Ask HN: Could you please brutally review my product website? | Hacker News Reader

Ask HN: Could you please brutally review my product website?

perfai.ai

27 pointsintesars109 comments

During YC, we relentlessly practice our “two-sentence description.” It’s roughly of the form “ACME makes soup taste better. We do it with a seasoning that chefs add to their broth.” That is, what you do, and how you do it (or an example.)

Yours might be “Perfai makes APIs faster. We do this by ?” The “how” is unclear.

Are the case studies from actual users? Why all of the disclaimers?

Your site mostly makes the claim that “slow APIs lead to churn.” If that is true, it is not obvious to me. I almost wonder if the copy was written by AI, as it uses plausible-sounding verbiage but is unconvincing.

I believe the real problem is that you haven’t talked to enough potential users about their API problems. If you were able to quote their pain points in their own words, I think the copy would more convincingly demonstrate there is a problem and how your solution fixes it.

I agree with all of your points.

1. Instead of doing a "two-sentence description" I did "problem-solution-result" style i.e. "Poor API performance results in negative UX and high customer churn. Deliver high-performance APIs to stop churn and boost 3x retention.". I like your suggestion "PerfAI makes APIs faster. We do this by" and I'll work on it based on your feedback.

2. Our CS team published the case studies at the last minute without a proper review. We'll fix it.

3. Most of the claims are actually backed by one of the Google white paper. We'll add its references to make it sound convincing.

4. You're correct we don't have customers we just started. We're in the process to getting potential users.

Thank you for your feedback!

Kudos to you for asking for honest feedback. I don't mean to pile on; only to clarify something about the pitch on your website.

The thing with "problem-solution-result" style descriptions is that they can rely on (insider) knowledge about both the problem and the solution. Your pitch / tagline works great for people who know about both... even though this is unlikely to be the case outside of people working for your company. I.e. you want readers / potential customers to imagine how your solution applies to their problem, not to the one you've described for them.

If a user / potential customer is tech-savvy enough to know about API testing, it's a good bet that they won't just take it on faith that AI testing will hand-wavily solve their problem. I think you can trust that your potential customers don't need to be heavily sold on the idea that API performance is critical. Instead, I recommend focusing on coming up with marketing copy to address the following questions: * What does your AI testing do that internal engineers and a set of tests can't? * How can perfai.ai augment engineering efforts? * Can perfai.ai find things that traditional test-suites miss?

You sort of address this in the "No Code / No Config" section, but it's none too clear and takes some digging to figure out. Speaking of which "bringing the concept of Shift-Left to API active performance" is inside baseball.

Hope this helps!

I changed the heading to:

AI-Powered API Performance Testing No-Code, Self-Learning!

Since AI learns how to interact with the API and validates every path in it. Is the new sub-heading any better than the previous one?

Also, shortened all the other sections.

Thank you so much for the feedback!

You are on the right track, but you haven't reached your destination, so to speak.

"Value" isn't a property inherent to the copy on your website (or any text, for that matter), it's a property of the relationship between the copy and the reader. One way in which people recognize value is when they gently realize that they are (at least a little bit) wrong.

What are people wrong about re: API testing?

You currently have:

  > AI-Powered API Performance Testing
  
  > No-Code, Self-Learning!

None of this implies that I have a problem. And this is because if I am building a website, I already test my API.* So it doesn't matter that your product is automatic, no-code, and self-learning. Instead, you have to point out how my testing of my API is incomplete.

*Or so I may think. I might be convinced that I am testing my API even doing something as simple as '> rake routes'...

But users will likely respond to the following value propositions:

  > Most people don't even know they aren't testing their APIs properly...

  > Doing it right is actually quite tricky, due to... (e.g. effort, time, money)
  
  > By relying on unit tests, engineers miss the forest for the trees...
  
  > For example...

^^ Build this out first, and a snappy way of wording it, as well as a heading / tagline will follow.

The more you understand your true audience, the more you can "tune" this copy to specific issues they might have in their API test coverage. If I build a Rails site in a month by myself, am I likely to need your product? What about if I'm at the scale of Shopify? Or is your intended audience somewhere in between?

I used to have this line earlier. I changed it after criticism from the other users. I agree that your suggestions is valid to have a reason as well.

I'll add this line back. "According to Google poor API performance leads to negative user experience and high-churn. Deliver High-Performance APIs with our Self-Learning and No-Code platform."

I think you are better off with only the second sentence. The first sentence is something your users likely already intuit.

thanks

re: 3 - It may be true that google has a white paper describing the impact of api speed on retention. But beware, your users will need to have experienced this pain themselves, not theoretically, for them to be interested. They will need to have identified that latency (or churn - what is the actual problem you’re solving for users, that their api is slow out that they have churn?) is a problem that they want to solve right this second.

Make the value proposition extremely clear, and try to really help the user focus on a pain they have right now and that you can solve it.

I added this text "According to Google, poor API performance leads to negative user experience and high-churn. Deliver High-Performance APIs with our self-learning and no-code platform."

OP you're getting a lot of criticism in this thread, and 100% going through it will be the best thing you have ever done. People are doing God's work here and you are to be applauded as well for wanting this honesty in the first place. Good luck.

I agree the author should be applauded. That said, trying to please everybody on HN is impossible.

It would be more useful if these comments explained why they think things are important, so the OP can assess the importance of feedback within the context of their own site. Too many of the comments assume all websites should be built in the same way, and criticise anything that isn’t.

> Too many of the comments assume all websites should be built in the same way, and criticise anything that isn’t.

I think they're smart enough to ignore that or at least realize that a total rewrite is silly.

Thank you so much!

I'm going to respond to each and every comment.

Thank you so much!

- A very slow website is trying to sell me a product about optimizing my product. That immediately kills my first impression

- More specifically, the blocks takes ages to display after scrolling on mobile

- There is a horizontal scrollbar on mobile, which is annoying (Kiwi Browser)

- You should drop the fancy JS and make plain HTML pages, or it's going to hurt the performance and SEO.

- After quickly reading the webpage, I do not understand what this it about. How can a third party product make my hand crafted 10-endpoints REST API faster? Is it some kind of CDN? Or analysis tool to help navigate big APIs? (in the second case, you should clearly sell it as such, because it does not actually make anything faster by itself)

1. The website is horribly slow. We're new to Webflow. We'll work on improving our website performance.

2. We haven't completed mobile compatibility yet. We're going to work on it as well.

3. We assumed we're clearly saying what we do but it seems it's not clear enough. We're going to work on it.

Thank you so much for your feedback!

> We're new to Webflow.

Don't use Webflow. You really, really do not want a Webflow page to be the first impression of your product. Use something that actually builds to static HTML.

If we couldn't improve the load time and other performance issues we'll dump Webflow and rebuild the entire site again.

Thanks for the feedback!

You're probably looking for a "static site generator" :)

A good one should do all the work of adding your header and footer to every page, allowing you to re-use components, etc., but the result won't require JavaScript or anything like that unless you specifically opt to add it.

Kill the "appear on scroll" imo.

What it results in is scrolling, seeing a completely blank screen for a second or so while scrolling happens and updates are deferred, and then showing something. It could have been there all along.

We'll kill the appear on scroll added it to our task list.

Thank for your feedback!

On mobile the page is too wide so you can accidentally scroll right to an area with no content pretty easily.

Also on mobile the button below "Get a free assessment done by our expert performance engineering team." only shows the letter "S"

This is on iPhone 14 pro

The website is not ready for mobile. We're working on it.

Thank you so much for the feedback!

Too many animations. Probably loading in a lot of custom js. Honestly keeping it simple is probably the best bet it already looks pretty generic for these kinds of sites the animations don't really help. Often the goal for those is subtle which is why people spend a lot of money on ux designers to do just that. Overall it's fine but some stuff should probably just be removed. Your site isn't really accessible but that shouldn't matter for your userbase.

We're using Webflow with one of the premium templates. I'm not sure whether it's Webflow that is slow or the template is making the site slow. We just started the website a month ago. I'll take a deep look at the performance and experience.

Thank you so much!

The website is slow to do an initial load and when scrolling it takes longer than a second to load the next section. The website seems mostly static, (except for the PerfAI terminal) it shouldn't be this slow.

It looks like you're using Webflow, not sure if it's possible to make it faster.

Lighthouse on my MacBook Pro 13-inch 2020 i7:

https://imgur.com/a/dSPmea2

It is similar to a presentation deck, with plenty of empty whitespace, tiny low-contrast text, some Pioneer plaque-like graphics, and in the end the punchline is "Request a demo".

Just set a price, don't ask people to request a demo because they won't.

Some what I would suggest is:

- More contrast between the text and the background

- Less text

- Put the most meaningful text earlier

- Create "plans": basic, pro, enterprise. Keep the "Contact sales" button only for enterprise.

I agree. Always hate when I have to inquire.

The first thing is a prompt to enter a Swagger definition url. Without context, how do I know if I am singing up for DDoS as a service or something legit? It doesn't tell what it does but prompts me...

The page assumes that the reader knows what it's about. That I care about putting a swagger url, and get... something as the output. Tell me what's the goal, what's the value of the output.

I'll add what it's going to do.

Thanks for the feedback!

My advice is to take HN suggestions with a serious grain of salt.

I agree, some feedback I need to ignore but most seem helpful.

Thank you so much!

Why do I care that it is powered by AI? The copy seems smarmy and scammy.

If you built something that will make APIs with an OpenAPI spec faster, can you say that in a way that doesn't seem like you are selling snake oil?

I think they are using the AI to generate test data but the site is so obtuse I am not sure. When I hear performance my first thought is throughout, latency and such. If it were a search engine people might consider accuracy as a kind of “performance” and if it were selling stuff the effectiveness it has at that could also be considered “performance”. (e.g. a salesperson gets “performance-based” pay.)

So they are doing really bad at explaining themselves and it’s particularly dangerous in a field like A.I. that attracts so many bullshitters.

Yes, you're correct. AI is going to do things that otherwise an engineer had to do including:

1. Writing a custom Performance plan for the API

2. Learning how to invoke API endpoints

3. Applying Performance Top-10 coverage/categories, etc.

Thank you for your feedback!

what’s a “performance plan?”

It's Performance Testing Plan a simple report that includes total effort, cost and skills required to do a continuous performance testing.

AI is going to do things that otherwise an engineer had to do including:

1. Writing a custom Performance plan for the API 2. Learning how to invoke API endpoints 3. Applying Performance Top-10 coverage/categories, etc.

What I need to do is explain how AI is helpful.

But thank you for your feedback!

I'd say you need to start again. Make sure your content is solid first, then your layout, then your color scheme.

Only play around with JS and animating once you're sure it's absolutely needed. Most animated reveals are wishful attempts to look high tech that just irritate the visitor. If you're in doubt, just follow how a major company like stripe.com does it.

Thank you so much!

The copy needs an edit pass. Remove useless words and condense sentences.

What is the AI doing here?

Can I use it if I don’t have “swagger”? I don’t have swagger - I only have Typescript.

Who is providing the “AI”? Is this an LLM thing or like, your own deep learning model? Hard to say I want to give API access to random vendor.

Step 1 is “call us” - I would never bite.

* As others were saying “no-code” implies non-tech. I’d focus on “continuous” or “api performance testing on autopilot”.

* Marketing websites need to be mobile first. Your first impression will be on a phone even if the product isn’t usable there. The demo should be a one click on an example domain and you might even fake the formatting.

* Put the case study logos on the home page for social proof

* A much simpler and more useful solution to API testing is to scrape or record latency data from prod. With that data I don’t have to worry about authentication or making sure the automated tests are hitting a realistic account setup. Those are the hardest parts of doing synthetic performance testing and you expect the customer to do those for you.

Yes, we're fixing the mobile version.

We're still struggling between "No-Code", "Continuous", "Active" keywords to describe what we do.

Thank you so much for the feedback!

> Neglecting API performance results in poor UX and high customer churn. Deliver high-performance APIs to stop churn and boost 3x retention.

Remove the first sentence.

Also the viewport width is broken on mobile (chrome on android). I shouldn't be able to zoom in and out on mobile either.

Docs on the viewport meta tag: https://developer.mozilla.org/en-US/docs/Web/HTML/Viewport_m...

These are the most obvious things I noticed within 5 seconds and I stopped after this.

We'll follow the feedback.

Thank you so much!

> I shouldn't be able to zoom in and out on mobile either.

In fact you should be able to.

> user-scalable

> Controls whether zoom in and zoom out actions are allowed on the page. Valid values: 0, 1, yes, or no. Default: 1, which is the same as yes. Setting the value to 0, which is the same as no, is against Web Content Accessibility Guidelines (WCAG).

in proposal phase:

https://www.w3.org/WAI/standards-guidelines/act/rules/b4f0c3...

From a practical perspective, what makes this bad is the screen reader on a mobile device reading aria-labels for items that aren't inside the viewport, not to mention the fact that it's too easy to zoom unintentially either by browser configuration or accidental screen touches. Contrary to popular "wisdom", web browser zoom is one of the most anti-accessibility features out there, because there's no good way to detect page zoom from the Web API implementations of the most common browsers. What is more accessible is zoom on the entire screen buffer, not the web page. Both android and iOS support this (with focus change support) as well as many desktop environments.

thanks

Sure. I have Javascript blocked by default and all I see is a blank screen and the title in the browser bar "PerfAi".

So you are communicating exactly nothing to me about who you are and what you do.

You don't care enough to even throw an alt-text <noscript> there with some kind of "hey we're x, we do y, why don't you turn on the JS so we can show you, or alternatively, here's a link to a PDF brochure, or send us an email at example.com."

I've been known to unblock things I find interesting, but you're not even giving me a hint of motivation to do so.

Hey!

I like that it's relatively clean. No huge images or animations. It has a bit of Google feel to it with focus on central text field and limited distraction.

Being that simple, it should be blazingly fast. It appears it uses code to... make things slower? That's going to divide audience, to those who like cute effects and those who just want to get to it. Who's more likely to be your buyers?

It was not at ALL obvious to me that I should scroll down. Again, it evokes google, there's a search box, and four sections / buttons / numbers at bottom that seemed to indicate end of content/page.

Your pricing/faq structure is confusing to me. Pricing goes to FAQ, FAQ goes to pricing#FAQ. There's no pricing either way :D

I don't know if there's enough description right up front, or an easy enough link, to understand what it does / what's your sales pitch / why this is worth the time. But I like that you can try it right away with minimal hassle. Goes back to my first point though - if your audience are no-nonsense techies, reconsider the value of intentionally slowing / fading things in :)

FAQ letters are either somewhat low contrast/hard to read if cursor is not on them, or very hard / extremely low contrast to read if cursor is on them (which it typically would be if you click on it). Likely to be another pet peeve of the HN audience :)

I feel it is not expected behaviour for "Products" button to not do anything, only open your choice of specific product lines. My expectation would be that I can hit "Products" root button to see overview and comparison of your products. I don't want to chase them one by one - I don't have the information to make informed choice yet. Edit: Clicking those two links just scrolls me on the front page which is somewhat unexpected behaviour (it's structured to make me feel it's a separate page), but more importantly it's strangely slow. It pauses then scrolls then fades in. It feels like the automatic gearbox in older cars - nothing, nothing, oh wait did you press gas? Hmm should I do something about it? Oh! Yes! I'm a transmission! I should downshift! Yes! Here you go! ==* :->

With that in mind, the more I use the site, the more little delays accumulate. I feel like in 1998, netscape could instantly go to anchor inside the page. I feel like in 2023 on my Fibre gigabit, I'd expect clicking a button that does the same thing should be... like, faster than instantaneous :->

That being said, I cannot evaluate how good your product is - I did not interpret that to be the gist of your question, just the website itself.

Thank you so much for the feedback!

> According to Google

For the first three words of the first sentence on the page, this feels unnecessarily defensive. It diminishes what's being said as a fact by making the reader work out if they trust Google in this area. It almost sounds like you don't believe it.

If API performance is important, you should state it, not pass along a rumour. If the work done to show that is someone else's, cite that in a footer.

Thanks, let me think about it.

The cards look pretty rough at certain widths https://ibb.co/CmWWdj9 (946x8690 in this example) and you have a horizontal scrollbar at all viewport widths below 768px https://ibb.co/dj4hB4K

1. Completely broken with javascript disabled. Which is to say, completely broken.

2. When permitted javascript, it appears to use it to... intentionally fill in page contents slowly? It's hard to fully convey how obnoxious this is.

3. I assume you're aware that the "Pricing" link at the top goes to a page that contains no information whatsoever about pricing?

4. For the FAQ page (which is somehow also the pricing page, despite the lack of pricing?), having to manually open and close each question is cumbersome and annoying. Or rather, it would be only cumbersome and annoying, except that you have once again introduced the "intentionally slow down displaying content" javascript, which makes it outright infuriating.

5. Clicking on "Products" does absolutely nothing. Hovering over "Products" brings up some obnoxious little micro-lightbox bullshit.

6. Clicking on either of the items in the obnoxious little micro-lightbox bullshit takes you to... the front page of the site. The entire site appears to only have two pages, so why are you insistent on pretending that it's four?

7. The fact that one of the first problems that this site with wretched UX claims to be able to solve is "Poor UX" is some absolutely top shelf comedy. My compliments to whichever troll managed to sneak that joke in.

Piling on to #2, get rid of the fancy animations and fade ins. For a developer audience especially this is meaningless posturing that's impressing no one.

I liked it :)

did you like it so much that you initiated a consult?

I have no use for the service...

I'm a hard disagree here as a dev. The timing is too slow for me (i'd half or third the delay) and perhaps lower the transition duration a hair. But the effect itself was quite nice and gave me a strong sense of polish and quality.

I'm in-between you and the comment you replied to on whether or not to keep the animations or if they add value. I kinda don't care one way or the other.

However, I think the current implementation is a real problem. I'm using an old phone. I scroll down and there's nothing but white for a while until things start popping up. It's blank long enough for me to start wondering what's wrong. At that point, I'm not really thinking about the content.

We need to fix the mobile compatibility issues.

Thank you!

I like the animations but most folks here seems to hate it. Since we're using Webflow with a template it might not be that easy to remove animation.

We're going to improve the load time by half.

Thank you so much!

I agree. It needs to fade in much faster.

We'll do that. It's added to our task list.

Thank you so much!

I'd advise OP to keep #1 going: As a fledgling company you don't want to waste time time on the kind of person who sticks a 2x4 in their bike tire and then yells at you for not being more accomodating.

The light chasing affect before each page load is hopefully a bug (it's ok to show that once, but every page change gets old)

Overall clearly the site is incomplete, I think OP should go through the site on Chrome and Safari and clean up the broken links. I suspect OP wanted a brutal review of their core offering (ie we'd put in API links)

Unfortunately it's not easy to get people to do that OP: You're asking for API access without showing what the product does. I'd move your screenshot of a terminal up as a hero image (for inspiration: https://www.warp.dev/warp-ai) and hopefully you'll get a better conversion rate to actually trying it out.

Also maybe a creative "cheat" here, OpenAI plugins require publicly accessible APIs. Maybe instead of asking people for their own schema, offer reviews of the publically available plugin APIs (which there are a ton of)

Seconding this. Don't listen to HN nerd shit. Listen to paying customers. Nobody with a budget gives a fuck about "javascript disabled"

Their call on the first bit, but I can assure that that your last statement is at least off by one. My budget for this year was eight figures, and extends to things like what this product appears to be.

And while I'll admit that I wouldn't make a purchasing decision based exclusively on javascript fuckery, it definitely does mean that my first impression is to doubt their technical judgment.

Why should a website with static content need JavaScript to display?

I didn't say it should. I'm saying don't worry about it if it doesn't matter to your customers.

Customers definitely care whether a site loads instantly or in 10 seconds.

That has very little to do with having javascript on your page.

It has everything to do. Sites without JavaScript load fast unless you really screw something up.

thanks

Thank you so much!

1-2. We're using Webflow and it seems these guys required jQuery and similar javascript libraries. 3. Pricing link is broken we didn't realize it. 4. Agree, we'll change FAQs to full view. 5. Yup, it only shows on hover. We can change it to click I think that's much better. 6. I thought having important links in the nav-bar makes it easier for the users. Anyways I'll take another look and fix it. 7. Lol!

Thank you so much for the feedback! It seems lot to improve!

Mobile site scrolling horizontally within the viewport, ie; scrolling is broken for mobile break-points.

The second section, just below the fold, was clearly designed by an incompetent idiot, as it renders badly on mobile. Same goes for section entitled How To Get Started. Comparison is misspelled. Section renders badly. Fire your designer, preferably into the sun.

lol...I take the blame we never did a website in ages.

Thanks for reviewing it!

1. Get rid of the JavaScript. All of it. If you need JavaScript, you're doing it wrong.

2. Information density is way too low, the entire page could fit on one 1080p monitor.

3. I don't know what I'm even looking at immediately after opening the page. What is the information you want to convey? Put it front and center on the page, make it the first thing I see. That interactive text box is useless to me, because I don't know what anything is.

4. Gray text on white background is an insult to peoples' eyes. Use black text on white for maximum readability. You're trying to convey information here, this is not an art piece for your high school project.

It doesn't load at all on Android Chrome. Just completely empty scrollable page. When refreshing, something blinks for 1 frame and then it goes back to blank page. Only after disabling RethinkDNS adblocker it loads.

Can you please answer these questions: What am I selling and how is it helpful?

Beats me. I lost focus after about 10 seconds of scrolling and the multiple animations.

At which point I bailed.

Appears to be something that will make my APIs better, presumably by load testing them.

Unclear how AI helps here. I'd rather know what the product does and why it is better than scripting curl or something like that.

I scrolled on mobile and took 3 seconds to appear. Awful ;)

It takes forever to load. I'd scroll down and think I'm at the end only for something else to pop up like 5 seconds later

The screenshots are tiny and show nothing of use. You could either remove them, or give them more space. Lose the window chrome/buttons, Mac screenshots are wasteful; just focus on the content.

Entering a long URL is cut off by the blue button.

Ensure that 169.254.x.x type URLs aren't being processed by your checker first.

Overall I think I like it, it does explain well what it does.

The FAQ section is seriously lacking color contrast. If you want people to read paragraphs of text, make it actually readable.

There is a lot of text in the FAQs. We'll shorten it.

Thanks for the feedback!

CTA is below the fold on mobile. Company name appears twice in large font on mobile. Hamburger menu animation time is too slow.

> CTA is below the fold on mobile

Device screen aspect ratios are not consistent, especially on mobile. This is a very annoying problem for many websites. There is a simple solution though: make the entire masthead section square. Cut the gordian knot.

Two questions

1) Why should I care?

2) What are you going to do for *me*?

Features < Benefits

Thank you so much!

I just wanted to thank everyone again for your valuable feedback. We have addressed several issues, but I believe the most critical ones are the mobile fixes and improving the load time.

I personally upvoted, commented, and thanked each of you individually for your feedback.

To everyone - Thank you so much!

No-coding means it's focusing on not techies, then interface is 90's hacking terminal.

It's not consistent

Yeah, you're write. I need to think how to say "Platform is AI-powered it self-learns and it requires you to write no code or no config" at the same time make it appeal to the techies.

Thanks for the feedback!!

“You’re right” not “you’re write”

“The platform is AI-powered and self learns”, not “Platform is AI-powered it self learns”

Others have pointed out that it needs trimming down, but I think it’s just clumsy all the way through. I’d recommend getting a copy editor to do a pass on it. What you’ve got so far is a great starting point for them because you’ve already decided what you want to say and where, and they’ll be able to reduce the copy, make things clearer, make sentences read and parse more easily, and reduce that clumsiness.

Also be careful when copying the new text back into the site, I saw a few errors that look like they’re from copy/pasting out of a document into a site backend.

Also for the love of god ditch webflow.

My only comment is you have a lot of courage. This is admiral. I am terrified of doing the same.

I believe this exercise really helped us a lot.

The website doesn't respect prefersReducedMotion, making it extremely annoying to navigate.

This is my attempt at your tag-line/call to action:

API Performance Monitoring: Detect high-latency APIs.

Speed Matters

Two small remarks:

  * In the pricing page, you might want to remove the `.00`, to make prices easier to read
  * In the demo, there is a typo in the 'console' output: `defnitions`.

For me on mobile it was scrolled halfway down the page when it loaded. Agree with the other commenters that the slow-loading content is really annoying and caused me to lose interest very quickly.

1. Ai based load performance testing.

2. Slow websites stop users from using the system.

Thank you so much!

The tiny screenshots in the bottom of the page should be right at the top and bigger showing what this product can provide. Let your product do the selling.

Reply on news.ycombinator.com