Ask PG: How would you have started Reddit/HN without your existing audience? | Hacker News Reader

Ask PG: How would you have started Reddit/HN without your existing audience?

131 pointsstaunch30 comments

Both Reddit and HN were seeded by people who read your essays (I was there for both).

How would you have advised "The Muffins" to overcome the chicken and egg problem without the assistance of your audience and YC's press attention?

I know they simulated user activity using fake accounts to overcome the "ghost town" effect. How could they have attracted the first critical few thousand users without leveraging an existing audience?

From your office hours talk I suspect you may suggest finding an especially enthusiastic niche audience at first. Certainly Reddit was all startup/hackers.

The problem is: how do you actually go about getting that targeted audience on the site in big enough numbers to jump-start the virtuous cycle?

(I'm really curious to hear PG's answer, but also love input from others as well.)

View on HN

How could they have attracted the first critical few thousand users without leveraging an existing audience?

Borrow someone else's existing audience.

There are a couple ways to do this. The most straightforward is literally asking them for it. Somewhat surprisingly, people do say "yes" to this. (Joel Spolsky had a subreddit back in the early days, to share links with his fans. </trivia>)

There are various flavors of this in many services. Lady Gaga says all good little monsters follow her on Twitter, etc.

There are, of course, less savory options. The go-to option for many people is spamming Craigslist.

I suspect PG/YC's involvement led to Joel's subreddit, at least indirectly. It also wasn't particularly successful IIRC.

Can you think of ways of borrowing an audience that don't rely on existing status (like being the host of a TV show[1])?

1. http://www.youtube.com/watch?v=W1_YoG7lqI4

This.

When you get right down to it, all marketing is finding an audience that someone else has, and borrowing it. If I'm buying AdWords ads, it's Google's audience, if I'm getting press in the NYT, it's theirs, if I'm putting up billboards in the middle of London, it's the M4's traffic into the city.

The key is threefold: 1. audience interest & profile compatibility, 2. clear conversion points, and 3. a UVP.

Breaking those down:

1. What you're pushing out, and the interests of the audience, have to be aligned. So if you're launching a new HN-style area, you need to know who your audience are (probably dissatisfied HN-ers, some people from Reddit, programmers etc), and identify where those people are (HN, GitHub, Engadget, Bit-tech, Twitter...) and how you can reach them (ad platforms, publicity on blogs that they read, referrals, recommendations from trusted sources).

Understanding your users is key, both in being able to find them, and being able to serve them when they reach you.

2. Know what your points of conversion are. Is it getting someone on to the site? Is it having them sign up for something? Registering an account? Downloading something? Leaving comments? Starting threads? All these things are valid conversions, but it's important to know which ones you want to look at, and which ones are just going to be distractions.

Also, what are you going to do with this data? How about setting up a notifier for your first 250 users, when you reach that point, so you can email them all in person and thank them for coming along? Think about what you want to do with the data you collect, before you collect it. Data without purpose is just time invested for no gain.

3. What's your UVP (Unique Value Proposition)? Why should people come to you instead of wherever else is available? Audience quality? Signal to Noise ratio? Curation of content? Design? Features? What's the benefit that you offer over someone else?

If you can't convey this clearly, you're going to have problems. It's rare to get something like Twitter's early days where downtime and a confusing message is overcome by audience enthusiasm and the community doing your messaging for you. Ensure that you know what you want to tell people, and how you're going to do it.

Hope this helps.

That's a great contribution and really worthy of a blog post in it's own right.

But what's with the first (single word) paragraph? It seemed to me as particularly arrogant... really threw me off. I'm sure I misunderstood you intention, but it almost prevented me from reading the rest of what you had to say.

But I'm glad I got past that.

It is a +1 mechanism for places that don't have voting, so far as I can tell. I have seen it quite often over on reddit. It used to be rather frowned upon here but the removal of visible vote numbers is likely to encourage it.

Basically what stonemetal said. With no vote count, it's a visual indication that I'm in agreement with what the person I'm replying to has said, and extending upon it.

"This" as the start of a reply is shorthand for "the comment I'm replying to makes a very good point that I agree with entirely" (and usually "and here's my take on what (s)he wrote"), rather than "what you are about to read is the answer you are looking for".

"This." is shorthand for "This is how:"

Another way would be to create a lot of captivating content, capturing their email addresses, and turn all those readers into the audience. So give them something they want to read first, then make them communicate w/ each other. Or make sure the social network you want to create has value even if there's NOONE else in it.

You can also troll around in related forums and just start personal relationships with people. Start your own curated newsletter.

I also would recommend scraping fake posts from forums rather than coming up with your own. Put fake usernames, etc. Otherwise it ain't gonna be scalable. I rather invest my time in marketing than creating fake discussions.

"I also would recommend scraping fake posts from forums rather than coming up with your own. Put fake usernames"

Not this.

I agree with the first two paragraphs. But scraping content and removing attribution is unethical.

Over the years I have heard many anecdotes of successful startups doing shady things in their early days. [1] It makes me think this is an industry norm. Do what it takes to succeed. The ends justify the means. (As long as it's just a white collar crime.)

Am I in the minority for noticing this? Or for caring to the point of construing an action such as content scraping without attribution as going down an ethical slippery slope? Do other entrepreneurs consider this viewpoint as being weak and uncompetitive?

More than any of my other comments, I'm hoping to get replies here rather than silent views or votes.

[1] Facebook and the Social Network has been the most sensationalized. But for my argument, let's remove instances of intra-company squabbles. They are often he said, she said; a company's interactions with the public should be easier to interpret.

Content scraping happens all the time, and the benefits outweigh the risks (worst that happens is a cease and desist). If the alternative is investing lots of time creating fake users and fake discussions, I rather scrape and use that precious time for marketing. The content you're scraping technically isn't copyrighted by the website - it's user generated content.

If all you're worried about is getting caught and copyright, then you are missing the point regarding ethics.

I am well on my way to 5000 members on http://www.weekendhacker.net

I started by posting it here, forrst, startupguild and HN FaceBook group then I got picked up by a couple of other places and used that to then ask yet other places to have a look.

People have so far been very positive and that have helped me gain even more traction. It also help that I have actually connected all projects in need of help with someone who have offered to help.

Next step is to write about all the things I have learned from it. I am working on the website (got people to help me there in the spirit of WH)

What I have learned is 3 things.

1. Create as little friction for sign-up as possible. Be concise. Be personal. Be honest. The majority of my signups read the FAQ.

2. Think about social very broadly. For instance with WH I am sending out a mail with the projects structured, curated etc. Instead of people having to go to the website all the time, they receive a mail with the projects. So right now I am not depending on traffic. I am depending on making sure that everyone who have a project get offered help. (100% success rate so far). If it works by mail it will work by other means too.

3. Create boundaries for what your site is about. WeekendHacker is for very small projects. It might expand later on but now we are keeping it simple and exploring how far that will take us.

Hope this helps. I will make a bigger post about it all and then numbers next week.

How exactly did you learn any of that? Random guesses? Maybe the thing you should have learned here is to put the FAQ on the same page?

Simply because it is a mailing list at the moment and still chugging along doesn't mean this is superior. You basically created something naturally suited to being a website, but forced it into into mailing list form.

As for 'not having to go to the site all the time', there is an established solution for this. RSS.

Maybe I have experience, maybe I track things, maybe Email is more popular than rss. But why the attitude?

I probably would have suggested that they get all the other startups in their batch + a few friends each + more than a few of the Reddits' friends. 20 + 40 + 30 = 90, which probably would have been enough.

I strongly believe it's a matter of what you offer. You can tell everyone about your site, but if something of quality isn't there, people have no reason to stay. There are plenty of options.

My community site (hubski.com) is starting to get some energy after a few months. Personally, I believe it's because my view of the site has changed a bit. A bit of digression...: I started the site to teach myself programming. I grabbed the HN code, figured it out, and began to change it. I was an early Redditor, have been on HN some time, and I had a number of ideas I wanted to try. Slowly but surely, I began to build something I personally wanted to exist. I now know exactly where I am going. As for a seed community, I was very lucky to have some friends interested in giving feedback and to mess with it. Some were Redditors too. My wife is a loyal active member as well.

I only have a few months of experience, but IMO best way to build community is to engage and importantly, enjoy the site. Take the time to post the kind of things that you really want to see (not just filler content), and definitely take the time to get to know users. I am glad I did, because I've met some very cool people in the effort. You can't fake community. Don't waste your time trying. People can see through it. It's like a restaurant. If the food is good, people will come back.

I don't know what our trajectory will be, but I don't expect a steep climb. Actually, if you want a steep climb, you are in the wrong space, as a steep climb is antithetical to a quality community. At least the type I am interested in. If you are going to build a community, enjoy the process. -If you don't, you probably won't succeed.

To start something like HN or Reddit without an existing audience is simple, since this is the kind of service that you can start enjoying even with 10 selected users.

So if you are an hacker, just invite a few of your friends and start sharing links, and the community will start growing...

Soon or later I want to create some kind of no profit organization to create something very similar to HN but with a more open model (not YC focused and so forth) and with more features that can be interesting to experiment with. For now I don't have time for family and work issues, but hope to find some time in the next months.

A lot of the jump starting on Reddit came from irc.freenode.org via word of mouth about a new aggregate voting site that needed stress testing, while it was certainly a predominately hacker crowd it wasn't overly dominated by people that read PG essays, I myself wasn't that aware of the PG connection until several months later.

petewailes made some interesting comments, in my humble opinion reddit was likely helped along more by leveraging an existing community that already communicated well than it was by PG essays - although all things helped and I'm sure the essays played their part as well.

Another option, albeit expensive and time consuming, is to start snowballing by reaching out to users in person. For example, we ran many in-person programming workshops for kids using scratch.mit.edu and now we have more than 800,000 users.

I'm really curious to hear PG's answer, but also love input from others as well.

Thanks for emboldening my reply. I am not at all in pg's league as an influencer of online communities, but I've had a personal website with information and a point of view since 1995, and have been very active in online communities since before most members of the general public had ever heard of the Internet. (I started out on commercial online services.) Any online community allows an opportunity for a member to become conspicuous by contributing good content. Quite a few online communities are organized around an initial common interest everyone has (e.g., homeschooling or education reform or gifted education in the communities I'm most active in) and every community broadens its topic scope over time as people form friendships and share other interests.

Were I to set up a social news forum (presumably as a subpart of my personal website, to which I have devoted very little maintenance attention for years), I would announce that first to my 566 Facebook friends, most of whom I have met in online communities. (Many of them I have since "face met" at conferences around the country about the issues we all care about.) Many of them would be good moderators of a forum, and would be happy to help in return for finding new, good content and sharing that with a broader community of readers. I'll have to try the experiment, as soon as my busy technical adviser (my oldest son) squeezes some time together to upgrade my website.

P.S. Is there already a good site of this nature for news and discussion about education policy, or will I have to build my own? There is no sense in reinventing that wheel if a good wheel is already available.

What we are asking here is a subset of the question "how to I advertise?," with the condition that we don't have existing users to use as an asset.

Seems to me there are two different issues to deal with:

1) How minimise the negative consequences of not having users ("ghost town effect").

- One option is to structure the product so that the number of users is not obvious. Since at this point there is little value to showing all the users of the system at once, or highlighting aggregate activity, perhaps it's fine to not include these features, or at least not make them prominent. So a UI like StumbleUpon might fare better than one like Digg, where the aggregate activity is shown prominently.

- Or, pick another MVP with lower critical mass, and then build the community later. Delicious was great because it was useful if only you are using it, but better with others. (though still no where as good as it could be.... Chad, i wish you luck!).

2) what assets can we use (or create) to acquire users efficiently?

Zince we don't have users to work on word of mouth, you have to have something else. Some options:

- A unique value proposition that will excite someone with influence to give you access to their audience. - a unique product structure - a quid pro quo for the influencer

- A "story" that will entice a blogger to want to talk about you. - make it controversial (e.g. blippy ) - make it human interest ( e.g. man sells spot on iPad line using AirBnB).

- A unique marketing opportunity - the greatest ever was when Facebook opened up their platform to intense viral (spammy) promotion. - You may not find a gold vein that rich... but there are always more opening up. Perhaps you were early to find a burgeoning high quality community, and built a persona there with influence.... - find a news story and attach yourself to it. (reputation.com connected with people who had some crazy bad google results, to the degree they could get into the news. )

- An asset you have already. OK, you're probably not PG if you are reading this, but maybe you: - Know a lot of people in area X (which is why that is where you should focus your community). - Have a good story to tell, and can link it to your product - Know ONE person who has enough influence to help you if your pitch is good. As a VC I felt like this was a role central to the job, whether I planned to invest or not... everyone loves to help someone that they think should and will succeed.

That 'niche strategy' sounds exactly like the advice from "Crossing the Chasm", by the way.

Which plateform and theme would HN members recommend for creating a Hackernews like on a nice subject ?

Pligg ? http://www.pligg.com

Anymore suggestions ?

Not a platform, but you can create your own channel on this blatant ripoff of HN : http://textchannels.com

Remember that HN is a blatant ripoff of Reddit and was created because the founders of Reddit wouldn't give pg his own installation (this was before it was OSS). Even Reddit was a "rip off" of Slashdot. Just because it's a social news aggregater does not make it a rip off of hacker news.

Arc: http://arclanguage.org/install

This. Or the Reddit source: https://github.com/reddit/

http://hotarucms.org

Having tried and failed two social news sites, and seeing tons of people try social related startups via H&F, I think I have a little to add. I wrote my own niche social news sites (on Django), and they both failed to get initial audiences.

Cuuute.com was going to be a social news site for cute shit (cat, dog pictures, etc...). It failed because I created a text based social news site rather than a image based one, and because after working on it for 4 months, I couldn't look at a kitten picture without wanting to vomit.

Newsley.com was my second attempt. It started out as a social news site for financial news. I worked hard at getting an initial audience for about 6 months. Finance types generally didn't want to talk about financial news, they want the news that's important to them, and they want it now. So, I pivoted to turn Newsley into a financial news search engine. That was going really well. Traffic was (and still) doubles every 6 weeks even though it's crap, alpha and buggy as hell.

<aside>Hackers & Founders started exploding this winter, and we had to choose either Newsley or H&F. Having a chance to hack Silicon Valley is a bit too enticing to pass up, so we're focusing on H&F right now. You'll hear from us soon.</aside>

The large social news sites, Digg, Reddit, HN have all had large geek audiences at inception. Digg had Kevin Rose's following from TechTV. Reddit and HN got a _huge boost from the people that read pg's writing.

The only other social news site that I'm aware of that's gotten anything close to successful has been Tipd.com. That was started by Digg power user refugees, and focused on a highly monetizeable audience.

Normals don't get social news. They don't get the concept of voting, checking back on a regular basis, and contributing to the conversation. If you're approaching a non-geek niche (like I did for both my social news sites), you're facing an uphill battle. That may change in a year or two as Reddit grows, but Reddit's a juggernaught right now, and it's done a great job of sub-reddit communities.

I firmly believe that starting (like I did) with a social news technology, is the wrong approach. I took a technology and tried to jump start a community around that technology. What I should have done is to create a community and then create technology to support and scale that community.

Adam Rifkin is doing that at 106miles.net for the 106 miles meetup. I'd put a lot of money on him succeeding. Adam has been building the 106 miles community for 6 years and recently opened it up to the public the last 18 months. His team only recently started writing code to support online conversations for that community. Their approach is to build online tools that help their physical community continue interacting.

Start with the community first, and build the technology to support it.

Reply on news.ycombinator.com