HN Explorer

Ask: OK, I've got my Web 2.0 website. How do I promote it?

37 points by curtis ↗ HN
OK, I've got my Web 2.0 website. How do I promote it?

This is really two questions. The first is: How does one promote a website in general? The second is: How do I promote this particular website? The first question is interesting to a broader audience than just me, but if anyone has any suggestions for the second, I'd certainly be appreciative.

The website is called Golimojo (http://www.golimojo.com) and the service it provides is Wikipedia "autolinking". Basically you can take a webpage and have Golimojo examine it for noun phrases that can be linked to Wikipedia articles whose titles match those phrases. Golimojo is smart enough (I'm using "smart" loosely) to autolink nouns like say "Iraq" or "Iran" but not common words like "page" or "men". It works better on some pages than others. I did most of my testing on political blogs (lots of name dropping) and the online versions of newspaper and print magazine articles (poorly linked even today, fifteen years after the World Wide Web was invented).

Now out of the billion-odd Internet users out there, probably only a tiny fraction of them would find Wikipedia autolinking really useful. Still that could be thousands or even tens of thousands of users. The question for me, is how do I find them? In particular, how do I find them without putting a lot of effort and money into the process?

One obvious target market is Wikipedia editors who might like a tool to help them identify missing links in Wikipedia articles. But there seem to be dozens of Wikipedia mailing lists, and I'm not sure how a service that looks like it might be commercial would be received (I have no commercial expectations at this point).

Any ideas?

36 comments

[ 3.6 ms ] story [ 82.1 ms ] thread
iamdave
curtis9

I want to shake your hand. This is probably the best example of utilizing Web 2.0 that isn't ambiguously superficial, no obscenely huge banners, flashing colors of "beta" texts. You're attacking the technology at the core. I congratulate you heartily, sir.

I would promote this service personally by NOT making it social (yet), for one thing. There are lots of possibilities in socializing this type of software, which is truly unique and very helpful for research endeavors; all of them good. This may kill some of the buzz you might get from TechCrunch or Mashable on the blogosphere, as it is a great app but I wouldn't rush to make it social just yet. Find the user incentive to socialize it, understand the needs and demands, then bring in the sheep.

Great job by developing a bookmarklet: you see many blogs out there with autolinking, those double underlined links that popup little hover boxes and kill usability by displaying a screenshot of what they're going to see. Maybe even develop your bookmarklet so that it will bring up a smaller, unobtrusive hoverbox with the page title of where you're going (eg. Wikipedia: Tibet) for the end user.

You are correct in estimating that not very many people would find this useful. You may target it at education based blogs, or even yes, wikipedia editors. I know one guy who was just elected to an editor position and I'll be more than willing to ask him if he'd find any use in it.

Finally, commercialization really should be your last thought. 37Signals put it best, solve your problems first and the rest will fall into place. Build to solve problems, not create new ones with financing.

Just my two cents. Awesome application overall, I'm very impressed overall, most notably at your method of attack; sticking with the core technologies and solving a problem from a different angle. Color me surprised if your work goes unnoticed by the big wigs out there.

curtis
Hey thanks!

I'm not even sure how to make it social, so that's certainly not high on my list of things that need to be done. A lot of the design is such as it is because I was scratching my own itch, and I have no love of in-your-face UIs (I didn't fully succeed here, but I can't justify a lot more work unless I get some actual traffic.)

iamdave
If you have the need for such a designer, I'm open for contracts or would be more than willing to outsource the work to some of my associates in case I get tied up by then.
indrax
To make it social, let people define other (non-wikipedia) documents that are tied to terms, at least for linking up their own work. This will make it more useful for SEO, and specific unique domains. Then let people share wordlists, and select wordlists to use on a per-post basis (or maybe your code already has the smarts for this.) The result should be that if I write a post about an arcane topic, the jargon is linked to my favorite references on the subject.

Something to consider: instead of embedding direct links, use redirection so you get a record of every time a link is clicked on.

richesh
make it into a facebook app and see if that increases usage.
curtis
I'm not sure Wikipedia autolinking would be popular or useful on Facebook, although some other form of autolinking might. I have a coworker who's put a lot of work into a Facebook app, I'll mention it to him and see what he thinks. (My personal position towards Facebook is that it seems like a great idea in principle, but I've never found it all that interesting in actual practice)
marrone
hey congrats man, it works quite well. I like the concept. You should see if you could get this incorporated into opensource forum software and the like. Most of these pieces of software have user preferences, one of which could be to have your hotlinking turned on by default. I think forums are a good place to start since it is a basically large amounts of text, almost none of which would be linked properly, lending itself well to your tool
curtis
That's an interesting idea. I don't really have a lot of time to do much engineering these days due to the demands of a real job. It's certainly worth some thought though. Thanks.
ojbyrne
Haha. Step 1 - post it on Hacker News ;-)
FiReaNG3L
What about an API for websites that want to detect links to wikipedia in their content?
FiReaNG3L
Also, if you do this, it would be nice to only link to the first occurrence of a wikilink (tried it on some astronomy pages, and the first occurrence of 'Earth' is nice to have, but not the 10th :P)
curtis
I've certainly considered that idea. There are two big questions here: 1. What does the API look like? 2. How do I promote it to that particular audience?

Right now websites could use the existing service to manually add links to new content. Certain features might make such usage quite a bit easier, for example the ability to take text markup (MediaWiki markup, MarkDown, etc.) and spit it out with links added in the appropriate syntax. This would still be manual, but it would be pretty easy to use.

A fully automatic API could be something as simple as "include this .js file". That's simple, but not very efficient (caching would help of course), and doesn't present the best user experience (there would generally be a delay where the text is visible in the browser, but links haven't been added yet.) A server-side API would certainly be better, but then you have a whole bunch of different server types to support: PHP, Python, Ruby, Java, ASP.NET, or even application specific like WordPress or Drupal, etc.

I could probably roll a client-side JavaScript inclusion API in a weekend, but it would probably take another couple of weekends to implement decent caching.

A server-side solution is ultimately the right way to go, but is there a way to get maximum bang for the buck? Just expose an XML API and let other people do the server-side integration?

FiReaNG3L
A server-side XML api would be ideal; javascript loads on every page and would add significant latency to page loads, which is unacceptable (and wasteful, and tough on your servers). Find someone who can code a Drupal and a Wordpress module and you're pretty much set.

The only question with an XML api is how do you benefit from it?

curtis
> The only question with an XML api is how do you benefit from it?

As long as the load doesn't blow out my hosted server, then just some branding on the client might be worth it.

andreyf
cocks head

I'm not sure if I see the utility in this... maybe I just don't look things up on wikipedia enough, but Firefox already has a "look <selection> up on Google" in the right-click menu, and most "important" nouns will have a wikipedia link on the front page of Google anyway.

As for promotion, in the "level of connectedness" we now have, word-of-mouth seems to promote products really well. Make something impressive, and people will tell their friends about it. At this point, I won't tell my friends about your site, I probably won't remember it in 15 minutes (unless something on-topic comes up). Do change my mind ;)

So I'd say ... keep working on the idea/implementation until it wows people, at which point users will find you.

jakewolf
Pretty freaking cool. Make a plugin for wordpress. Getting on people's blogs would give your great visibility and wordpress developers/designers are always talking about new and useful plugins.
marcus
Just add the option to make the link nofollow so blogs that don't want to squander their PageRank prowess can keep the linking for links they actually want to approve manually.

Another thing you might want to add is to build a wikipedia based spell checker for the plugin for misspelled names etc.

ryanspahn
I constantly right click on words or subjects to read their wiki entry. I wished there was a firefox extension in which I right clicked the word or subject and in doing so would bring up the entry in a box (CoolIris like) without leaving the page.

I see you have a firefox extension - does it do something similar to the above?

As for promotion, ironically this post here on Y News is a great place to start! Its a highly trafficked site with a demo that ranges from the tech influential to the next Sergey/Brin!

curtis
I think this is a good idea and I'd like to do something similar. The current version just adds ordinary hyperlinks to pages and I'm too time constrained to do anything about it right away.
rksprst
This functionality is built into IE 8. You can select text and look it up on Wikipedia, Facebook, etc...
rksprst
Was I voted down, simply because someone doesn't like IE or Microsoft? I was pointing out a feature of IE 8 that directly competes with his website.
PStamatiou
btw, the frame that loads the website provided by the user is way too small (100px-ish) in Safari 3.1 OS X.
curtis
Ah, I see what you mean. Try it now.
wastedbrains
I would created plugins for popular blog authoring software movable type and wordpress and such.

I might also be cool to make an API so developers could call your service on single words or sections of text from their applications. I might not want whole pages autowikied, but it could be cool to have sections of articles or something. Offering a simple and easy to use webservice can drive adoption of your service in ways that you could never even begin to imagine.

I would also try to find some high profile bloggers in your target audience (in your case politics maybe), and write personal emails to them about how the service could be good for their readers and ask if any would be interested in trying it out, or if there were any features they would really need before the service would be useful for them (such as which wiki links were click through, showing what their readers don't know enough about or are most interested in)

anyways just some thoughts off the top of my head.

8en
I might consider changing the name to something more direct - for example, wikiautolinks, or wikilinkster, or wikilinkr, or wikiglue
elai
This is definitely better as a software extension to websites & browsers, not a website that you go and visit, like flickr or something similar.
rokhayakebe
this is a great idea. i always find myself searching for people or products mentioned within articles. great. could it be like snapshot where we could see maybe a bit of the content then if we are interested, we can click for more and be redirected to wikipedia.
astine
Promoting a website? It's easy. Submit it to mailing lists, search engines and discussion boards. For additional measure, you can even try purchasing advertising and making a youtube video but this is unnecessary. If people like it, they will spread the word, and usage will lift it higher on search engine results lists.

I was thinking that this should be turned into a firefox extension, but I see that you already did it.

tim2
That's not how it really works, when it works.
NewWorldOrder
To me, initially it was unclear when somebody would want to use this. After some thought I figured that something like this could be cool for a tech document or white paper you're reading online with a lot of unfamiliar or esoterically techie jargon in it.

So people who have to read about difficult concepts in varied areas may find this useful. This could be anyone from students to professional consultants. So finding forums where students are trying to get help with various topics could be a good place to start.

doubleplus
I know you're not asking for advice, and I figure you're still working on it, but I'd be more likely to use something like this if it didn't only search for capitalized proper nouns. i.e., If "san antonio" or "set theory" generated links too. I wouldn't mind waiting a little longer for the script to do the extra work. http://www.golimojo.com/proxy.html?christiansands.livejourna...
curtis
It's not strictly limited to proper nouns -- try "jury nullification" or "obfuscated code" for example. However, if the matching title is a proper noun, it requires exact case match. This is a side-effect of relying heavily on case for matching in order to prevent accidental or overzealous matches. This is in fact a fairly thorny problem, at least without some sort of semantic parser.
doubleplus
hm... okay i typed a few more into the same link and it picked up a couple... i just figured it was only proper nouns b/c those worked every time. thanks...
nickhac
Hire a Startup Marketing company :)

We work with Australian Web 2.0 startups to help them grow traffic, conversions and customers.

It seems in most scenarios the skillset needed to build a great web service, and the skillset required to market, promote and sell that service are be totally different.

Something to be wary of tho, the majority of PR/Marketing/Advertising companies really dont get startups at all, and can be a very fast way to drain your scare capital without offering measurable results.

Any questions, happy to help out.

http://www.ShiftedPixels.com.au