repeat or not” as a forcing function for the founder, not as another signal for the model. If Insights can infer the outcome from the data, the lighter structure is probably better. I’d still keep a small “what I’d do next” note for cases where the result is ambiguous — a post gets clicks, but from the wrong audience, for example. That keeps the decision visible without pretending the journal entry proves causality.
it makes sense, but why not let the insights tell you if it really worked and you should repeat or not?
my concern is "whether I would repeat it" doesn't add much value, because it's obvious that if it worked you're going to repeat it, log it, and the system will know
so to me the most valuable structure is "what I tried, what I observed afterward"
I agree that building it in journaling would prompt the founder to focus on action-results so I'm going to add it in 👍 thanks for the suggestion
the thing is, insights always look at results. Sure, sharing your opinion in Journaling is relevant as it's additional datapoints we could work with, but we're always looking at hard data first (financials, analytics) so we see if your actions have had any impact.
The interesting part is that the journal isn't just a record of what happened. It gives the metrics context by capturing what you actually tried, which should make the later insights much more useful.
exactly :)
That’s the part I found most interesting too. I’d be curious what you learn once people have enough history for those patterns to become visible.
The instinct behind Journaling is right, and I think it generalizes past growth specifically. Most founders already have some written trail (Slack, a doc, commit messages). It's just scattered and un-cross-referenced, which ends up being nearly the same problem as having no trail at all when you actually need to look something up.
What I'd want from a tool like this: not just "log it," but make the log queryable against a decision, not just a date. Two years from now the question is rarely "what did I write on March 3rd," it's "why did we pick X over Y," and that needs the entry tagged to a decision, not just timestamped to a day.
Small thing worth testing: does the habit survive week 3? Journaling tools live or die on whether the second entry happens without a prompt. Curious what retention on the feature looks like a month in.
makes sense, also, fairly easy to do
I'll post updates about this, if the clients understand the value I believe they will do it
Would genuinely like to see that when you post it. Retention data on a self-reported habit is rare to find written up honestly, most of what's out there is survivorship bias from teams that only mention the win. Good luck with the rollout.
thanks :) will keep you posted
I’m at this exact stage too. I’d keep the journal brutally simple: channel, experiment, expected signal, actual signal, and next decision. The “next decision” matters most—otherwise it slowly becomes a diary. I’d also separate replies, signups, repeat usage, and payments, because they’re very different levels of evidence.
makes sense, for now it's just a free text input but I'll probably add templates for whatever you want to log -- any log is important especially if you want to track your journey from 0 to 1
this we already do 💪
I’d keep the first template tiny: what I tried, what I expected, what happened, and what I’ll do next. Anything longer and I’d probably stop logging it myself.
yep, makes sense :) I'll make the templates optional and keep free text otherwise so it's not too demanding on the inputs
The part that resonates with me is having a record of the decisions behind the numbers. It’s easy to look back at a traffic spike and convince yourself you know why it happened, when in reality you might have changed three things that week.
Even a really simple log like “what I tried + what happened” would make those patterns much easier to spot later. Especially when you’re running several acquisition channels at once, having the context next to the metrics seems way more useful than another dashboard full of numbers.
yep, exactly -- adding urls next to the log may help with cross-referencing content with outcomes
The dated log is the one thing analytics can't reconstruct — tools show what happened, never why you did it. We keep the same habit, and it's the only reason old decisions still make sense once data arrives. The log turns numbers into questions worth asking. We're building analytics around that: https://amami.dev