Riffit Month 3: shipped two features and 12 blog posts. The funnel did not move.
I'm building Riffit, WhatsApp invoicing for Indian freelancers. Solo, designer by trade, no coding background. A lot shipped this month. The numbers barely noticed. Here's both halves.
What shipped this month
Two product launches plus a distribution push. Contracts went live inside Riffit: I built 4 contract templates, then shipped just one, a standard project contract, to keep the surface small while I learn how people actually use it. Also live: a client content generator, plus a round of code refactors under the hood.
On distribution, the blog went from 13 to 24 live posts. 12 new articles in 30 days, plus title and meta rewrites on 7 older posts that were ranking on page 1 with broken click-through. To make that pace survivable on a 3 to 5 hour weekly marketing budget, I built a content operations system on top of Claude: 11 small skills that pull live Search Console data, decide between new posts or CTR fixes, and draft Reddit replies matched to a corpus of 23 of my real comments. Marketing compressed from write-everything to review-and-post.
The numbers
18 real signups now, up from 16 a month ago. 3 of those came in the last 30 days: one has already sent 2 invoices, the other two have sent none. Overall, 10 users have sent at least one real invoice, 16 real invoices total, and still exactly 1 user has crossed 3 invoices. ₹0 MRR. Search is the bright spot: 30 clicks and 1,578 impressions in the last 28 days at average position 7.2, up from roughly 21 clicks a month last quarter. Half of those clicks are people searching "riffit" by name, which is encouraging and also a reminder that the non-brand engine is still small.
What I learned
I made content production efficient and measurable, and the funnel barely noticed. Activation 0 to 1 works when people show up (median signup to first invoice has been about 7 minutes), but invoice 1 to 3 is where everything stalls, and no volume of blog posts fixes that. Also embarrassing: UTM capture is still broken. The database columns exist now, and the count of attributed signups sits at zero. Every channel decision I made this month was made blind.
What's next
Two specific things before the July update: personal WhatsApp messages to the 5 users who sent exactly one invoice and stopped, to find out why there was no second. And UTM capture working end to end, verified with one real attributed signup, before I write another content batch.
Question for the IH community
If you've solved a "first action works, second action stalls" gap: what actually moved the second-action number for you? Product nudges, personal outreach, or something else entirely?
The part that stood out to me is that you may be treating this as a traffic problem when the data is starting to look more like a repeat-usage problem.
You already proved someone can sign up and send the first invoice.
The expensive mistake now would be spending another month improving acquisition while the real bottleneck sits between invoice one and invoice three.
I wouldn't answer the product-vs-outreach question casually because the fix depends on why users stop after the first invoice.
Happy to put the tighter activation path in writing if useful. There is a real risk of collecting more users before understanding why the current ones stall.