SEO 3x'd my traffic. Signups dropped to zero. Here's what actually drives first customers in vertical SaaS.

Before you file this under intent mismatch, one thing worth ruling out: is the page actually working for the traffic that showed up?

I had almost this exact shape — traffic climbing, signups flat — and spent weeks on the assumption that the audience was wrong. It wasn't. I put session replay on it and watched people land on mobile, hit a hero with no visible CTA above the fold, scroll a bit, and leave. Separately I found a chat widget that was broken, and people were rage-clicking it over and over before bouncing. Neither of those shows up in analytics. Both look exactly like "wrong audience" in a funnel report.

The reason I'd check it in your case specifically: SEO traffic skews far more mobile than community traffic does. Your Aug 1-7 cohort came from BetaList/IH/PeerPush — people on a laptop, deliberately browsing tools. Your Aug 8-22 cohort came from Google, on phones, mid-task. If your mobile page has a problem, the first cohort would mostly never have hit it and the second one would hit nothing else.

Worth splitting those 8K impressions by device before you lock in the conclusion. If mobile bounce is way out of line with desktop, it's mechanical, not motivational.

Your split between "SEO brings readers, discovery platforms bring triers" matches what I'm seeing, except my version has an extra wall in front of it.

I have the inverse of your problem and it points at the same conclusion, so it might be worth adding here.

We sell through a marketplace — the Shopify App Store — so our discovery is gated by someone else's search algorithm rather than by Google. In three weeks our listing got 96 sessions total. About four a day. In that same window we got two paying customers, both from our own network. So: you have 8,000 daily impressions and no signups, I have paying customers and no impressions. Neither of us has a traffic problem in the direction we assumed.

The thing I'd add to your framework is a step before it. When I finally audited where we actually rank, it turned out we're on page one for exactly one search term — a Spanish-language one, position 13 out of 772 competing apps — and nowhere on the English equivalent, which has 1,115 apps and most of the real demand. The marketplace wasn't ignoring us. It was showing us in the one query almost nobody runs. "No signups" and "no impressions" are different diagnoses, and I spent three weeks treating mine as the first when it was the second. Before optimizing the funnel, it's worth confirming which half of it is actually broken.

The part of your post I'd like to press on: BetaList, LaunchPact and PeerPush all assume you own the signup. Has anyone here used them for a product where the install happens inside someone else's marketplace, and the thing you're really trying to unlock is that marketplace's own search ranking?

The uncomfortable version of your finding, for us: the only thing that reliably converts at this stage is emailing one merchant at a time and offering to set the thing up for them personally. It doesn't scale and I'm not sure it's supposed to yet.

This matches what I'm seeing almost embarrassingly well. I'm in the middle of a launch push for a small browser game, and the only visitors with real intent so far came from directories and communities where people browse to try new stuff. Search impressions are climbing at the same time, but those people came for an answer, not a product. The 'each engine has a different KPI' framing is exactly what I needed - saving this for the next time traffic looks great and the signup counter doesn't move.

The “traffic ≠ tryers” distinction is probably the most important part here.

3x traffic sounds like a win until you look at the intent behind those visits. Someone searching “California rent increase rules” has a completely different job-to-be-done from someone browsing a product-discovery platform looking for a tool to try.

I’d be especially interested in tracking the funnel by intent/source, not just channel: landing page → engaged session → signup → activated user. That would make it much easier to see which sources are actually producing customers rather than just sessions.

Pre-PMF, I think that distinction can save founders a lot of time chasing impressive-looking traffic numbers.

This matches what I just saw. I was hunting a cheap digital product in a trade niche. The people googling those keywords want a free checklist or a contractor, not a $29 PDF. Search leads in that space are $70-120.

Your Aug 1-7 signups from IH and BetaList vs zero from 3x SEO is the whole lesson. Traffic is not buyers.

This matches what I keep seeing. Search traffic wants the answer on the page. They got it, they left. The people who actually tried the product showed up because they were already browsing for tools.

Treating those as the same number is how you waste two weeks staring at impressions. Readers and tryers are different jobs.

I'd split this one more way: keep the pSEO pages, but stop judging them by signup rate until you've added one tiny conversion bridge. For a landlord reading a rules page, a useful bridge might be a rent-increase checklist or calculator that saves their state, then shows the SaaS only after they've done the job. That tells you whether the problem is intent, CTA, or the offer itself.

This split makes sense to me. I saw the same pattern while building DictaFlow. Someone coming through search may only want an answer, while someone from a community or discovery channel is often already willing to try a tool. It helps to treat those as different jobs instead of calling one channel good and the other bad. Are you tracking whether discovery users return after their first week, or just their initial signup?

Bon diagnostic sur reader vs. tryer. Un angle que je n'ai pas encore vu ici : ces mêmes pages pSEO qui "répondent et s'arrêtent" ont probablement le même problème face aux IA génératives qu'à Google.

Une IA qui répond à "how much does a property manager cost" va citer une source qui répond clairement à cette question précise — pas forcément la tienne, et même si c'est la tienne, elle cite la réponse, pas le produit. L'IA ne "visite" jamais ta page comme un humain qui pourrait remarquer un CTA au scroll suivant.

Le test que je ferais avant de conclure que le SEO ne sert à rien pour toi : prends 5 questions d'acheteurs réels (pas informationnelles — "best property management software for landlords", "alternative to [concurrent]") et regarde si ChatGPT/Perplexity te citent dessus. Si tes pages actuelles sont optimisées pour des questions génériques plutôt que des questions d'achat, ni Google ni les IA ne vont t'amener de tryers — ce n'est pas le canal SEO qui est mort, c'est le ciblage de mots-clés qui vise les mauvaises questions.

Ton framework reste juste, mais je le nuancerais : SEO = readers si tu cibles des requêtes informationnelles. Cible des requêtes commerciales et la même chaîne (Google + IA) devient un canal de tryers, pas seulement de lecteurs.

Great insight. Traffic quality matters much more than traffic volume, especially before PMF. 🎯

This matches what we see at SocialPost.ai: informational SEO traffic and buyer traffic are two different populations, not two stages of one funnel. The fix isn't more content, it's targeting bottom-of-funnel queries like "best property management software for landlords" instead of "California rent increase rules." I'd split your Search Console queries into informational vs. commercial intent before concluding discovery platforms are the only lever that works.

Your framework is right, which is rare, so let me push it one level past where you stopped. You're treating the three engines as parallel lanes with different KPIs. They're actually a sequence, and seeing that answers your own "do platforms only work once" question: yes, mostly. Discovery platforms are a finite, one-time burst per listing, the audience browses through once and churns. They get you your first ten. They physically can't be your engine, because the pool is small and doesn't refill. They're ignition, not the motor.

Which leaves SEO as the only one of your three that compounds. But you've written it off as "readers, wrong intent," and that's the miss. The person googling "how much does a property manager cost" or "California rent increase rules" isn't the wrong person, they're your exact ICP, a landlord, standing in the exact problem your product solves. They're just in research mode, not buy mode. You answer their question and let them leave. That's not a wrong-intent problem, it's a missing-bridge problem.

The fix isn't more traffic, it's an on-ramp. The article answering "rent increase rules" shouldn't dead-end at the answer, it should end at "or track this automatically across all your units, here's how." Same reader, one step toward tryer. Right now your SEO and your product are two separate buildings with no door between them.

So the reframe of your question: discovery got you to 6, but it won't get you to 60. The needle-mover isn't finding more tryers on more platforms, it's converting the readers you already have. What does the path look like today from someone landing on a "rent increase rules" page to actually seeing your product, is there one, or does the page just answer and stop?

The confound worth naming: you did not run traffic against no traffic, you ran posting against not posting. In week one you were present in places where people browse specifically to try new tools. In the fortnight after, you were absent everywhere and the only thing left was search traffic. Six signups versus zero is consistent with your conclusion, and it is equally consistent with "showing up is what converts, at any volume".

That said I think your conclusion is probably right, for a reason your numbers hint at. Programmatic pages get found by people with a question, not people with a problem they have decided to pay to fix. Someone searching a term your generator targeted wants an answer and then wants to leave. Someone on BetaList is there in a shopping posture. Same person, different intent, wildly different conversion.

Where I would push back on the fatalism: intent is a property of the page as much as the visitor. A page that ranks for a question and then offers the tool as the answer to that specific question can convert. A programmatic page that ranks for volume and then offers a generic signup cannot.

Out of 8K impressions, how many actually landed on a page and what did they do next? Impressions and sessions are very different animals, and the gap might be the whole story.

I’m seeing something similar. I have real users using the product, but none have reached the point of subscribing yet. So my question is the gap between “this is useful” and “this is worth paying for.” My guess is that usage alone isn’t enough—the product has to solve a painful, recurring problem. Did any of your six signups become repeat users?

This is a useful counter-story to the usual "just do SEO" advice. Was the traffic that came in from SEO fundamentally the wrong audience (wrong intent), or was it the right audience but something else in the funnel (onboarding, pricing, trust signals) that broke down before signup?

The paid version of your experiment says the same thing, which surprised me. We bought search ads on keywords that name the category we sell, so the highest intent money can buy, 42 clicks over the last two weeks. Signups: zero. Same result as your SEO fortnight.

What BetaList and IH gave you that neither search channel does is mode, not intent. Someone browsing a discovery platform is already in "look at new tools" mode. Someone searching, paid or organic, is in "solve this now" mode, and adopting a new SaaS is not solving it now.

The thing I would check on your six: how many activated. Six signups from one burst is a promising number and a dangerous one, because re-listing produces another burst and teaches you nothing new about whether the product holds them.

One small measurement trap here: 8K Search Console impressions isn't 8K visits. I'd split it into impressions → clicks → landing sessions → activation before deciding SEO produced zero signups. If the CTR was 1%, that's roughly 80 visits, so zero signups becomes much less surprising.

The reader-vs-tryer distinction still makes sense, but measuring from clicks rather than impressions would show how large the intent problem actually is.

Your own numbers made the argument: 3x SEO traffic, zero signups — and all your real signups came from discovery platforms. That's not an SEO problem, it's an intent problem: nobody googles their way into switching, but self-managing landlords complaining in their own communities are already in motion.

I hand-build prospect lists for founders. Want a free batch of the specific landlord communities and forum threads where your pitch actually lands? No strings.

This framework (SEO = readers, discovery platforms = tryers, community = credibility) is exactly the clarity I needed. I just launched a niche B2B tool and I've been treating SEO traffic and Product Hunt/launch traffic as the same funnel — clearly they're not. The "banned for 30 days for mentioning your product" bit is a brutal but useful warning too. Quick question: for the discovery platforms, did you find timing your listing (e.g., re-launching after adding features) actually reset the algorithm's attention, or was it more about hitting a fresh audience each time?

thanks for sharing :)
I agree with most things you said here, but one seems off: 8k daily impressions is too large for you to not drive any signups -- something must be off

what I'm implying is your website content may attract the wrong people or something deters them once they are on your website (I've checked and seen that you're offering a 30 day trial so you should have converted some visitors to users anyhow)

In comparison, one of my micro-saas has a very low traffic (450 visitors per month -- yes, you read that right) but it has a conversion rate from visitors to users of 5% -- every sass is different, but 8k daily and no signups feel way off

Updating my listings on discovery platforms with new features and tracking whether re-launching drives a new signup burst — or whether those platforms only work once.

I would focus less on traffic and understand why people avoid signing up even if you offer the 30 days trial

this is the same distinction that reframed how I think about my own week after your last post, and reading the specific breakdown here (BetaList, IH, LaunchPact, PeerPush vs. zero from SEO) makes the discovery-vs-answers split even clearer than before. these aren't just different channels with different conversion rates, they're capturing people at genuinely different points in their own decision process, one group is already looking for a tool, the other has a specific question and leaves the moment it's answered

the 30-day forum ban for mentioning your product is a real, costly data point too, worth weighing against the "credibility" framing. if community posts can only ever be credibility-building and never product-mentioning without real risk, that's a meaningfully smaller role for that channel than SEO or discovery platforms, worth being upfront about in the framework rather than implying all three are roughly equal legs of a stool

following your relaunch test with real interest, since "do discovery platforms only work once" is exactly the kind of question I'll need answered for my own eventual PH launch, curious if a second push on the same listing gets anywhere close to the first

The reader vs. tryer distinction makes a lot of sense. I’ve seen something similar with local services — someone searching for “how to fix a clogged drain” isn't necessarily looking to hire a plumber, while someone searching for “emergency plumber near me” has a completely different intent.

That’s something I’ve been paying more attention to with Truckee Plumber
separating informational searches from searches where someone is actually ready to contact a local service.

Traffic alone can definitely be misleading if you don't separate those two types of visitors

The reader vs. tryer split rings true, but I'd be careful about calling the two weeks a wash. Programmatic SEO at 8K impressions/day is mostly informational queries, and that intent gap won't close by waiting — it closes by segmenting. A "California rent increase rules" page can convert if the CTA is the next job in that reader's head (a free rent-increase notice generator, a state-specific template) rather than a generic signup. Right now those pages are probably asking a googler to adopt a property management system, which is a much bigger ask than the question they arrived with.

The other thing I'd check before concluding discovery platforms are the engine: those 6 signups came from ~13 sessions, which is a suspiciously good conversion rate and a very small sample. Discovery-platform visitors sign up because trying new tools is the activity — that's not the same as demand. The number I'd track is how many of those 6 imported a real property or lease in week two. If it's zero, then both channels are producing non-customers and the SEO/discovery framing isn't the actual constraint.

On re-launching listings: in my experience those platforms mostly work once for the "new tool" slot, but a feature-specific relaunch aimed at a narrow use case can get a second, smaller burst. Worth trying, though I'd treat it as one more traffic experiment rather than the thing that changes the $0 MRR.

Before rewriting the acquisition story off SEO, I would check one confounder in the Aug 1-7 burst: was anything else shipped that week, an email, a version update, a better landing page? A seven-day burst on discovery platforms is a timeframe, not proof those platforms carried it, and if it coincided with a launch email to an existing list, the channel attribution is less clean than it looks. Then measure the six credit-card users by week-four retention rather than count; at pre-PMF, keeping two of six engaged tells you more than the next six signups.

thanks for sharing this from the actual weeks. it made me think more about what I'm even measuring.

3x traffic and zero signups. that number lands.

I still don't have a first payment. the two weeks you stopped posting to build is the part that sits with me.

Your reader versus tryer distinction matches what we are testing with Speechara.Ai. Indie Hackers traffic can be smaller than broad search traffic, but the intent is often much closer to trying a product. We ended up watching the first useful transcript or translation, not just signup, because that tells us whether the visitor entered the workflow. I would split each channel by first meaningful outcome, return usage, and time to value. Have you seen whether the Indie Hackers and BetaList cohorts also retain better, or only convert better on the first visit?

One practical next test might be to align the CTA with the search intent instead of sending every SEO visitor to the same signup. For a query like “California rent increase rules,” a landlord checklist or calculator could capture a relevant email first; higher-intent queries could go directly to a product workflow. Then compare activation or qualified leads by query cluster, not just traffic. Your channel split suggests the mismatch may be between intent and CTA—not SEO alone.

This is the exact problem statement for measurement scope: you measured "traffic" (readers) but not "intent" (tryers). Both are real — you got exactly what your metrics said you should get.

The SEO visitors came to solve "California rent increase rules" and left. Discovery platform visitors came to try a new tool. Different platforms, different intent, different KPIs.

The fix isn't to abandon SEO. It's to measure what each channel is actually for. SEO measures readers (intent: get information, action: leave). Discovery platforms measure tryers (intent: try a tool, action: sign up). When you split the measurement, the strategy becomes clear without conflict.

Your framework nails this - each engine has a different KPI, not because you chose different goals, but because different audiences have different intents. The trap is conflating them into one "traffic" metric. When you do, you optimize for the average and lose both the readers and the tryers.

This is a brilliant breakdown. The "Traffic != tryers" distinction is something most founders learn the hard way. I am seeing the exact same split right now.

I launched Nexus, a Next.js 16 starter kit, a few days ago. My SEO traffic (Dev.to, Hashnode) brings in readers who read the technical write-ups and leave. But my actual demo visits and engagement came directly from Indie Hackers comments and the community here.

The discovery platforms have a specific intent: "I am looking for a tool to try right now." SEO has a broader intent: "I am looking for information."

That is why build cost matters so much. I spent months building Nexus with 626 tests and a 17-page admin panel so that when founders actually do find it via discovery platforms, the product is solid enough to convert them. If the product is half-baked, even the best discovery traffic won't convert.

Curious, when you were on those discovery platforms, did you notice a difference in conversion between founders looking for a complete solution versus those just looking for a quick, free tool to test?

The distinction between readers, tryers, and credibility is really useful. It explains why a big increase in traffic can look like progress while having almost no impact on early customer acquisition.

What stood out to me most is that the source of traffic seems more important than the volume at this stage. Someone searching for an answer to a specific problem has very different intent from someone browsing a platform specifically to discover and try new products.

Tracking each channel with a different KPI instead of treating all traffic the same feels like a much better way to evaluate what is actually working.

The two weeks weren't wasted, they were the control group. You just proved your pSEO pages rank for questions, not for the job your product does, which is a different keyword set entirely: "how do I track rent payments for 4 units", not "California rent increase rules". Same thing happened to me — traffic went up, nothing moved, because nobody googling a legal question is shopping. On the re-listing test, I'd keep expectations low; most discovery platforms give you one real spike and the second one is a third of the first. The thing that actually worked for me pre-PMF was going where landlords already complain to each other and just answering questions with no link at all, then letting the profile do the selling. Slower, but those signups actually stuck around.

That line about getting banned for 30 days from the biggest forum in your niche is a rite of passage for every indie dev! It's such a razor-thin line between sharing genuine value and getting flagged as self-promo. Your framework splitting "readers" from "tryers" is super clean, though. Re-launching on secondary discovery platforms with major feature updates is a solid play to trigger fresh signup spikes.

The reader vs tryer distinction is worth naming properly. SEO gets people in answer mode. They found what they searched for and moved on. That's not a traffic quality problem, it's intent mismatch. Pre-PMF, the only traffic that converts to trial is from people already in browsing-for-tools mode. Discovery platforms are that. Google organic almost never is, at least not at this stage.

One thing worth watching on the re-listing experiment: some platforms flag repeat launches. The 'new features' angle gives you a real hook though. What features are you leading with?

Matches what I see at much smaller scale: search traffic and community traffic aren't one funnel with two sources, they're two different populations, and adding them into "traffic" hides that one converts at roughly zero. Splitting the metric per channel was the fix for me too.

Living the same split right now marketing an AI product into a regulated niche (pharma quality teams), and your readers-vs-tryers cut matches our data almost exactly. Directory listings sent us tens-to-hundreds of visits and single-digit signups — we now value them as backlinks and legitimacy, not acquisition. The only thing that reliably creates tryers for us is showing the product actually doing the job, in-feed: a 30-second unedited screen capture of a real session with the claim in the first second. People who watch it through convert at a rate content clicks never touch.

One nuance on 'communities = credibility, never mention the product': in regulated verticals the buyers gather in fewer, quieter places, and what earns trust there is answering the compliance-shaped questions everyone else is afraid to commit an answer to. The product interest follows the person.

On your relaunch test — our experience so far is that discovery platforms are mostly one-shot per audience, but a feature relaunch with a genuinely changed headline promise behaves like a fresh listing on about half of them. Watching for your results.

This is a brutal but important lesson. I had the same realization with programmatic SEO for an API platform — we ranked well for technical queries and got thousands of impressions, but those users wanted documentation, not a billing solution. The discovery platform signups were worth 100x the SEO traffic because their intent was 'I need to find a tool' vs 'I need an answer.' One thing that is starting to work: putting contextual CTAs inside the SEO content itself, not in a header or sidebar. If someone is reading about rate limits, an inline mention of how your product handles that specific problem converts much better than hoping they navigate to pricing.

Great writeup — the signup-source breakdown is the part most people skip. One thing I'd add: your "which channel sent people who actually converted" question is exactly where analytics tools drop the ball. Most tools show sessions, not "these 6 came from BetaList". We built https://amami.dev to answer that kind of question in plain English.

This mirrors what I found building calculator tools in the UK finance space. I've got thousands of pages ranking for informational queries like "30000 salary after tax" and the traffic looks great in Search Console, but those visitors want an answer, not a product. They get it and bounce.

Your split is spot on. I'd add one more layer though: the discovery platforms seem to have a decay curve. BetaList got you 6 signups in week one, but I'd bet re-listing with updated features won't hit the same numbers. The audience on those platforms is browsing for novelty. Once they've seen you, you're not new anymore.

The community ban is real too. I lost access to a 17M-subscriber finance subreddit for posting one link. Genuinely thought it was helpful. Doesn't matter. Now I just answer questions without linking anything and let people click through to my profile if they're curious. Slower, but you don't torch the account.

One thing I'm testing: whether SEO pages can warm people up for a product if you put a soft CTA in context. Like if someone lands on "California rent increase rules" and you have a one-liner saying "we track this automatically for landlords on Leasebase", that's different from hoping they'll navigate to your pricing page. Not sure it converts yet, but the intent gap might shrink if the CTA is close enough to what they just searched.

Curious whether your BetaList signups actually activated or just signed up and ghosted. That's the next filter I'd look at.

This matches what I keep seeing on the link-building side too. People treat "more referring domains" the same way you treated 8k impressions — a vanity count.

Most competitor backlinks a new SaaS gets pointed at are press, investors, or customer subdomains you cannot copy. The handful that actually convert are usually still-maintained listicles or resource pages with a live submit path. Decorating a page nobody buys from just makes a prettier zero.

Your split is right at pre-PMF: SEO = readers, discovery platforms = tryers. I would only spend time on replicable links after you have one page that already converts from those discovery channels.

Also: community posts that do not mention the product is the correct trade. A 30-day ban is more expensive than a week of quiet building.

Your SEO pages answer reader questions, so the buying test is whether a model names you when someone asks the version of that question that ends in a purchase. Take three to five of those buyer questions, run each in a fresh chat on ChatGPT and Perplexity, and log the date, the question, and who got named. If you have never run that check, I can send a free five prompt snapshot for your category so you get a baseline before writing more pages.

I think there are different kinds of keywords. Some are just informational. Your offer has to be targeting a core commercial keyword. All your other pages should be served around this page to help it rank. For all the informational keyword pages, you should be actively advertise your core offering, given time, they helps your core offer page rank.

The strongest insight is separating traffic intent from acquisition intent. 3× more impressions can look like progress while producing zero customers if the visitors are solving an informational problem rather than actively looking for a product.