YC SUS: Aaron Epstein and Eric Migicovsky give website feedback
These notes were generated by AI and may contain inaccuracies.
Aaron Epstein went through YC in winter 2010 as co-founder of Creative Market, a marketplace for graphic design assets where designers sell fonts, Photoshop files, stock photos, and website themes. Now back at YC as a visiting partner to help startups with design feedback.
Two hours of Aaron's time to review websites fullscreen from zero knowledge, like usertesting.com with YC partners.
Hero Section: 'Grow your search traffic and outrank your competitors with just 20 minutes of work per day' – sounds like a lot of work, targets experienced SEO pain too early. Sub-headline: 'Instead of getting lost in endless reports and spreadsheets, SEO Buddy turns the daunting world of SEO into easy to complete daily missions.' Unclear what daily missions are. Good free trial CTA with orange buttons, fire emoji, '111 companies signed up in the last month' for social proof and urgency (like Airbnb/Booking.com stats boosting conversions). Gray subtext hard to read; shorten header, enlarge subtext.
Social Proof: Unclear if news clippings or companies; unknown names don't build trust, could be friends' companies. Suggest putting YC/Startup School affiliation on main page.
Negatives: 'SEO and rocket science but it's not easy to be successful' – too negative; flip to 'SEO is difficult, we make it easy,' end positive. Halfway down page, still unclear what differentiates SEO Buddy in crowded space.
Recommendation: Frame as 'what you're probably doing today' vs. shorter SEO Buddy process. Show 'world before/after SEO Buddy.'
Product Description: All-in-one SEO productivity tool: grow search traffic, outsmart competitors, control SEO strategy with backlink monitoring, competitor research, content manager, outreach tool. Screenshots good; move higher up. Crowded; focus on top 1-2-3 features and benefits, not features (e.g., don't say 'track backlinks,' say superpowers like 'rank higher, figure out what's working').
General: Strong, straightforward. Bottom 'Talk to me' CTA good; experiment with auto-pop up (no audio). Humanizes with founder story.
Hero: 'Computus, the world's best crypto tax tools' for crypto holders filing taxes. 'Find out what we mean' links to email – tell directly, don't make click. Bare-bones, no image, lacks trust for serious tax/portfolio tool. Copied Bootstrap template poorly.
Features: Generate tax documents, monitor expected tax obligations, make sense of crypto investments (be more specific).
Trust Building: Add logos (even unknown), recent signups, founder face/story like SEO Buddy. For landing page collecting interest/consultations, need logo, product images for legitimacy.
Goal Clarity: Drive to one goal at top (e.g., schedule consultation/email), not buried 'learn more' PDF – put content on site.
Advice: Website hurts credibility for user research; use warm intros or Webflow template for quick professionalism instead.
Mobile app for creative people fond of arts and crafts: discover trends, local events, introduce point of view, share feelings, meet soulmate. Buried lede – sounds like dating app? Vague terms ('join a mob,' 'stop the world from being regular,' 'mutual friend... truly your folks'). Imagery confusing; unclear if for in-person meetups around art or celebrity opinions/reviews. App screenshots in Russian; if Russian app, no need for English site. Actionable: Pull app screenshots higher for tangibility.
Hero: 'State-of-the-art audio and speech synthesis powered by machine learning.' Demo sounds robotic, not state-of-the-art (vs. Alexa/Google). Unclear use cases/API/integration.
Features: Cost-effective, 20 languages, text-to-speech (emphasize words), character voices for apps/PA. Tune in to Deventer demo. For developers (SDK?), but lacks how-to-start info.
Differentiation: Many speech-to-text APIs (iOS/Android built-in); highlight unique advantages. Pricing buried; call out. 'Start now' vs. 'contact sales' unclear – developers want quick SDK test/web interface like early Stripe.
Design: Iconography friendly but doesn't communicate product.
Positives/Recs: Unique voices/diagram. Pull editor/use cases to top. Drop 'powered by AI/machine learning' – focus benefits (auto-speak any sentence in 20 languages). Talk to users on attractions.
Hero: 'Welcome to the wonderful world of upcycling' – explains how to turn waste into profit, zero to market-ready product.
How It Works: Tell materials to upcycle, find best designer, agree timeline/price, provide waste, get upcycled products, sell via marketplace. Shop hundreds of upcycled products.
Issues: Unclear who for (everyone?); focus on max benefit users (e.g., companies with bulk waste like bottles → artists → sell). Examples resonate (recycled into products). T-Mobile (major label) case.
Structure Recs: Walk through full customer example (e.g., T-Mobile waste → artist → product → sold). Call out 'why/for who' clearly. Homepage steps don't explain. Big text blocks ignored; highlight key bits via examples.
Case Studies: Charity auction, nonprofit angle ok. Great photos/assets, but poor organization.
Business Model Guess: Partner with brands (T-Mobile/McDonald's) for bulk waste (free, they get brand value), match designers, sell cool products in marketplace.
Home Page Recs: Lead with marketplace/e-commerce of cool products (most visitors are buyers), add recycled/big-brand stories for value. Most won't be brands. Uses e-commerce platform (maybe Shopify); love familiar ones like Kickstarter/Shopify.
Users have mental models of how an e-commerce site should work or how a checkout process should work based on past experiences. The guest button is really small, and entering an email to sign out as guest is confusing. Map to what people are already used to doing so the product feels familiar and they won't have to relearn a whole new set of experiences.
Breezy: Quickly help you find the best and most affordable insurance for your pets. Clear, but the subscribe button is unclear—what am I subscribing to? Expect 'put in information about your pet and we'll give you quotes' or 'get a quote'. It's a non-launched landing page to collect interest. Pet insurance calculator—clarify it's health insurance. Personal story: Dog dislocated her leg jumping out of car window, cost way too much; pet insurance would have been nice.
Calculator for after pet accident: Type in vet bill. Too many options. Provide scenario example: 'This person came to us, their pet had this accident, we covered them, this is how much they paid.' Walk through specific example.
Mapping options to insurance terms like deductible, reimbursement, annual reimbursement limit confuses users. Users think: 'My dog dislocated her leg, $3,000 surgery, how much do I pay on each plan?' Frame in user mindset—simplify deductibles, reimbursement, limits. Add lower nav and footer bars for semblance of realness. Mental model for webpages missing key components. Add about page or section with personal story about massive bill saved by pet insurance. Minor design issue: Tab doesn't align to bottom of calculator. Image is good.
Pets food Kyle screams pets taxonomy tracks tracks anime business solutions reimagined. 'Rita match'—this is the place for me. Shrek tricks trips anomie offers consulting services with tools and XP. Way too much jargon, no idea what this is. Even scrolling down, no clarity. Stock photos come across as cheesy.
Get to the meat: Specialty? Clients worked with? How helped them? If no clients, past work experiences that drove consulting. Ideal customer—speak to them in header. Nobody searches 'business solutions reimagined'. Sites like this support in-person sales: Google company name, good .com. Frame for in-person follow-up, e.g., business solutions for transportation companies. Tiny point: Enable SSL—expected in 2020, Wix does it by default. About Us section lacks 'about you'.
Base site off competitors or others in same business using their elements. Friends in space can help.
URL not day one day I, throwing off. Day one day I make your app marketing campaign effective day one. Sign up for beta. Missing one-liner. Best one-liners identify who it's for and benefits. Getting it's for app makers wanting effective marketing campaigns, but 'day one' confusing—first day working on app or launch day? Does it take time to ramp up?
Gray text on white background, super small: Reduces acquisition costs by understanding user behavior on first day of app marketing. Also fights bots, detects fraud traffic. Main criteria: unique behavioral patterns. Unclear—top sounds like making campaigns effective day one (advertising), but now fraud detection for ads. Fraud harms budget, saves costs. Product might be unfocused, spilling into website.
Talk to users: How would they describe product? Use their words at top with screenshot or visual. No more SDKs—use raw data from popular analytics. Plug into Segment or Mixpanel API app developers have: 'We take your Mixpanel and Segment data and turn into actionable marketing advice.' Too many fields in form—put email in main header.
Connect common denominator future-proof local and global decisions on water. Local predictive analytics to plan, produce, protect. Multiple one-liners; best: 'Local water affects your global'—says what's happening. Kill others, make it bigger, center it. Predict local water supply? Why care? Track/predict problems affecting business.
Figure out who speaking to, use case—talk specifically. People make it broad to sound big, but sounds vague. 'Learn more' is mailto link—doesn't help. Have offline conversations, sense what resonates, common questions. Put answers on website top: Who for, what use for. Not to get emails. Unclear CTA: Email? Pay?
.org domain suggests nonprofit, especially water topics, but 'inform investors, plan future' not. Clarify: Switch domain or say 'enable investors better decisions on water resource allocation'.
Lots going on: Moving backgrounds, random alignment boxes, different fonts/colors/sizes—hurts hierarchy, credibility. Email form white on light blue, hard to read—use black. Button 'submit', not 'learn more' or 'sign up'. Animated graphics modern, but offset by text.
More new and repeat customers: Smart marketing system for restaurants automatically markets to customers based on engagement with brand. Subtitle better than title: 'with your brand'. Leads to questions: How? How engaged? Clear for restaurants.
Build customer database—'have data we can use it, no data relax we've got data'. Restaurant owner doesn't know databases/data. Complicated Wi-Fi? Tell how get more customers. Text on autopilot: Texting drives incremental visits. Product captures phone numbers (e.g., waiting list), auto-texts specials/promotions. Bring new customers—how? Referrals, prompt Yelp reviews?
Everything vague. Cut 'Boost Lee's customer acquisition tools'—what are they? Piece together from headlines/screenshots, could be off. Phone number up top suggests legit. Work from example: 'Local Joe's restaurant used product to double sales', walk through 1-2-3 steps. Benefits: Similar restaurants filter in, see relevance. Answer crisply: What is this? For me? Who for? Be specific, not broad.
Blinking thing left: Irving Texas linking—small, turns off quick, more noise. If real customers, say '35 customers using product', don't hide.
Protect ears in style. Hearing protection. Beautiful image of yelling, see earplugs. Earplugs reinvented: Loop reduces volume 20 dB without sacrificing audio clarity. dB logarithmic scale—even engineers barely know. For music events/shows to listen without ear damage.
Images don't match: Not at show, high-class vibe, rose gold, build quality, music. Wired article link not clickable—missed opportunity. Team has brand vision matching product design, but not use cases. Expect festival imagery vs. goofy orange earplugs. Imagery should reflect common use cases.
'Technology' worries—implies electronics/charging, but shape/form only. Clarify not electronic headphones, can't listen to music. Price $30, cheaper than expected for high-end brand.
Reviews sell it's real company. Store easy, different styles. Biggest: Imagery for specific use cases. FAQ link goes nowhere—not worth front and center.
Embed Google form in your website, embed inquiry form, make Google Forms look professional. Say 'forms' not 'bed feedback form'. Used product: Takes Google Forms, adds CSS styling to not look classic. No form on own site—missed opportunity. 'Try it free' button too small.
Best: Before/after side-by-side—recognize default ugly purple Google form vs. professional embed. Pull into header, bigger text: What it is, why use. Simple feature—test with 10-20 Startup School founders (everyone uses Google Forms, not $40 Typeform).
Get rid of rotating text. Just: 'Embed Google Forms in your website'. Instantly know if for you. Actually embeds (not just CSS), no Google URL. Think of as better Typeform ($50, scammy; Superhuman uses, others follow). Three features: Make form look like website, no Google URL, embedded.
Three important features: make the form look like your website, don't use a Google URL, use your own URL. The section above is a wall of text with small font nobody reads. "This should be your one yeah you have got more customers" doesn't convince they're your customers. Like the line "make Google Forms look great." Use as header: "directly embed Google Forms into your website so it matches your native look and feel and keeps your standard URL."
Pricing Suggestions
Pricing says "start your 14-day trial today." Site seemed free, no actions required. Simplify to free and paid instead of three options. Can embed Google Forms via iframe but can't style it, so this is a styled wrapper.
Outdated recruitment: keyword matching and work experience filters don't work in 2020. Tells about industry/competitors but nothing about Crawl. Jumps to detailed product points before explaining who it's for or what it does. "Enrich your CV portfolio" suggests for job seekers, but overlapping text confuses. "Let the jobs come to you" not clear from start. "Instead of hacking focus on building evaluate" confusing. Public profiles, evaluating work history/backgrounds. "Crawl all the job openings" sounds company-side. With NLP, matches to fitting jobs, handles HR process. Read most of page without knowing what it is. Also for developers and website builder.
Advice: Test with Others
Share with smart people/friends/family to see if they understand. Hard to put yourself in newbie mindset since you know it deeply. Identify missing gaps.
Improvements
It's an about me page for developers. Lead with screenshot of this as homepage. CTA: "build your resume web page with one click." Syncs with GitHub, LinkedIn, etc. Speak to developers at top. Website builder page better: "create your custom one-pager in seconds." Switch to this as main page.
"Office spaces here virtual and remote now we're from home just like you do so from office." English issues. Virtual office vs. Zoom/Slack unclear. Industry talk not helpful; focus on your company/why use you. "Single click videos seamlessly" - single click video resonates. Video link loops to site with no video. Manage remote employees seamlessly, collaborate with single clicks - vague, hard to visualize vs. current tools. Blurry screenshot interesting: team screen share/video while chatting. Highlight demo earlier, bigger screenshots. Be specific on features/benefits vs. current. Ask customers (e.g., in India) how they use/benefit, use their words.
Popup for newsletter blocks content. Food delivery/shopping site. No logo, name, description, location (Bay Area relevant?). Frame problem: Instacart for halal products, quick halal delivery, products other services won't bring (ethnic/halal). MVP with many products hard to test; start with one (e.g., fresh meats) for better signal. Too scattered: meats vs. physical products, lunch now vs. non-urgent. Shop buttons unclear (different products? Use filters).
"Repast the headlines your favorite articles and stories 100% distraction free." Unclear vs. RSS feeds, readability (browser-built). Gamification: points, winners, polls. Like Hacker News/HN-style for crowd-sourced articles. Headline suggestion: "find the most read or talked about articles on the Internet today." Embrace founders' perspective: "we started it because we wanted articles about X Y Z in better format." About: social reading platform to find/read/share articles/stories. Community-focused; founders active in comments/emails. Like social network for reading (see friends' articles/time spent). Why switch from Twitter/FB? Pull from other channels? Multiple value props: distraction-free, sharing, commenting, popular examples. Focus on what engaged users care about most. Can't see product without signup; show examples. Forces reading article before commenting (solves HN problem).
"Step into the life of these real heroines innate decisions and pave your own path." Sounds like game. Beta tester. Needs more info: what, who for, why use. Game? Historical/education for Indonesian history? Embed info: e.g., left side "game for young kids to learn Indonesian history," right 1-2 big screenshots (not 9 blurry/fast-scrolling). Copy like in-person pitch sequence: name, play as heroine from past, learn history. Twitter bio better: "empower women to reach full potential through gamification." Product: interactive storytelling app for girls, play real female role models from country, show possibilities of successful women. Put on homepage with download link.
"Record and send personalized reminders to share with aging loved ones." "Reminders" confuses vs. messages. "Want to check in but calling every day tough" - checking in is receiving status, not pushing. Use cases: remind mom heart meds post-breakfast, note to dad. How received? (iPad? Alexa? Automated call?) Users exist. Testimonial too feelings-focused, not specifics. Switchboard implies phone system (no learning). Bury key info under headers; people skip body copy. Put important in large/bold text. Instead of icons: 1. Record message, 2. Remind parent meds/schedule 9am daily, 3. Grandma gets call with your voice/custom message + summary report. Shows easy vs. telling "easy to use." Thrown into Typeform ($45+/mo); consider free Google Forms. Useful product; clarify opportunity/help.
Build websites/marketplaces/web apps without code using building blocks, 10x faster/cheaper. Overloads: for outsourcers? First-timers? Simplify: websites (marketplace/web app subtypes). "Most intuitive website builder for startups" - specify type, differentiate vs. Webflow/Wix/Squarespace. Haven't launched; cut features (email, integrations) for core use case to launch faster. Example: LaunchRocket - quick value for startups. Site built on it; for startup founders' landing pages. Show visual blocks/components. Editable demo: "try changing price, move fields" below. Solid design.
Lunch box delivery for school kids, India. Depends on country. India's system: people cook lunch at home, delivery people bring to offices downtown (leave early before cooking done). Incredible logistics.
It's cool if you're in Mumbai you can go to the train station and see people carrying them. This service exists in the context of the Dabbawala system. Lunch box delivery for school kids in India immediately lets people from India understand it. Almost everyone with Parsloe Jr. guarantees default comfort, fresh food, good health. It says you get to prepare the lunch. Do you cook the food or not? Are you the mom or the kid? Who's the person on this website? It's for a parent. They do delivery but you cook the food and put it in your child's lunchbox. It's the same system as for business people but for kids. Why can't you use the existing system? From a business perspective, advantages of delivering to schools: less routing because fixed locations for lots of people. Cool, like the WhatsApp thing—super key. Include a link that deep links into a WhatsApp conversation. Some graphics could have photos, especially at the top instead of balloons. Show someone in an autorickshaw delivering lunch to school. That resumes it. Start the call to action with WhatsApp. That's the right way in India.
Achieve your happiness goal, connect with amazing activities, enjoy your life. Sounds great but no idea what it is or if it's for me. What would I use? Be a better you. Very aspirational but way too broad. Focus in that block attacks but attentional. Everything starts big starts small. Changes your happiness starts with little happy moments. Apply that to your website: start with the small thing rather than the big thing. What are the small changes you help with that lead to big changes? Founders have big vision but embarrassed to share first step. Counterintuitive: better to share the small tangible thing people understand. Don't be embarrassed—make sure people understand the first thing before larger vision.
Look at the app: goal tracker. Words goal tracker not even on YAP site. Make your goals reality: create, visualize, achieve, share, find new interests. Say on website it's a goal tracker where you set goals, share them, achieve. Doesn't say it's an app unless you look at buttons. Explanation on app page better than webpage. Start by explaining app above the fold, then larger vision. These are links: good Instagram, aspirational goals.
Recommendation: start from explaining the app.
From idea to venture in a single app for entrepreneurs, freelancers, creative types. What do you help organize? Hard to see from screenshot—problems with screenshots: hard to understand levels of detail. Trick: zoom in on one specific part or create blocky version with gray text blocks to mask unimportant parts, focus on key. Best: screenshots as supporting evidence; description communicates what you do and why use it. Draws attention to picture over block of text. Too much: combining communication, collaboration, bookmarking. What's the one thing people use it for? Focus on that. Too much text, don't like white on black—feels too dark. URL tough to say verbally, e.g., explaining email: Eric at J-I-N-T-J dot com. Got lost at beginning.
Super theoretical rather than specific what app does to help today. You've built the app, have users—ask them the value they're getting. Growth Channel: AI powered digital ads for FinTech startups.
Get unlimited Forex and crypto ad campaigns for up to 12 platforms on small budget. Do you create campaigns, imagery, content, keywords? No payroll? Hourly billing. Activate your future. Help Forex brokers, funds, crypto exchanges save time, save money. What do you help them do? Literally what does product do? Help run digital ads. Unclear what 'run' means: buy ads, target, design? Too many items not telling positively what you do. How it works: one, two, three. Seems like marketing firm. Why FinTech? Something unique: domain experience, ran 300 campaigns for crypto exchanges, know it inside out. Lifeblood of FinTech: profitably acquire users. Start with that—why choose you over other ad agency?
Explicit: run ads on these 12 ad networks. Why better? Results, cheaper? Case study of customers. Do you have 10 years experience? Why bury in tiny text at bottom? Location important? Pay per click. Simplification needed. Struggling with who's it for and how helps.
Your online space for doing things together. What things? Who's doing them? Click 'try habittr'—it's an app. No description, missing key components. Link at top makes people leave page—don't know below fold. Counter: blinking arrow at bottom to scroll, or no top link off-page. Lots of 'try habittr now' buttons. Not clear it's app until App Store. iPhone photo but links to Google Play—say iOS available. Why try? Guess: set goals, invite friend for accountability. Event is deadline by when to do something, e.g., workout, learn JavaScript (different things). Text hard to read. Chat with avatar, update progress, share experience. Achieving goals together, keep accountable—say that at top. Help visualize what doing together and who for.
Show screenshots good but unclear. Join event that sparks interest. Don't need five screenshots—one or two. If users, find what they're using for, focus as main use case, find similar people. If no users, go to friends who workout or learn JavaScript.
Professional resume for influencers—like LinkedIn for influencers. Static screenshots vs GIF help. Show example or link to it, iframe. Control excitable to brands. Burying brand reviews and metrics—more exciting than sentence. Brand collaboration history, reviews from brands. Auto-rotation carousel: click or auto-roll. People don't click tabs—consider hidden. First two tabs more important than fans. Try it free—build page for free by linking Instagram/Twitter, see in 10 seconds. Makes it feel quick. Everybody create free page, link people; pay for brand features. More landing page than entry point. Create pages for top influencers, message them: built page for you from Instagram hashtags/ads. Link recognizable ones above fold—seems legit users (hack).
Helping find best funding opportunities, connect peers, learn field. Too much above fold: icons, funds hub, aims (brain aims?). Too much 'brain' play—use words users know like project. Still unclear: what's funds hub, aims? Prove you help get funded? Design with lots of colors, fonts, layout inconsistency reduces trust. Over-communicating—too much info, noise since didn't click at start. Start with most important thing, ensure resonates before other stuff. Paradox of choice: more options, longer decision. Get rid of everything, focus on one thing users do.
Let customers know what's up, get instant feedback. What does 'what's up' mean? Be specific: downtime, new feature? Like top bar 'recently updated' or 'what's new'. Embed your product on site as example with arrow. Change log—say 'add change log to site with one click'. Does it grab from Git check-ins? Dashboard unexpected. Feature to prompt nicer change log copy like Slack iOS. Who made it happen—cool, humanizes with faces. Let users know new update, share who worked on it, get feedback if liked.
Everyone uses Slack—'get Slack update system on your site'. Feedback interesting: quick loop vs monitoring Twitter/email. Switch order: feedback before dashboard to avoid confusion. Headline for 'learn from customers'? Dashboard with real-time feedback on changes. Faces make think simple rating (thumbs up/down), not open comment. Built a lot but page looks pre-launch. If out in week, say so. Don't overbuild—launch with just the box as v1. Every Startup School SaaS could use. Add about us, downloadable.
Smartest way to share any content anywhere, track results—too broad. What content? Enterprise: case studies how Siemens, Freshworks use. 'Content everything platform' means nothing—be specific to someone. Break cluttered inboxes, beautiful experience for sales emails, social, campaigns. Email builder? Track what happens after send—not just emails, attachments. Track attachment opens for salespeople sharing PDFs/PowerPoints. Host it. Put above fold: track attachment open rates, close sales faster. For salespeople. 'Send better attachments'. Why book demo to send better attachments? Pricing: $30/$50—not demo-worthy, too low for enterprise (purchase order annoying). Charge more if valuable, skip demos/sales process.
Docs and alternatives SEO effective. DocSend alternative—explain why better, simplify for immediate understanding: relevant if close sales over email.
Update one-liner on website from Thursday sessions feedback—don't let site fall out of sync with pitch practice. Most important: copy. If can't understand product, nothing else matters. Answer: what is it (literally, not big vision), who for, build trust, convince solves problems/valuable. Ask trusted smart people to review critically: who site for, who benefits, what product? Sit with ideal users, watch them use site/product, see trip-ups. If stumped, use UserTesting ($10-20) for random video sessions.
It's 11 p.m. at night and you just want someone in the world to look at your site. You can get pretty quick results with that website.
Another tool to try is install FullStory on your site. It lets you replay sessions of users using your product so you can see how people scroll. You can track the mouse to see where people's eyes are looking—people tend to do that because they browse around. You can figure out where people get stuck, what links they click on, do they make it to the bottom of your page or not, what buttons they focus on. It's a great way to learn.
Remember that site where it explained part of it on the second segment but then we had skipped down past it? You could actually learn if people do that by default on your site. Maybe we were the weirdos but maybe everybody does it, maybe not.
Thanks very much Sharon. Really appreciate you and we'll see you next week. Well not you but we'll see everyone on some surf school next week. Thanks.
Keep Y Combinator in your library
Save the videos and channels worth coming back to, and find them again in one place.
