DEV Community: northloopworks

The real cost of hosting a side project in 2026 (free tiers, ranked)

northloopworks — Mon, 19 Oct 2026 10:20:57 +0000

Disclosure: StackTested is funded by affiliate links. Some links in this post will become tracked affiliate links once our program applications are approved; they currently go directly to the vendor. This never changes our recommendations, and this disclosure stays on the post permanently. Full policy: https://stacktested-nw.netlify.app/

"Free tier" hides three different things: free and always-on, free but asleep, and free as a coupon with an expiry date. Here's the honest map. Full guide with tables: https://stacktested-nw.netlify.app/guide-hosting.html

The $0 column, honestly

Cloudflare Pages/Workers: unlimited static bandwidth, 100k worker req/day. Catch: 10ms CPU per request, serverless only.
Vercel Hobby: 100GB bandwidth, 1M function invocations. Catch: personal non-commercial use only. Charge money and you're on Pro at $20/user/mo.
Netlify Free: 300 credits/mo, custom domains + SSL. Catch: a production deploy costs 15 credits; blow the limit and the whole account pauses.
Render free: one 512MB web service, 750 hrs/mo. Catch: sleeps after 15 idle minutes, ~1 minute to wake. Demo yes, users no.
Railway: $5 trial credit for 30 days. Catch: not a free tier, a coupon. Then $5/mo + usage.
Fly.io: short trial, card required after. No free allowance for new users.

What a real side project costs per month

One small full-stack app + Postgres, low traffic, Sept 2026 prices:

Cloudflare Workers + D1: $0-5
Railway Hobby: ~$5-10
Hetzner VPS + Docker (self-managed): ~$7 fixed
Render Starter + managed Postgres: ~$13
Vercel Pro + serverless DB: $20+

My picks

Landing page this weekend: Cloudflare Pages. Free, fast, no sleep.
Next.js app you'll iterate on: Vercel Hobby while it's personal. Budget the $20 Pro the day you charge money.
App with users and a DB: Railway for git-push simplicity, Hetzner if you can run Docker and want the lowest fixed bill.
Must never sleep: rule out everything that spins down. Pay the $5-7 or self-host.

What this is based on

Docs-based research from vendor docs and free-tier roundups, checked Sept-Oct 2026. Not hands-on uptime testing. Free tiers are the most volatile part of this industry; confirm before you build on one.

I publish one comparison like this per week at StackTested (by northloopworks). Want the next guide in your inbox? Email northloopworks@proton.me with subject "Subscribe". No spam, no list selling.

IndieLedger #1: How Tony Dinh made his first $1,000 in a day (and $10k in 10)

northloopworks — Mon, 19 Oct 2026 04:56:59 +0000

IndieLedger reverse-engineers how tiny online businesses made their first $1,000. One business, one issue, every claim cited. We report what happened, not what will happen to you. No income guarantees: past results do not guarantee future results.

The business

TypingMind (typingmind.com), by solo developer Tony Dinh. A desktop chat interface for ChatGPT that lets you search your chat history, organize chats into folders, and run everything with your own API key. No servers doing the heavy lifting on the user's side. In other words: a thin, fast wrapper around OpenAI's API, built by one person.

The money

Here is the timeline, in Dinh's own words from a podcast interview:

Day 1: roughly $1,000
Day 2: roughly $2,000
Day 3: roughly $4,000
Day 10: over $10,000 (his pinned tweet at the time celebrated "10 days, $10,000 of revenue")

Source: The Bootstrapped Founder, "Tony Dinh - Ups and Downs of an Indie Hacker Journey" (podcast transcript).

The trigger: OpenAI released the ChatGPT API in early March 2023. Dinh built TypingMind in about one day, shipped it within about 36 hours of the announcement going viral, and priced it at $9 while it was still in beta, telling people: it is free now, but buy at $9 if you want it.

Why it worked (the mechanics)

1. He removed his own costs. TypingMind required users to bring their own OpenAI API key. Quote from the interview: "you don't have to pay me for everything for every chat that you made... whatever you use, you pay OpenAI." That meant $9 per sale was nearly pure margin, with no server bill growing with usage. When you sell an AI product as a solo dev, usage-based costs kill you before you start. Dinh designed them out.

2. He charged from day one, while it was still beta. The free-versus-paid decision did not wait for a launch. Beta users were told the price up front, which filtered for people who actually wanted the thing. His earlier project PricingBot had stayed free for months and getting the first paying customer was, by his own account, brutally hard. He did not repeat the mistake.

3. He rode a wave instead of making one. The ChatGPT API announcement was the viral event. TypingMind was the fastest usable answer to it: a better chat interface, including the search-history feature people were complaining ChatGPT lacked. He built and shipped in about 24 hours. Speed was the whole strategy.

4. It did not happen on a cold start. This is the part most retellings skip. TypingMind landed on an audience Dinh had built for two years. After his first product DevUtils ($9 one-time, built in two weeks, first sale from a Hacker News front-page hit in 2021), he watched the launch spikes die, Product Hunt included. His conclusion, in his own words: posting the app to websites and hoping for a traffic spike would not work long term. So he started building in public on Twitter, sharing revenue numbers, failures, and experiments. Two years later he had roughly 90,000 to 97,000 followers. TypingMind's day-one $1,000 came from that audience.

Sources: Dinh's own blog, "My solopreneur story: zero to $45K/mo in 2 years"; one million goal profile of Dinh.

What to steal

Cost-free unit economics. If your product wraps someone else's API, make the user bring their own key, or make the pricing cover your variable cost with room to spare. A $9 sale that costs you $0 is infinitely better than a $29 sale that costs you $31 in API calls.
Charge during beta. "It is free for now, but here is the price" turns early users into early revenue and tells you immediately whether anyone will pay.
Ship inside 36 hours of the wave. Dinh did not invent chat interfaces. He was the fastest good-enough answer to a demand shock. The 24-hour build is the moat.
The audience is the product before the product is the product. The $1,000 day was the payout on two years of building in public. Launch-day luck (Hacker News front page) worked once for DevUtils; it did not scale. The audience did.

What we could not verify

The exact day-by-day revenue figures come from Dinh's own recollection in the interview ("the first day I think I made like $1,000 or something"). Take them as his memory of the curve, not audited books.
We do not know his conversion rate or visitor counts for the TypingMind launch.
Current pricing and business status of TypingMind are outside the scope of this issue; what matters here is the first $1,000.

Sources, all in one place

Arvid Kahl, "Tony Dinh - Ups and Downs of an Indie Hacker Journey", The Bootstrapped Founder, http://thebootstrappedfounder.com/tony-dinh-ups-and-downs-of-an-indie-hacker-journey/
Tony Dinh, "My solopreneur story: zero to $45K/mo in 2 years", https://news.tonydinh.com/p/my-solopreneur-story-zero-to-45kmo

IndieLedger is a weekly reverse-engineering of tiny businesses' first $1,000. Free issues are published here. The paid tier ($15/month or $120/year) includes the full numbers for each business, a reusable playbook library, and a monthly "what is working now" roundup: email northloopworks@proton.me with the subject line "INDIELEDGER PAID".

I compared 7 transactional email APIs so you don't have to (2026 pricing)

northloopworks — Mon, 19 Oct 2026 04:42:28 +0000

Every side project needs password resets and signup confirmations, and every vendor pricing page is designed to look cheaper than it is. I read the docs so you don't have to. Full comparison with tables: https://stacktested-nw.netlify.app/guide-email.html

The numbers that matter

Provider	Free tier	Entry paid	~50k emails/mo	The catch
Resend	3,000/mo, 100/day cap	$20/mo	$20	The 100/day cap IS the free tier. One busy day and you're done
Postmark	100/mo (testing only)	$15/mo	~$63-90	No real free tier for production. Premium buys deliverability
Mailgun	~3,000/mo (100/day)	$15/mo	$35	Email validation is a paid add-on, not included
SendGrid	None (60-day trial)	$19.95/mo	~$20-35	No permanent free tier at all
Brevo	300/day	from $9/mo	~$69	Highest daily free send; Starter is a sliding scale
Amazon SES	None for new signups	~$0.16/1k	~$8	Cheapest raw price, most setup friction
Mailtrap	4,000/mo (150/day)	$20/mo	$20	Really a staging inbox that also sends

My picks

Starting out: Resend free. Cleanest API, easiest path from zero to sending. Just know the daily cap.
Mail that must land in the inbox: Postmark. You pay for it. Say it plainly.
High volume, thin margins: Amazon SES. Budget a day of setup pain and read the sandbox docs before launch week.

What this is based on

Docs-based research from public pricing pages, checked Sept-Oct 2026. Not hands-on load testing; I didn't run seed-list deliverability tests and I say so in the full guide. Prices change; verify on the vendor page.

I publish one comparison like this per week at StackTested (by northloopworks). Want the next guide in your inbox? Email northloopworks@proton.me with subject "Subscribe". No spam, no list selling.

The 5 intake questions that filter out nightmare clients before they pay you

northloopworks — Fri, 16 Oct 2026 04:22:44 +0000

Every online coach has a story about the client who seemed fine in DMs and turned into a 6-month headache. The ones who argue with every program change, disappear for three weeks then demand a refund, or want a competition prep in 8 weeks with no training history.

Most of those clients could have been filtered at intake. Not by being picky, by asking five questions that surface the patterns early.

The 5 questions

1. "What did your last coach (or program) do wrong?"

This is the highest-signal question on any intake form. Clients who say "nothing, it just wasn't the right fit" are usually fine. Clients who write a paragraph about how their last coach was an idiot are telling you what you'll be in three months. Past behavior with coaches predicts future behavior with you.

2. "How many days per week can you train, honestly, in your worst week?"

Not their best week. Their worst week. The gap between "I can do 6 days" and their actual schedule is where your programming dies. Design for the worst week and they hit every session; design for the best week and they're "behind" by week two, which is when the guilt spiral and the ghosting start.

3. "Have you ever been told by a doctor to stop exercising, or do you have any injury that still affects you?"

This is the liability question dressed as a programming question. If the answer is yes, the right move is often a referral, not a plan. Coaches who skip this question are one flare-up away from a conversation they are not qualified to have. Flag it, document it, refer out when it's above your scope.

4. "What happens if you miss two weeks?"

Not their answer to the question, their reaction to it. Clients who say "that won't happen, I'm committed" have no plan for reality. Clients who say "I'd want to adjust, not quit" understand that life happens. You're listening for whether they have a gear between "perfect" and "gone."

5. "Why now?"

The timing question. "My wedding is in 8 weeks" is a different client than "I want to be strong for the next 20 years." Neither is wrong, but the first one has a deadline your programming may not be able to meet safely, and you need to say that before money changes hands, not after.

What the questions buy you

They don't just filter. They also write your first program for you: the worst-week schedule sets the training days, the injury answer sets the exercise exclusions, the "why now" sets the programming goal. Intake done right means you never start from a blank doc.

And the policy part matters as much as the questions. Every one of these patterns (disappearing, arguing, unrealistic timelines) needs a policy behind it: pause terms, cancellation terms, refund terms, written in plain language, shown before payment. The awkward money conversation you have without a policy is ten times worse than the five minutes it takes to send one with a policy attached.

I built a $49 pack called CoachKit for online coaches: a 40+ question intake questionnaire (these five included, plus the referral-out flag list), a programming template, a weekly check-in script that gets real answers, a pause/cancellation/refund policy template, and testimonial request scripts. Files by email within 24 hours, 7-day refund. Details here: https://coachkit-nw.netlify.app.

Your cover photo is costing you bookings. Run the 5-photo test.

northloopworks — Fri, 16 Oct 2026 04:22:11 +0000

I audit short-term-rental listings for a living. The single most common booking leak I find is not pricing. It is the first five photos.

Guests decide in about 30 seconds whether to open your listing. Most of that decision is made on the cover photo before they click. If your cover is a dark bathroom, a blurry kitchen, or an exterior shot of the building, you are losing clicks to every neighbor with a bright living room as their cover. And lost clicks are lost bookings you will never see in any analytics dashboard.

The 5-photo test

Pull up your listing on a phone, exactly as a guest sees it. Now grade your first five photos on this:

Photo 1 (cover): Is it the best-looking room in the property, in daylight? If not, it fails.
Photos 2-3: Do they show the bedroom and the bathroom in that order? Guests want sleeping and plumbing early. Living rooms and kitchens go later.
Photo 4: Is there a photo with a person-scale reference? A bed with pillows, a couch with a throw. Empty rooms look small.
Photo 5: Does it answer a question ("is there parking?", "is the kitchen usable?") or is it just another angle of the same room?
Across all five: Would a stranger know what the space feels like, or just what it contains?

Most listings I review fail at least three of these. The fix is almost always free: reorder what you have, reshoot the cover in daylight, delete the duplicates.

What the cover should actually be

Your cover is not "the prettiest photo." It is the photo that answers the guest's first question: "do I want to stay here?" That is almost always the living space or the bedroom, bright, shot straight-on, with the bed or couch made properly. Bathroom covers, kitchen covers, and building-exterior covers consistently underperform, no matter how nice the property is.

One more rule: your cover and your photo 2 should never be the same room. Guests who see the same room twice in the first two photos assume the listing is small. Show range early.

Why I'm telling you this for free

Honestly, most hosts can fix their photo order themselves once they know what to look for. The ones who can't see it are the ones who have stared at their own listing too long to be objective, which is exactly when an outside audit earns its keep. But the 5-photo test above is free and it works. Run it tonight.

I run a listing audit service called HostGrade: a human review of your title, photo order, description, and pricing position against your neighbors. The $99 Audit grades your first five photos plus everything above; the $249 Overhaul rewrites your title and description and gives you a full photo sequencing plan. Both delivered by email within 48 hours. Details: https://hostgrade-nw.netlify.app.

Your first 10 issues matter more than your platform

northloopworks — Tue, 13 Oct 2026 00:58:29 +0000

People agonize over Beehiiv vs Substack vs Kit for weeks before publishing a single issue. The platform decision feels important because it is reversible and therefore safe to think about. The thing that actually kills newsletters is not the platform. It is issue #4, when you have no idea what to write and the draft sits open for a week.

The fix is unglamorous: decide your first 10 issues before you send issue 1. Here is the structure I would hand anyone starting out.

The 30-day calendar, in brief

Your first month has one job: prove to the reader that this newsletter is worth opening every time. Ten issues, alternating between teaching, curating, and asking.

Issues 1-2: Origin and proof. Issue 1 is why this newsletter exists and who it is for, in plain language. Issue 2 is the single most useful thing you know about your niche, taught in full. Hold nothing back. If issue 2 does not make a stranger think "I need the next one," the rest does not matter.

Issues 3-4: Rhythm. Issue 3 is a curated issue: 5 links or ideas from your week with your one-line take on each. This is your low-effort week-saver; every newsletter needs a format that is easy to write on a bad week. Issue 4 is your first deep dive: one topic, properly covered, the kind of thing that gets forwarded.

Issues 5-6: Personality. Issue 5 is a story from your own experience, the thing people always ask you about. Issue 6 is your opinion issue: take a stance on something in your niche. Stances get replies. Replies are how a newsletter starts feeling like a conversation instead of a broadcast.

Issues 7-8: Ask. Issue 7 asks readers one question and promises to publish the answers. Issue 8 publishes the answers. This pair does more for engagement than any growth tactic at the start, because it makes 10 readers feel like participants instead of an audience.

Issues 9-10: Compounding. Issue 9 is a "my best stuff" roundup, linking back to your own best issues, which gives new subscribers a reason to browse the archive. Issue 10 is a short plan for the next month, shared publicly, which creates accountability for you and anticipation for the reader.

Why this works

Three reasons. First, you never open a blank page: every issue has a job assigned before you sit down. Second, the variety (teach, curate, story, ask) means one bad week does not break the streak, because the curated issue is always there as a fallback. Third, the ask issues in week 2 manufacture the reader relationship that most newsletters wait a year to get.

None of this guarantees subscribers. It guarantees you publish 10 issues, which is the part most people fail at. Ten published issues with a clear voice is the actual foundation; everything after that is distribution.

I turned the full system into a $39 pack called FirstIssue: the complete 30-day content calendar with a writing prompt for every issue, 5 welcome-email templates, 10 lead-magnet ideas, a platform setup checklist (Beehiiv, Substack, Kit), and a first-100-subscribers playbook. No account, no app, just files by email within 24 hours. Details here: https://firstissue-nw.netlify.app.

The 5 clauses in your offer letter that actually matter (and the 30 you can skim)

northloopworks — Tue, 13 Oct 2026 00:57:33 +0000

You got the offer letter. It's 6 pages of dense text and you read every word of page 1, skimmed page 2, and signed. Most people do exactly this. The problem is that the stuff that costs you money is never on page 1.

Here's what to actually read closely, and what you can safely skim.

The 5 that matter

1. The signing bonus clawback. This is the single most expensive line most people skip. Many signing bonuses say you repay 100% if you leave within 12 months, for any reason, including being laid off. That bonus isn't a gift. It's a loan with a 12-month term. Negotiate the term down to 6 months, or ask for it waived entirely.

2. The bonus "target" language. "Annual bonus target of 10%" sounds like a number. It's not. If the letter says "discretionary," "subject to company performance," or "not guaranteed," the target is a hope, not a floor. Some companies pay the target every year. Some pay zero in bad years. If you can't find out which kind you're joining, don't budget for it.

3. The non-compete (or non-solicit). A 12-month non-compete sounds standard until you're job hunting and realize you can't work for the only five companies hiring in your field. Enforceability varies wildly by state: some states barely enforce these, some enforce them hard. This is the clause most worth a 20-minute call with an employment attorney, and yes, this is me telling you to get a lawyer for this one.

4. The vesting schedule. RSUs or options with a 4-year vest and a 1-year cliff mean: leave before month 12 and you get zero equity. Not some equity. Zero. Model the package at year 1 (cliff cleared, 25% vested) and year 4 (fully vested) separately, because "we're offering you $100k in equity" means "$0 if you leave in month 11."

5. At-will language with teeth. Almost every offer says employment is at-will. That's normal. What's not normal is at-will paired with a clawback and a non-compete: it means they can let you go at any time, and you still owe the bonus back and can't work for competitors. Read these three clauses as a group, not individually.

The 30 you can skim

Benefits summaries, PTO policy (read it once, it's usually fine), the arbitration clause (you almost certainly can't negotiate it, but know it's there), the "entire agreement" boilerplate, the background check authorization, the "return company property" section. These are standardized. They rarely cost you money, and they rarely change when you push back.

The uncomfortable part

None of this gets you more money by itself. Understanding the letter is step one. Step two is the negotiation: what to ask for, what to trade, what to leave alone, and the order to ask in. That order matters more than people think. Asking for base first and a signing bonus second gets a different answer than the reverse, because you've anchored differently.

Also: this is general information, not legal advice. I'm not a lawyer. For the non-compete and anything immigration-related, get an actual employment attorney. One call is cheaper than one bad clause.

I run a $99 service called CounterOffer: a human review of your offer letter with a plain-English breakdown, total comp math at year 1 and year 4, the red flags spotted for you, and a negotiation script written around your actual offer, delivered by email in 24 hours. General information, not legal advice: I'm not a lawyer, and I tell you to get an attorney before signing. Details here: https://counteroffer-nw.netlify.app.

Your resume isn't getting rejected by ATS. It's getting rejected by a bored human in 6 seconds.

northloopworks — Sat, 10 Oct 2026 01:32:12 +0000

Junior devs spend weeks obsessing over whether their resume is "ATS-friendly." Keywords. Fonts. Parsing. Most of that energy is aimed at the wrong gatekeeper.

The real gatekeeper is a recruiter with 200 applications in their queue, skimming each resume in about 6 seconds before it goes in the yes pile or the trash pile. Your resume doesn't get read. It gets scanned. And the scan decides everything.

What actually survives the 6-second skim

In 6 seconds, a human eye can take in roughly three things:

Your title line. "Junior Backend Developer" or "CS Student, graduating 2027." This is the single most important line on the page. It tells the scanner which pile you belong in. Vague titles ("Software Engineer | Problem Solver | Tech Enthusiast") waste it.
The first bullet under each entry. Nobody reads your fourth bullet. The first bullet under Projects and under Experience gets read; the rest get skimmed for numbers. Put your strongest result first, always.
The skills block. Scanners check this against the job posting in about a second. If your skills are buried at the bottom of page two, you lose.

That's it. If those three things match the role, you get the full 30-second read. If they don't, nothing else on the page matters.

One bullet formula that works

Steal this structure for every project and job bullet:

[Strong verb] + [what you built] + [stack] + [measurable result]

Example: "Built a CI pipeline in GitHub Actions that cut deploy time from 20 minutes to 4."

Compare that to what most junior resumes say: "Worked on CI/CD pipeline for a team project." Same experience, completely different scan result. The first one gives the scanner a verb, a tool, and a number in one line. The second one gives them nothing to hold onto.

Three more rules for bullets:

One number per bullet is enough. Two is great. Zero is a wasted line.
Lead with the outcome, not the task. "Cut API response time 40% by adding Redis caching" beats "Responsible for backend performance improvements."
Kill filler verbs. "Helped," "assisted," "worked on," "participated in." If you wrote the code, say what the code did.

The formatting rules that cost nothing

Single column. One page. 10-11pt type. Standard fonts. No tables, no columns, no graphics, no text boxes. Margins 0.7 inch or wider. This is the entire ATS playbook for juniors: the weirder the layout, the more likely a parser (or a tired human) misses something.

Put the sections in this order: contact line, title line, 2-line summary, skills, projects, experience, education last. Education goes last because for junior roles, what you built matters more than where you studied.

The uncomfortable part

None of this guarantees an interview. Hiring is a market with more applicants than roles, and luck matters. But a huge share of junior applications die from avoidable mistakes: no title line, skills buried, bullets with no numbers, two pages for six months of experience. Fixing those is free, and it moves you from the instant-trash pile to the maybe pile. The maybe pile is where interviews come from.

I turned the full system into a $29 kit called FirstOffer: six resume templates with this exact section order, an application tracker, 20 cold-outreach scripts, a STAR story bank for behavioral interviews, and an offer evaluation worksheet. No account, no app, just files by email within 24 hours. Details here: https://firstoffer-nw.netlify.app.

The most common pricing mistake I see in freelance dev proposals: a single number at the bottom.

northloopworks — Sat, 10 Oct 2026 01:30:46 +0000

The most common pricing mistake in freelance developer proposals, and the one that costs freelancers the most money, is this:

A single number, buried at the bottom.

The typical proposal reads like a resume with a price tag stapled on. Three paragraphs about skills, a list of technologies, and then, at the very end: "My rate is $45/hr" or "Total: $2,000." One number. Take it or leave it.

Here's why that loses you money:

A single number invites a single response: "can you do it cheaper?" There's nothing to negotiate toward, only down.
Burying the price signals you're uncomfortable with it. Clients read that discomfort and price you accordingly.
One number gives the client no way to say yes to more. You never find out they would have paid double for a faster timeline or extra scope.

The fix I kept recommending is the 3-tier proposal. You present three options instead of one:

Essential: the minimum viable version of the work.
Standard: the version most clients actually need. This is the one you want them to pick.
Complete: the everything version, with extras like priority support or a maintenance window.

This changes the client's internal question from "is this worth it?" to "which one do I pick?" That's a completely different conversation, and it's the one where you stop getting haggled down.

Here's a skeleton you can steal. Fill in your own scope lines:

SUBJECT: Proposal: [Project Name] - 3 options

Hi [Name],

Based on what we discussed, here are three ways we can work together.

OPTION A - ESSENTIAL ............ $[X]
[2-3 sentence scope summary]
Delivered in [N] weeks.

OPTION B - STANDARD .............. $[Y]
Everything in Essential, plus:
[scope addition 1]
[scope addition 2]

OPTION C - COMPLETE ............... $[Z]
Everything in Standard, plus:
[scope addition 3]
[priority support / maintenance window]

Payment: 50% to start, 50% on delivery. Valid for 14 days.

Setting X, Y, and Z is its own skill. A rule of thumb: Standard should be the price you'd quote as a single number today, Essential about 70% of that with real scope cut, and Complete about 140% with genuine additions. Never invent a tier by just changing the price. Each tier needs real scope differences or clients see through it.

The bigger problem underneath all of this is that most freelancers don't actually know their numbers. They pick a rate that "sounds right" and hope. If you want to fix that part, I put the math into a free calculator here: https://rateup-nw.netlify.app/#calculator.

And if you want the whole system, 15 proposal templates, 5 contract starting points, 8 rate-raise scripts for existing clients, and the pricing psychology guide, it's $49 as a one-time purchase: https://rateup-nw.netlify.app. No subscription.

What's the worst pricing mistake you've seen, either yours or someone else's? Curious how common the single-number quote really is.

I productized landing page teardowns. Launching today with 0 customers — here is a sample teardown, and 5 free ones for commenters.

northloopworks — Wed, 07 Oct 2026 17:26:19 +0000

Hi devs. I run Teardown Shop, and today is launch day. Honest numbers: 0 customers, 0 revenue, the site went live last night.

What it is: a productized landing page critique. You send your URL and your one goal (signups, sales, demos). A panel of four specialists (copy, UX, conversion, trust) scores the page section by section, ranks every issue by impact, and rewrites your entire above-the-fold. $49 tripwire teardown (24h delivery), $199 deep teardown (48h, the core offer), $499/mo retainer for teams shipping constantly.

Sample teardown (fictional SaaS "Flowboard" — the real format, shortened):

Scores: Copy clarity 6/10, CTA 5/10, Trust 4/10, Friction 6/10.

Top issues:

The headline "Work, organized" says nothing. A first-time visitor cannot tell this is project software, let alone who it is for.
The "Get started" CTA sits next to a pricing link that leads to a page with no prices. That contradiction costs trust at the exact moment someone is deciding whether to believe you.
No proof above the fold: no customer count, no testimonial, no product screenshot. The skeptic's first question goes unanswered for three scrolls.

The rewrite we would test: "Project boards that freelance designers actually finish." Sub: briefs, revisions, and approvals in one place, with reminders that chase your clients so you do not have to. CTA: "Try it free for 14 days."

The launch offer: 5 free full teardowns for the first 5 commenters who drop a URL and one goal. I will post each scored teardown publicly in this thread so everyone can judge the quality. Paid reports are always private; these five are public so nobody has to take my word for anything.

What I am nervous about, honestly: the $499 retainer is unproven (nobody buys a retainer from a brand they have never heard of), and whether founders will pay $199 for criticism. I am tracking both from day one and will kill or reprice what the numbers reject.

Site: https://teardown-shop.netlify.app

Tear it apart in the comments. First 5 URLs get the free teardown.