Many web apps are heavily dependant on network effects, and are completely useless without a decent user base already in place. When faced with this issue, the common wisdom is to target a small niche at first. Over time, you can branch out to other users. There are many examples of this, such as Facebook starting with just Harvard students for example.
The strategy makes a lot of sense, but the question is what niche should one choose? In our particular case, we need to sign up people who browse many of the same web pages. This is necessary because our app, Jellly, is meant to suggest related links as users browse the web, so there better be other users who saw some of the same web pages earlier and already suggested related links.
The “hacker” niche, as represented by this community, may be a good choice here, but I’m not sure. We had thought about targeting bloggers, as they are often all looking at the same blog posts and articles; they provide free marketing and they could get into an app like this (it’s much better than using track-backs for sure), and they’d be happy to get some traffic out of it too by linking their own posts to other relevant stuff on Jellly. But the overall blogosphere is huge; the target needs to be more focused than that at first. Hacker-bloggers?
To better understand Jellly and our problem, you could take a peek at our FAQ. It may be more instructive though to see the crappy demo app I threw together in a few hours for our (failed) YC application, as long as you don't consider it to be characteristic of the actual app. It really just demonstrates how it might work for finding content: http://prettyobscure.jellly.com/
Yeah, I had looked into their API before and it's a good option. Using tags to figure out the related links could give decent results to start off with. The results will be suboptimal though, as any automated approach would be, since "relevance" is more complicated than simply having a few matching tags. I hadn't seen moreofit.com before, seems to give pretty good results, so I may have been underestimating how far you can take tag analysis.
I think it's a very interesting idea, but I definitely see your problem of how to get initial content. I think a good niche would be programmers of a certain language, maybe ruby, javascript (or even just jQuery). Try to hire a couple of the more popular bloggers in the space to start seeding it with data and promoting it some and I think it would take off. If I was looking at a certain rails plugin it would be really interesting to see a list of links to tutorials for that plugin as well as some competing plugins that I should checkout.
I'm just one data point... But, while I think I grok what you're doing, I'm not sure I grok why I should want it. When I read articles, maybe 1 out of 10 makes me want to explore further. When I do want to explore further, there's usually a pretty natural way to do it-- the google search box in my browser. Google uses the link graph to find stuff related to the keyword. You're proposing to have users do it, right? The Google path doesn't seem to be broken to me... Other than the rare time that I can't think up the right search phrase to get good results.
I'm sure you sometimes click on hyperlinks besides those that show up in search results. The point is to make navigation of that manner via related hyper-links better, which is somewhat orthogonal to search we think. Sure, you can search and find anything, but it's not quite the same as just browsing by clicking links.
Google uses the link graph to find stuff related to the keyword. You're proposing to have users do it, right?
I'm proposing to have users create a new link-graph, separate from the web's own inherent link graph. Keywords have nothing to do with it. :) - Looks like my home page isn't as clear as I thought.
I'd suggest a screencast on how to use it. Is it a separate site? An add-on that allows me to explore/suggest sites based on what I'm currently viewing?
I think I'll be adding my demo to the home page soon, once I polish it. Otherwise, it can be pretty confusing.
There is probably going to be more than one way to use it, but the main way would be a browser extension. A web toolbar is also going to be there for people who don't like installing extensions, to get them started fast. You can see the demo of that (only shows you how you might browse through, had no functionality to add links): http://prettyobscure.jellly.com/
Let me know if it makes any more sense after trying the demo :)
I think you've done this backwards... problems that need solving tend to come with their own niche - the people who want the solution.
Don't get me wrong, this is an interesting app, but I don't see a demand for it (people are more than happy with Google mostly), and I'm curious exactly how you'd expect to turn a profit with it as well? Adverts and mass adoption?
I don't think this is true, but it does sound that way from the way they explained it. Your app may be targetting a certain demographic, but you can start marketing heavily in a certain niche. Facebook and Harvard students was the example he used. But a lot of the location based services start in one city and then expand. Groupon and Foursquare are examples of this.
I think you've done this backwards... problems that need solving tend to come with their own niche - the people who want the solution.
I actually intentionally worded it backwards from the usual as a minor form of headline bait, which failed. The app really should have broad appeal among people who frequent the web in general. That group of people is just too varied and spread out to help each other find links through Jellly. Hence the need to focus on a smaller, more closely knit group.
Not all apps fit the "solve a pain point" model that paid web applications go through. While pain killers are great money makers for sure, sometimes you don't know you'd want to use an app till you've tried it and see why it would be useful. These apps have a lower chance to succeed I guess, but they can be great when they do.
I know there's nobody dying to have this service. But Twitter comes to mind as another application that people thought was mostly useless. That is, until some of them used it (I know most people still think it's useless, but many obviously don't).
Advertising fits perfectly here, since the service is there for relevant links. The advertising would consist of relevant text links, similar to search ads. Mass adoption would be necessary for large profits, but I believe it's possible to break even much earlier, using advertising networks to start advertising early on.
Do you have any beta sign ups? If there is a recurring theme that might make it easy to select a niche. Another thing you can do is select a niche, any niche, that you are already active in, like gaming or cooking or something like that.
Hacker bloggers is not a bad place to start if you are already active on HN and other places. The only thing I would caution is that hackers tend to be much more critical than the average person. Like some are willing to be early adopters but others expect everything to be perfect. Expect to hear a lot of "You should fix this bug/implement that feature - it should be EASY!" I think that tech-savvy but non-hacker people might be more open to "Is this useful?" discussions.
As a side note I find your logo to be really unclear. Honestly I think you should change it if you want it to really catch on, but if you're not willing to do that, then you might put "jellli.com" text underneath it.
We just started telling people about the site two days back via Facebook/StumbleUpon and I guess HN now, and we have just above 40 signups so far. Conversion is pretty low, but that's expected without any real content. Need to fix that, add the demo perhaps. Part of asking what stuff people are into is to try to find a recurring theme, so I totally agree about that.
I think your comments about hackers are on target mostly, but I don't mind the criticism. It's really better than having friends tell you how great what you made is.
The logo was not worked on too much, and I'm not exactly good with graphic design. I'm definitely going to get a professional to redo it in the future. I'm not sure it would make a huge difference to whether people sign up for beta, but I can see how it might.
I'm going to take a leap here, but I'm pretty sure that -- on the internet -- hacker/blogger/programmer/designers are not a 'niche'. There are numerous communities for all of these people. Instead you should focus on groups of people who have a hard time connecting. There are plenty of communities that have no real internet equivalent. A small list of actual niches:
- UrbEx (Urban Exploration)
- Board Game Enthusiasts
- Cooking (This could be a big one. Food blogs are really popular, but disconnected.)
There are a lot of people out there, and if you can focus on long-term user acquisition (which rules out the 'try every new thing and then drop it a month later' tech crowd), you'll be a lot happier.
Very interesting. I think it would be difficult to help create an internet presence for those who don't already have some, whether in a community or not. But in the case of disconnected blogs and other content, where a presence does exist, there may be great potential to help unify all the bits spread all over the web. I'll look into this more, but I may have a hard time finding appropriate niches.
By "niche", I had really meant a group of like-minded people. There may be a more appropriate word. I don't want to stick to a particular niche forever, but expand to all possible topics eventually. The "niche" is just required to make sure there are network effects. People who try a new thing every other month are ok with me, as long as they help bring in others who are more likely to stay. Besides, some tools are rarely dropped by the tech crowd if they are truly useful, and I would hope that my application fell into that category.
I think that it's a great idea. One thing that I'd like to comment on: crowdsourcing > algorithms, especially with social going mainstream now.
The demo was great, but I much prefer that it's a bookmarklet instead of an inline frame. "Related sites" is not something that I need to see all the time, and it takes away vertical screen real estate, which can be annoying to some users. If you're set on using a widget, use the side of the screen to maybe house a button, expandable on click.
I agree about crowdsourcing > algorithms, obviously.
Yeah, I'm planning to have a bookmarklet, iframe and browser extension UIs all available. The final UI is in no way set, and probably won't look exactly like the demo. Lots of people don't know how bookmarklets work, but they'd know how to use the toolbar. And many people are wary of installing extensions still, so a web-only version is good to have. Only problem is the confusion having multiple interfaces may cause some people. Will have to A/B test and see.
BTW, do make sure you sign up, haven't gotten any more signups since a while back :-)
Many web apps are heavily dependant on network effects, and are completely useless without a decent user base already in place. When faced with this issue, the common wisdom is to target a small niche at first. Over time, you can branch out to other users. There are many examples of this, such as Facebook starting with just Harvard students for example.
The strategy makes a lot of sense, but the question is what niche should one choose? In our particular case, we need to sign up people who browse many of the same web pages. This is necessary because our app, Jellly, is meant to suggest related links as users browse the web, so there better be other users who saw some of the same web pages earlier and already suggested related links.
The “hacker” niche, as represented by this community, may be a good choice here, but I’m not sure. We had thought about targeting bloggers, as they are often all looking at the same blog posts and articles; they provide free marketing and they could get into an app like this (it’s much better than using track-backs for sure), and they’d be happy to get some traffic out of it too by linking their own posts to other relevant stuff on Jellly. But the overall blogosphere is huge; the target needs to be more focused than that at first. Hacker-bloggers?
To better understand Jellly and our problem, you could take a peek at our FAQ. It may be more instructive though to see the crappy demo app I threw together in a few hours for our (failed) YC application, as long as you don't consider it to be characteristic of the actual app. It really just demonstrates how it might work for finding content: http://prettyobscure.jellly.com/