Shipped 2 PWAs in ~2 weeks → 3 signups, ¥0 MRR. The bottleneck wasn't what I expected.
Hi IH 👋
Two weeks ago I launched two small PWAs (a read-it-later app and a family subscription tracker, both Japanese-first). I went in worried about the usual thing — "will people convert to paid?" Two weeks of data later, my actual problem is one step earlier than that, and it's been a useful gut-check, so I wanted to share the numbers honestly.
The numbers so far:
-
3 signups total across both apps
-
0 paid → ¥0 MRR
-
A 4-article SEO cluster on the same topic (Zenn, Qiita, note, Dev.to) → ~0 public engagement (1 like on one, 1 comment on another, the other two zero at 24h)
My first instinct was "the funnel / paywall must be wrong." But you can't have a conversion problem on traffic you don't have. The real bottleneck is discovery — almost nobody is reaching the apps in the first place.
The mistake I made: I spent ~95% of the two weeks building and ~5% on distribution. Classic. I assumed that once the landing pages were live, search would slowly pick them up.
The concrete thing I didn't know: I checked Search Console and both of my landing pages come back as "URL is unknown to Google" — brand-new, low-authority subdomains just don't get crawled for weeks. So my own pages are getting a structural zero search traffic right now, and there's no quick fix on my side (it's indexing + time, not a code bug — the pages serve 200 fine).
The one channel that is live is borrowed authority: dev.to / Zenn posts get indexed basically same-day because the domain is already trusted. So for the next few weeks that's my only real path to traffic, and I'm leaning into it instead of waiting on my own SEO.
What I'm changing this week:
-
Stop optimizing the funnel until there's traffic to optimize.
-
Write for high-DA platforms with clear search intent, not for my own (unindexed) blog.
-
Actually measure traffic before I touch conversion again — I was flying blind on the "no traffic vs. no convert" question, which is embarrassing to admit but probably common.
One of the two apps is a Pocket alternative, which feels timely since Pocket is shutting down — that's the one piece of search intent I have working in my favor right now, so I'm focusing the content push there: https://readbox.dev-tools-hub.xyz
Question for you all: at launch, where did your first non-zero traffic actually come from? Owned SEO, a single platform that worked, a community, somewhere unexpected? I'd genuinely like to hear what moved the needle first for you — and I'll be in the forum this week returning the favor on other people's launches.
the bottleneck being one step earlier than you expected is the most useful thing in this whole post, most people stare at paid conversion and never notice nobody reached that screen. 3 signups isnt about pricing at all. its a top-of-funnel problem, and SEO clusters are a slow burn that wont show for months. for japanese-first niche tools id be way more aggressive about going where those exact users already gather, niche communities, dev twitter, replies, before betting everything on search. whats the traffic doing, are people landing and bouncing or just not landing at all?
The traffic question is the honest answer here: not landing at all. Both apps are still unindexed in Search Console, so the 3 signups came from the IH post itself, not organic. The SEO cluster was written before the pages were even crawlable, which is roughly the mistake in one sentence.
On Japanese-first communities — you're right that I've been underweighting that. I've been defaulting to English-language channels out of habit when the actual target user is in places like Zenn comments and specific Slack communities. Moving that up the priority list this week.
the read here is 3 signups in 2 weeks for two brand-new apps just isnt enough traffic to learn anything about pricing yet, dont over-index on the ¥0. ive shipped stuff that took 2 months of pushing before the first real signal showed up. id pick one of the two apps and go all-in on distribution for it instead of splitting attention across both, two half-launched things is harder than one done properly.
Usage data on the 3: honestly sparse. One came back once, the other two haven't returned in the dashboard. Which probably tells you more about the product-market fit signal than the ¥0 does. You're right that the ¥0 itself isn't the signal — the "aren't coming back" part is. Taking that as the actual thing to fix before I touch pricing.
The focus-on-one note also lands. I've been running both in parallel and the split attention is real.
Sparse usage on the first 3 is honestly better signal than the 0 MRR, it points straight at activation. The one who came back did something the other two didn't, even if it was random. I'd just message all 3 and ask what they hoped it would do the second they signed up. The gap between that and what they actually hit on day one is usually the whole thing. Did any of them reach the core action at all?
The thing I'd be careful with is assuming the bottleneck is discovery just because the traffic number is low.
With 3 signups, it's still very early to know whether the problem is traffic, demand, positioning, or something upstream of all three.
The expensive mistake here is spending the next few weeks optimizing for more visitors before you're sure the current signal means what you think it means.
I wouldn't make that call casually in-thread because it changes what you should measure next and what conclusions you're allowed to draw from the data.
That's a fair pushback, and honestly the sharpest one I've gotten so far.
You're right that I can't cleanly separate "no traffic" from "no demand" at n=3. The signal is too small to be confident either way. What I was trying to say — probably not clearly enough — is that I haven't even been able to run the demand test yet because almost no one has seen the landing pages. Both pages are still "URL is unknown to Google" in Search Console, so the 3 signups came from outside of search entirely.
My framing of "discovery is the bottleneck" was less "I've ruled out demand/positioning" and more "I can't start measuring demand yet." The distinction matters and I conflated them in the post.
Point taken on not over-indexing on more traffic before validating the current signal. I'll be more careful about what I claim the data says. Thanks for making me sharpen it.
That's the distinction I was trying to get at.
The part I'd still be careful with is that "I can't measure demand yet" and "I need more traffic first" can lead to very different next decisions.
I wouldn't make that call casually in a thread because it's easy to collect more data and still learn the wrong lesson.
If you'd like the tighter version, drop your email and I'll send it over properly.
The distinction you're drawing is useful and I think I conflated them in the original post. "Can't measure demand yet" and "need more traffic first" are both true but they're different problems — one is a sample size issue, the other is a prioritization call. I'll pass on the email, but appreciate the offer. If you ever write the tighter version publicly, I'd read it.
Fair enough.
Good luck with it.