So I’m reading Wikinomics right now and I came across a great concept: “treat advertising as if it were content”
That got me thinking a lot as to what the implications for such an approach would mean:
- Allowing comments to be attached to ads
- Giving people the ability to share ads (like a digg for ads)
- Forcing ad-creators to approach advertising from a completely different perspective, etc.
It’s just something to keep in mind…
I was thinking a lot about our product today. I believe that our value-proposition needs to be quite simple – our product creates a higher rate of clicks per ad.
By analyzing real-time content created by users, our algorithm does a better job at targeting ads to particular users and thus the users respond more positively.
Can we do this?
Google already does this in gmail but they have deeper pockets than one part time noob web developer. I wish I had a high volume site to test personalized vs. contextual ads or a even a good report on it.
I’m not 100% sure. It’s a gamble on a hypothesis, but the reality is we won’t know until it’s sitting on a live site and we’re getting real feedback. We’re not ready for that at this time (no plugin). I could probably jury rig google ads off of a single users twitter username and have cookies remember them.
The best way to get those measurements is to have a plugin live on website for a period of time and compare it to normal google ads. I could host them side by side on my blog once I get the oauth working.
I’d prefer having a product we can turn some dials on before serious testing begins. I don’t want development to stall fine tuning once it’s live.
Testing on your blog is a good start I think.
More good news got an invite to Google Wave. Should have that to play with in a couple of days (another potentially very powerful social media).
API http://code.google.com/apis/wave/
doc: google app engine documentation http://code.google.com/apis/wave/guide.html
That is very good news.
Got python installed in netbeans, and stand alone. Got friendfeed API lib and sample codes (had to manually cut and paste into local text files, there web versions had gunk/html mixed in). Installed google app engine code. Having some problems getting their code to compile with jython in netbeans, no luck with command line python builder either. Hair pull out exceeded for the night shutting down.
Don’t pull your hair out – the morning always brings new light 
Got some interesting feedback from the friendfeed developers. They said speed wise most apps are limited by data transfer rates and memory, independent of the coding language. So there’s no strong push for scala/java (faster) vs. python/rubyonrails. My gut instincts are leaning towards whatever works now but I’d like to leverage the faster web code scala/java if we do any heavy database processing (clustering) or fast visualization on the fly.
At this rate I’m going to learn at least the fundamentals of scala/lift, java, php, python, and ruby with mysql before we’re through.
Google app engine plugins for netbeans: http://ekschi.com/technology/2009/06/22/developing-for-google-app-engine-in-netbeans-ide/
Ok, so I did research on registering a company tonight. Essentially we have two logical options: the C-Corp and the LLC.
LLC: This is slightly simpler and cheaper to set up but more difficult to raise money, gain outside investors, and eventually go public.
Pros – Owned by members, not shareholders; liability protection; no double taxing; less strict laws about meetings, boards, etc.
Cons – Unlikely that a venture capitalist would invest in LLC; cannot take company public; complex operating agreements; restrictions on transfer of ownership; different laws in different states
C-Corp: This is a bit more complicated to set up but is generally the preferred method for venture capitalists and other investors.
Pros – Limited liability; most appropriate for IPO; easy to set up stock option plans; easy to transfer ownership and rights; structure that venture capitalist requires
Cons – Double taxation; high administrative compliance costs; directors held accountable; well defines laws and regulation
I think that the choice is pretty clear – if we plan on growing this thing quickly and looking for investors early on than the C-Corp makes the most sense. S-Corp limits ownership to only 100 people, sole proprietorship doesn’t make sense, and partnership is also to limiting.
We can set up the company pretty easily and cheaply through LegalZoom (http://www.legalzoom.com/) – I know a few satisfied customers.
so C-Corp wins, legal fees etc totalling?
Legal fees (using LegalZoom) and state fees come to about $300. https://www.legalzoom.com/legal-incorporation/incorporation-packages.html
I think I’m correct – take a look.
wow that’s pretty inexpensive. I like. Do we need to find any other founders to start it up or can we cut them in for a stake later on?
Yeah, it’s not bad. I suggest we wait just a little longer to see how things develop – i.e. if we find an additional founder soon, etc. I’m pretty sure that we can add others later but it’s probably easier to do it form the beginning.
What’s this?
documentation, I’ll clean it up
Great conv with Jesse Stay of socialtoo who may have exactly the type of system we’re looking for as part of his software (he’s working it now though). I’ll come back and review what he has later and see how we can hook into it. http://staynalive.com/articles/2009/07/20/friendfeed-opens-up-the-firehose-to-developers/#comment-12967187
What type of software is it?
Supposed some groovy mashups to social media here: http://www.programmableweb.com/
friendfeed’s new API: http://blog.friendfeed.com/2009/07/friendfeed-api-v2-real-time-oauth-file.html
Users want tools like lazyfeed similar to the search aspect of the project, to slurp up their google reader and find similar stuff
messel 10:15 am on July 26, 2009 Permalink
Reddit actually does this already and it’s a interesting approach.
I had a related thought on monetizing the real time topical conversation discovery tool. First the tag analysis is performed on a user. Then when they log into our site (or API) virtual chat rooms are formed about topics happening at that moment and those conversations related to your tags show up as “joinable”. On the right hand side there’s space for relevant ads to that topic as well as users who pay for their messages to endure longer
messel 10:16 am on July 26, 2009 Permalink
You can always manually search for chat topics and join that way
messel 10:16 am on July 26, 2009 Permalink
Or you can find conversations by user
messel 10:28 am on July 26, 2009 Permalink
Just an update, still reviewing oauth for facebook and Twitter with python. We still need an independent user profile generator. Hoping to find some good generic code for that , think I found some php that will get us up and running.
messel 4:05 pm on July 26, 2009 Permalink
Vlad could you look into licensing, particular the apache 2.0 license. I think I’ll be using much of that type of licensed code for the prototype software framework: http://www.apache.org/licenses/LICENSE-2.0
Vlad 10:34 pm on July 26, 2009 Permalink
Ok, I’ll look into it.
Vlad 5:25 pm on July 27, 2009 Permalink
After reading the license it seems that we’re good to go. We just have to be transparent about using Apache.