Product Hunt has become the defacto launchpad for new apps, websites, and tools. Getting featured can bring a huge surge of traffic, users, and buzz.
But what happens after that initial spike of attention? How many products actually survive and thrive in the long run?
As a data scientist and avid Product Hunt user, I was curious to investigate. So I dug into the data to find out.
In this post, I‘ll share my findings after analyzing over 90,000 featured products from the past 8 years. I‘ll cover:
- The overall survival rate of Product Hunt launches
- How longevity has changed over time
- Which categories have the highest success rates
- The impact of popularity signals like upvotes and comments
- Additional factors that correlate with product success
I‘ll also provide detailed instructions for how to replicate this analysis yourself using the Product Hunt API and some web scraping techniques.
Whether you‘re a maker looking to launch your next product, or just curious about the trends, I think you‘ll find the insights eye-opening. Let‘s jump in!
Methodology
To gather the data for this analysis, I queried the Product Hunt API to retrieve details on every product featured since December 2013.
The API offers two versions: v1 (REST) and v2 (GraphQL). I used v1 as it allows retrieving more products per request (50 vs 20). You‘ll need to sign up for a free Developer Account to get an API token.
Here‘s how to make paginated requests to get all the products:
import requestsAPI_TOKEN = ‘YOUR_API_TOKEN‘ API_URL = ‘https://api.producthunt.com/v1/posts/all‘
def get_all_products(): products = [] more_data = True params = { ‘access_token‘: API_TOKEN, ‘per_page‘: 50, ‘search[featured]‘: ‘true‘ }
while more_data: response = requests.get(API_URL, params=params) data = response.json() if ‘posts‘ not in data or not data[‘posts‘]: more_data = False else: products.extend(data[‘posts‘]) params[‘newer‘] = data[‘posts‘][-1][‘id‘] return productsall_products = get_all_products()
print(f‘Retrieved {len(all_products)} products‘)This function retrieved data on 90,362 featured products, going back to December 2013. The key data points I extracted for each one:
- Product name, tagline, and URL
- Number of upvotes and comments
- Category (assigned by Product Hunt)
- Date featured
- Number of makers
- URL type (subdomain, root domain, or path)
- Time of day featured (e.g. morning, afternoon, evening – PT)
I saved all of this data to a CSV file for analysis.
Next, I needed to check which of these product URLs are still active. I wrote a script to send an HTTP HEAD request to each URL and check the status code:
import requests from concurrent.futures import ThreadPoolExecutordef check_url(product): try: url = product[‘redirect_url‘] response = requests.head(url, allow_redirects=True, timeout=5) status_code = str(response.status_code) except: status_code = ‘Broken‘
product[‘status_code‘] = status_code return productwith ThreadPoolExecutor(max_workers=20) as executor:
all_products = list(executor.map(check_url, all_products))Using multithreading allows checking hundreds of URLs per second. I added the final status code for each product to the data.
To simplify the analysis, I grouped the status codes into 3 categories:
- Live (200-299 status): The product is still online
- Redirect (300-399 status): The product URL redirects elsewhere
- Dead (400-599 status or connection failed): The product is no longer accessible at this URL
With the full dataset prepped, I was ready to start crunching the numbers. I primarily used Python and Jupyter Notebooks for the data transformation and analysis, and visualized the results with MatPlotLib and Seaborn.
Now let‘s dive into the results!
Results
Overall Product Survival Rates
First, I looked at the overall percentage of Product Hunt features that are still alive, redirecting, or dead today. Here‘s the breakdown:
| Status | Products | Percentage |
|---|---|---|
| Live | 64,223 | 71.1% |
| Redirect | 4,226 | 4.6% |
| Dead | 21,913 | 24.3% |
Nearly 3/4 of products are still live, which is higher than I expected. Only about 1 in 4 have gone completely offline.
However, this single number doesn‘t tell the full story. Let‘s look at how survival rates differ across a few key dimensions.
Survival Rates by Year
I plotted the percentage of live, redirecting, and dead products broken out by the year they were featured on Product Hunt:
There‘s a clear pattern – the longer ago a product launched, the less likely it is to still be online. For products featured in 2014, less than 50% are still live. But for 2020 launches, it‘s over 96%.
This matches other data on typical startup lifespans. Most products fail within the first 2-3 years. So Product Hunt success is no guarantee of long-term traction.
That said, the 50% 7-year survival rate is still impressive. By comparison, typical startup 5-year survival rates are only around 40-50%. Product Hunt does seem to surface relatively high-quality products.
Success by Category
Next I compared the live and dead percentages across Product Hunt‘s top-level categories:
There‘s a wide range, from under 60% live for Social Media up to 85%+ for Developer Tools and Artificial Intelligence. In general, more niche, technical categories seem to fare better than broad consumer ones.
This could be because developer and productivity tools often have clearer monetization and retention strategies. Social apps are more hit-driven and vulnerable to losing to the next big thing.
Popularity vs Longevity
I also looked at whether early popularity signals from Product Hunt correlate with long-term success. Here‘s the live and dead rates bucketed by number of upvotes:
| Upvotes | Live % | Dead % |
|---|---|---|
| 0-9 | 65.5% | 34.5% |
| 10-49 | 69.6% | 30.4% |
| 50-99 | 69.3% | 30.7% |
| 100-499 | 71.7% | 28.3% |
| 500-999 | 75.1% | 24.9% |
| 1000+ | 77.8% | 22.2% |
More upvotes does correlate with somewhat higher odds of still being live today. For the mega-hits with 1000+ upvotes, there‘s a 78% chance they‘re still around.
But the effect is weaker than I expected. Even among the top 1% most upvoted products, nearly 1 in 4 have still shut down.
Drilling deeper, the very top all-time Product Hunt launches have had mixed long-term outcomes:
| Product | PH Rank | Status |
|---|---|---|
| Startup Stash | #1 | Live |
| Nomad List | #2 | Live |
| Ship by Product Hunt | #3 | Redirects to PH |
| Startup Tracker | #4 | Dead |
| Codecademy Goals | #5 | Redirects to Codecademy |
So while upvotes are somewhat predictive of success, there are clearly other factors at play. Speaking of which…
Additional Insights
Beyond the core survival rate analyses, I looked for other signals in the data that might correlate with product longevity. Here are a few interesting findings:
URL Type
I classified each product URL as either a subdomain (e.g. app.example.com), root domain (e.g. example.com), or subdirectory path (e.g. example.com/app).
Root domains had the highest live rate at 75%, followed by subdomains at 70% and paths at just 58%.
Having a dedicated domain (and not just a single-page site) may be an indicator of a more complete, fully-built product.
Solo vs Team
For each product, I counted how many individual makers were listed. Solo projects made up about 60% of launches.
Survival rate clearly increased with team size, from 70% solo up to 92% for products with 6+ makers.
Building and maintaining a successful product long-term likely requires a range of skills that are hard for a single person to sustain.
Name and Tagline Length
Just for fun, I also analyzed the length of each product‘s name and tagline to see if there was any impact on survival rate.
For both names and taglines, longer lengths correlated with better odds of still being live. Whether this reflects having a more descriptive, thoughtful brand or is just a quirk of the data is unclear.
Launch Time
I also looked at survival rate based on what time the product was featured (Pacific Time).
Interestingly, products launched in the evening (6pm-midnight) had the best live rate at 75%, followed by morning and then afternoon.
This may be because evenings tend to have less competition for attention, so products launched then get more sustained visibility on the front page and in the newsletter. But the effect is fairly small.
Takeaways for Makers
Based on all this data, what advice can we offer to makers looking to launch on Product Hunt? Here are my top suggestions:
Pick a focused, niche audience and use case (like developers). Avoid overly-broad consumer social products.
Build something that is a sustainable business, not just a fun side project. Have a real monetization strategy.
Spend time on your branding and web presence. Get a dedicated domain. Clearly explain your value prop.
Try to build a complete, fully-featured initial product, not just a landing page. Prove your commitment.
Consider launching with a co-founder or small team. Solo founders have to wear a lot of hats to succeed long-term.
Have a post-launch plan to maintain momentum. An initial spike from Product Hunt won‘t necessarily translate to traction.
Of course, these are general guidelines, not guarantees. Launching a successful product is always hard and a bit unpredictable.
But by learning from the patterns of past successes and failures, you can tilt the odds more in your favor. The data shows it is possible to build something that resonates with the Product Hunt community and still thrives years later.
Conclusion
Getting featured on Product Hunt is an exciting milestone for any maker. But it‘s just the beginning of the journey.
This analysis revealed that over 70% of products are still online today, with survival rates steadily dropping over time. Differences in category, popularity, team size, and other factors seem to impact the odds of long-term success.
But ultimately, what happens in the months and years after launch depends on you. Keep shipping, keep improving, keep marketing, and keep delivering value to your users.
With determination, skill, and a bit of luck, you too could join the ranks of Product Hunt alums that are still going strong half a decade later. So what are you waiting for – go build something awesome!
