Looking for feedback
Hey IH,
Rick here. Building CloudWise (https://cloudcostwise.io), an AI cost-optimization agent for AWS SMBs spending $5K–$50K/mo.
The honest numbers:
- 8 months in
- 0 paid customers
- 0 connected AWS accounts
- 17 days until my self-imposed go/no-go gate (May 15)
I have a working product (free tier, Shield $19/mo, Agentic $49/mo). What I don't have is real users telling me what's broken. I've been in build-mode too long.
Looking for 5 beta testers running AWS in that spend range. Beta gets:
- Free Shield tier for 90 days
- Direct Slack/email to me for bugs + feature requests
- 30-min onboarding call so I can watch you use it
What I want from you: brutal feedback. What's confusing, what's missing, what's noise, what made you bounce. I'd rather hear "this is useless because X" than polite silence.
Comment or DM if you're in. Happy to answer anything about the stack, the
business model, or the fact that I'm doing this solo with a hard deadline.
Trending on Indie Hackers
I have questions for you, not really a feedback on your product (sorry in advance)
What distributions strategies did you try? What worked or didn't work?
I see you have many blog articles on your landing page. Did it help getting more traffic?
No apology needed, these are the right questions.
Distribution strategies tried (8 months, in rough order):
SEO/content — wrote ~30 blog posts targeting "aws cost optimization
<vertical>" keywords. Honest result: ~200 organic visits/mo, 0 conversions.
The traffic exists but it's tire-kickers and people researching for a
blog post, not buyers. Articles helped credibility (when prospects Google
me, there's a body of work) but didn't drive direct revenue.
Paid ads — tried both LinkedIn Ads and Google Ads. Burned budget, no
conversions. My read: at $19–$49/mo price points the LTV doesn't support
paid CAC, especially with a cold audience that has to also connect an
AWS account (high-trust action). Paused.
Hacker News Show HN — posted twice. Both completely ignored, fell off
the front page in <30 min. Lesson: HN doesn't reward "AI cost optimizer
for AWS SMBs" — too crowded a category, no novel technical hook in the
title. Saving the next attempt for when I have a sharper angle.
Cold email (Instantly.ai, $97/mo) — paused yesterday. Tired GMass for a month but I didn't get any reply. Regarding Instantly.ai, the AI agent
hallucinated stats in every personalization ("I noticed your team runs
47 EC2 instances" — completely fabricated). Switched to hand-written
founder emails this week. 2 sent so far, 0 replies yet (sent today).
LinkedIn outbound — many partnership DMs sent over past 4 months to AWS consultants for a 20% rev-share program. Didn't work.
LinkedIn organic posting — I've been consistently posting technical content every day for 3 months.
Communities — I've been building in public on LinkedIn mostly and got many good ideas from the some engineers. A couple of IH post over the past 4 months. FinOps Foundation Slack pending approval. But non of it converted into paid users.
In progress: AWS Marketplace listing (multi-week registration process,
unclear if it'll move the needle but worth the option value).
Honest meta-lesson: I spent 6 months on product and content, ~2 months
on distribution. Wrong ratio. If I were starting over, I'd start outbound
in month 1, not month 6, and I'd skip paid ads entirely until I had
proof the funnel converts on warm/cold-outbound traffic.
Happy to go deeper on any of these.
Thanks a lot for the details! How much did you spend in ads?
Idea you can still try (if wasn't tried already in the cold emails): look for companies who hire AWS experts, and contact them saying you can reduce their AWS costs
Ad spend totals (rough, across both channels over ~3 months):
"reduce aws bill", and a few long-tail "<vertical> aws cost".
CPC was $4–$11. Got clicks, ~6 trial signups, 0 conversions.
10–200 employees in tech. CPC was $14–$22. Almost no signups
reached the product, most bounced from the landing page.
So ~$2,700 total for 0 paid conversions and ~6 free signups that
went cold. Lesson learned the expensive way.
On the hiring-signal idea, that's actually sharp and I haven't
tried it. The mechanics:
SMBs (LinkedIn Jobs, Wellfound, Indeed)
that warrants a full-time hire
Before they start, here's what 30 min in your account would
surface, at $0 to you."
Two things I'd want to validate before going hard on it:
HR will gatekeep aggressively.
they may want the new hire to "own" cost work, not pre-empt it.
Going to add this to my outbound queue this week. Will report back.
Thanks Thom, this is a more useful idea than 80% of the "growth
hacks" content I've consumed.
Looking forward reading the outcome!
this needs a live in-person offline demo.
I've been manual personalized outreaching for almost 2 months and demoed to many, but it seems I haven't hit the right ICP yet.
Your site looks clean, but asking for a live AWS connection right away is a massive trust hurdle that likely scares off cautious founders. Without a "playful" demo or a preview of the actual dashboard, that first step feels like a high-friction black box. Showing exactly what the "Agentic" optimization looks like before the signup would help lower the barrier and build instant credibility. What's the most common security question people ask before they're willing to connect their account?
Appreciate your review. That means a lot to me.
You're closer to right than you might think, most of this ships, some of it has discoverability problems.
What's actually live today:
"Try Demo: No AWS Account Needed" button on the homepage hero (shipped Apr 15). It drops you into demo mode with seeded sample data — no AWS connection, no real account scan. Quick signup is required (standard) but no AWS credentials.
Demo video embed below the hero (shipped Apr 15) showing real waste detection in action.
Air-Gapped Mode at /air-gapped: for teams that won't connect AWS at all even after signup. Upload a CUR file and CloudWise analyzes it. Zero connection required, ever.
Full IAM permissions breakdown page categorized by service (CUR, EC2, S3, RDS, Agentic write-role). Shipped Apr 6. But it's behind signup right now. Moving it public this week so anyone can see exactly what we ask for before they decide.
Public CloudFormation templates (read-only + remediation write-role). Always public:
To answer your direct question on most-common security questions before connect (5 things in order):
"What permissions does the role need?". Read-only IAM:
ce:GetCostAndUsage, ec2:Describe*, s3:ListBucket, rds:Describe*, dynamodb:Describe*, cloudwatch:GetMetricData.
Zero write. Full categorized list at /setup/permissions (going public this week).
"Where do you store my data?". Per-tenant isolated, encrypted at rest, deletable on disconnect.
"Can I scope to a sandbox account?". Yes, deploy the CloudFormation in any account.
"What about Agentic write permissions?". Separate stack, separate role, 15-min IAM sessions never cached, curated allow-list, every action requires explicit human approval, full rollback. Explicit denies on IAM/VPC/KMS/secrets.
"Are you SOC 2?". Not yet. This is the real blocker for buyers above ~$20K/mo AWS spend. Honest answer.
The fact that you read the page carefully and still concluded the demo and Agentic preview didn't exist tells me the homepage isn't doing its job on discoverability. That's on me. Fixing.
Coming from a UI/UX design background, my perspective is naturally a bit design-heavy. To add to the feedback: I noticed that CloudeWise uses quite a few colors and text blocks, which tends to scatter the user's attention.
Most general users (myself included!) don't scroll down very much. Because of that, the top section really needs to grab the eye instantly. Right now, the heavy amount of text and information makes it feel a bit less engaging compared to other sites.
If you can streamline this 'above the fold' area, the site will become much stronger! I'm sorry my feedback is mostly limited to the design side, but I hope this helps.
Lily, this is exactly the feedback I needed. You're right.
Counted ~17 elements and 6 color families above the fold. That's overload before anyone reads the headline. Those were piled up overtime.
Working on it now: cutting the pill badge, the 3-bullet list, the itemized waste rows, and consolidating to 2 colors so the eye lands on headline -> CTA -> outcome in 2 seconds. Will ship this week.
Thank you! Design feedback from someone with your background is rare and useful.
Thanks! Despite my feedback, your product is genuinely excellent, and I have no doubt it'll do great. Rooting for you!
If you’re open to it, I’d love to get your thoughts on our team’s product as well. It’s bunzee.ai. Please feel free to play around with it and give us your most brutally honest feedback. Whether it's about the UI or specific friction points where you think users might drop off, any insights would be hugely appreciated!
Lily, thank you. And yes, absolutely.
I played with bunzee.ai and here's my honest take:
What's strong:
"Stop guessing. Trust real data". That headline is clean, concrete, and leads with the pain. I understood the value prop in 3 seconds. The free + no-credit-card positioning is exactly right for indie hackers doing pre-validation research.
Where I got stuck:
The input box is the first action I need to take, and I didn't know what to put in it. A market name? My startup idea in plain English? A competitor URL? There's no placeholder text, example, or micro-hint to guide me. I hovered, stared, and almost bounced before trying a generic category name. That cold-start confusion is probably your #1 drop-off point. Users who don't know exactly what to type will leave rather than guess wrong.
The demo categories shown (LLM, AI Music, Code Review) feel big-company-scale, which made me wonder: does this work for niche ideas? If I'm building something in "AWS cost optimization for SMBs," will it find meaningful data or return empty? Showing one small/niche success case in the demo would reassure founders with non-obvious markets.
The score output ("54/100 - Wait & See") is interesting but I wanted to know: what moves that number? A one-liner on the 3–4 signals driving the score would make it feel data-backed rather than a black box. Right now I trust the headline but not yet the score.
Minor UX: the social proof bar (TechCrunch, Crunchbase, App Store, Play Store, Reddit, ProductHunt, GitHub, Chrome) is long. If all of those are earned press/listings, great, but if some are aspirational it could backfire with skeptical technical founders. I'd cut to the 3 most credible ones.
Bottom line: The core insight is solid and the data angle is genuinely differentiated from "ask AI what you should build." The product just needs to hold my hand through that first input. Fix the cold-start UX and I think retention improves significantly.
Good luck. Rooting for you.
I can't thank you enough for such detailed feedback. I've asked for feedback many times, but no one has been as thorough and thoughtful as the CloudWiseTeam. I honestly don't know how to fully express my gratitude!
I will definitely be sharing your intuitive and sharp insights with our team so we can brainstorm and implement better solutions. Because Bunzee.ai is built for both founders who already have an idea and those who don't but want to build an MVP, we are working hard not to lean too heavily in just one direction.
While it’s true that no single app can satisfy every single person, we will always strive to build a product that works for everyone. Wishing the absolute best of luck to both you and our team!