The first screen a new user sees is usually the least designed screen in the whole product.
They sign up, land on an empty dashboard, and have to imagine what it looks like when it's full. A lot of them close the tab right there.
What helped me was picking one action that means this person gets it, then measuring how long people take to reach it.
Sample data, a prefilled example, a demo project, anything that kills the blank page would go a long way to address the dead experience concerns.
1
117
Your support inbox is the cheapest research you will ever run.
One habit that changed things for me: for every message, write down one word for the cause sitting underneath the complaint. Confused by pricing, lost in setup, expected a feature that isn't there.
After about 30 of those you will have a better understanding in which direction the product needs to go.
1
1
98
Building the thing got about ten times cheaper in two years. Getting a stranger to look at it costs exactly what it always did.
The bottleneck moved and a lot of us are still tuning the part that already works.
What's actually pulling people in for you, outside of posting here?
67
Read this in my first month of building and I still come back to it every few months.
Paul Graham, do things that don't scale: paulgraham.com/ds.html
The bit that stuck with me is that recruiting users by hand is how you learn what to build. Reach out to real people !
48
5% monthly churn sounds harmless.
Run it out though. In 20 months you have replaced every customer you have. You are paying to acquire people just to stay level.
Work out your real number before you spend another month on content or ads. It changes what you should be fixing.
1
47
If you're past 10 paying customers, I want the boring version of the story.
What did you actually do to get the first 10, and would that same thing still work today?
A lot of the advice going around is from a time when those channels were empty. Nobody went back and updated it.
32
Most useful thing I read on Indie Hackers today, and there's no win anywhere in it.
Founder shipped v2 with zero paying customers. Two weeks after launch he had 9 signups. Two were him and his cofounder. Two were friends being nice.
His cold DM reply rate was 37 percent, which most people would post as traction. Nobody paid. His own read, people were curious, not buying, and the product had rough edges he was quietly hoping nobody would notice. They noticed.
Then he found checkout was silently failing for some of the people who actually wanted to pay.
Two things I'm keeping from this. Reply rate is not demand. And go buy your own product end to end today, because you genuinely do not know what's broken until you do.
indiehackers.com/post/i-ship…
2
2
71
Small framework that made my docs hurt a lot less.
It's called Diataxis. The idea is that every page in your docs is exactly one of four things, and mixing them is why docs feel confusing.
Tutorials, learning by doing.
How-to guides, solving one specific task.
Reference, dry facts only.
Explanation, why it works this way.
Once you name the type, the voice writes itself. Reference goes terse. Tutorial holds your hand. You stop writing pages that try to teach and specify at the same time.
Extra reason to care right now, agents read your docs before humans do, so clean structure means better answers about your product.
diataxis.fr, and the HN thread has good real world notes from people who've done the migration
news.ycombinator.com/item?id…
1
55
If you're building anything that needs to be found, AI search is turning into a real distribution channel and most of us are still optimizing like it's 2019.
A team ran an analysis across millions of citations from ChatGPT, Google AI Overviews and Perplexity.
Wikipedia came out as the source ChatGPT leans on most. Reddit came out first for both Google AI Overviews and Perplexity.
So the practical move isn't just publishing more blog posts. It's showing up in the places these models already trust: genuine Reddit threads in your niche, comparison pages, directories, industry roundups.
Spend twenty minutes checking which prompts your competitors get cited for. It's usually humbling.
(worth noting this is the vendor's own research, so take it as directional, but it tracks with what I've been seeing)
leapd.ai/blog/ai-visibility/…
2
47
Best thing I read this week on validating an idea.
Takes about ten minutes.
This founder sized his market properly first. Upwork has 18 million plus registered freelancers. Sounds enormous.
He then defined active as two projects won and $6k earned in the last six months. The real number came out around 30,000.
Then he checked spending power inside that slice. Roughly 16,000 earning $2k a month or more, about 5,000 earning $5k plus.
Then he paid $500 for one month of his biggest competitor's product just to see if it was beatable. He says if it had actually been good, he would not have built anything.
He hit $10k MRR in 60 days. Now at $20k with two people.
Registered users is a vanity number. Go find the active paying slice.
indiehackers.com/post/tech/e…
47
Read this one twice today.
A guy is doing about $15k/mo across roughly 30 tiny sites, and his whole method is that he never creates demand.
He finds defunct companies whose domains still rank and still pull traffic, buys the domain, and builds the thing visitors were already showing up looking for.
The warning buried in it is the good part: Google will de-rank you if you don't build what people expect that site to be.
He also says ads were a total waste, and that most of the leads Google and Meta sent him were bots. His bot filter is charging $1, because bots will fill out anything but they won't pay.
indiehackers.com/post/tech/h…
25
Posting updates & content in parallel with building the product, this whole solopreneur thing is certainly going to give me imposter syndrome !
1
55
I wonder what all the people who bought Mac minis in February to run OpenClaw are doing now.
Have they successfully value-maxed every moment of their lives while their agents worked through the night, or are those Mac minis just expensive paperweights now?
54
Really good insights. I do think that, in the coming years, individuals with strong problem-solving skills will have a real chance to bring their ideas to market much faster.
But @patrickc and @stripe, please unblock Stripe onboarding for solopreneurs who want to build and sell globally from countries like India.
New from Stripe Economics: The Age of the Solopreneur
stripeeconomics.com/p/the-ag…
1
316
What sort of resource allocation is Anthropic facing with Fable 5, meanwhile OpenAI is chill with GPT 5.6 ?
Curious how they are spending on the tokens.
Beginning July 20, Claude Fable 5 will be included in all Max and Team Premium plans, at 50% of limits.
Pro and Team Standard users will continue to have access to Fable via usage credits, and will receive a one-time $100 credit.
Demand for Fable has been challenging to predict, which is why we rolled it out to subscription plans in stages, extending access several times as we secured additional capacity.
63
So Claude's dedicated services to the Govt, rarely goes down. It's only for the wage-scramblers.
1
102
With CueLight, your screen, camera and audio stay separate after recording.
So you can resize your webcam, move it or completely change the layout without recording the demo again.
Currently working on this part to make the experience better.
1
79
First preview of the app.
Will provide a solid in browser, editing and export option.
Started working on the presentation view so people can record their cams into the layout.
93









