http://www.optimizely.com/

that might help, you can probably rewrite the optimizely generated javascript for each page w/ some perl, to do things as you want.

Perfect, thank you

"Run tests on multiple pages

Want to test a change to your site-wide navigation bar? Want to change the way five thousand product detail pages look? With Optimizely you can easily specify sets of URLs on which each experiment should run, making multi-page experiments a snap!"

I am looking to collect feedback from my users in the most effective way possible. Are there any strong books/articles out there regarding this subject?

i learned a little bit about collecting customer feedback and tweaking your product from the lean startup about what he calls a "concierge minimum viable product." one of the main takeaways for me was not to worry about scalability at first.

http://www.amazon.com/Lean-Startup-E.../dp/0307887898

aytm.com - lots of people seem to like it.

If you are doing a survey you can try SurveyMonkey. It's free for 10 questions per survey & 100 responses per survey.

I'm trying to gauge demand for a new website before it's constructed:

Any thoughts on putting up a home page that tells people what the site intends to do and asks people to give their e-mail address to be notified when the site is online?

We would use adwords to get sponsored links and basically just use this as a way to pay for market research and gauge demand. If a lot of people "subscribe" then we would build the full website. The keywords are not going to be that expensive per click.

In my opinion they are usually a waste of time (people will have varying opinions here though). The reason why in your case it will be pointless is as a measure of potential demand it's a terrible way to gauge it.

It's very hard to get a person to give you their e-mail address on a real, established web site. You really think very many are going to give it on some under construction page?

I'm not sure why you would use that to gauge interest - why wouldn't you use the number of clicks you got on your ads? Those were people interested enough to click.

Its not a bad idea

http://www.startuplessonslearned.com...-for-your.html


google

adwords minimum viable product

you should be able to find a bunch of stuff to help you to decide whether its worth it; I had a really good step by step article, but I cant find it. Ill post if I do

This is pretty standard.

If you want to do absolutely zero work, use launchrock.com or unbounce.com

Otherwise, decide if you want to fake it with a "buy now" button that says "sorry, not finished, please give us your email" or just ask people to sign up to the beta.

Whatever you do, make sure to a/b test the landing pages and adwords and keep the spend low initially - it takes around 6 weeks to really understand what kind of keywords and landing pages will work. From there, plug the numbers into a spreadsheet to see if your business is viable.

I was about to mention Launchrock (forgot about unbounce - have used neither) in reply to Markuis, it does what you are asking, and people do seem to use it, I see it mentioned on Betabeat quite often when they profile startups.

I think sticking up a landing page that collects email addresses for interested people is a good idea, but I have no opinion on the specific idea Markuis put forward of just describing the site. You know what I might do in this case, is to write up the landing page as if the site WAS live and in invite-only beta, spin the email address field as "request an invitation", not "tell me when this product exists", and see how that goes.

There is some discussion of this here actually (found this link after I wrote that last paragraph hah):

http://news.ycombinator.com/item?id=2441828

Incidentally, I don't understand WHATSOEVER how the actual site Launchrock is a viable business. All it does is lets you create a fast landing page to capture email addresses, and yet I see it mentioned relatively frequently, and yet I've seen it referred to as a "really hot startup" a few times. But all it does (unless I am missing something huge, in which case they need better marketing), is creates 1 page for you and adds a mailing list. I don't get it, they have 800k in funding, which I guess isn't crazy, I mean it definitely SHOULD be a profitable business at some point, but it's just bizarre to me. God I hope this means there really is a new web bubble, and I hope it peaks in like 3 weeks after my new site launches.

Oh okay, I guess Launchrock's hook is they create landing pages that encourage people to get their friends to sign up too, the more people you send, the quicker you get an invite. Clever, I think Hipster.com did this notably (yeah, see: http://techcrunch.com/2011/01/17/hipster-2/ ).

So they don't JUST create vanilla launch pages. Although still kinda? Am I crazy?

Sweet, thanks a lot for the feedback! I'll check out launchrock and unbounce. That article guids linked to is really useful too.

You should really check out Kickofflabs instead of launchrock or unbounce. They provide much better customer service, and make everything much easier than launchrock. And yes, I have used both services.

Launchrock somehow got the funding, but they are way behind kickofflabs. Once you sign up, you will see all the different features that they offer, it's not just an email collecting landing page with referral links, it has much more capabilities than that.

great thread. enjoyed reading parts of it. quick question: is there a good site or journal with scientific studies out there on people's behavior online with respect to content?

for example, a site that has articles and research that say stuff like "such and such study says people are five times as likely to read past the paragraph if you use headings through the article."

i'm looking for a serious website, not one of those "INCREASE YOUR TRAFFIC NOW BY BUYING THIS $49.95 E-BOOK".

like a Pub Med of online consumer behavior and marketing, but not SEO.

edit: found this on Clickz since posting. it's a start. http://whitepapers.clickz.com/

http://www.useit.com/
the author has a phd in human computer interaction and has tons and tons of stuff that you'd be interested in.

i do sometimes see criticism of his work and ideas, but it is strongly data driven (frequently using eyetracking, which is very cool) and is pretty much the opposite of scammy ebook type stuff.

i just love reading stuff he wrote in the 1990s (!) that seems super obvious, yet so many advanced, well funded sites get these things wrong in 2011.

for example this is from 1996 and i think 2, 3, 4, 8, and 10 are still commonly made mistakes by large web properties.

and this from 97 is still spot on

thanks for the link. great stuff. looking at this section now on web writing.:

http://www.useit.com/papers/webwriting/

looks like a lot of relevant data and research

agree with you completely. i'm putting together a PowerPoint on how nonprofits can improve their web content and those were some of the gripes i had based on my review of some sites.

If you have a site you want to measure that kind of stuff for your content then you could use http://www.clicktale.com/ It will show you all mouse movements, how far down people scrolled, what they typed, etc...They have a free trial.

this is an open source alternative that can track at least some of that stuff (mouse movements and heatmaps of clicks for sure)

http://www.openwebanalytics.com/

haven't gotten around to using it on anything, cuz i've been lazy to switch from piwik/GA, but it looks promising

We are in a bubble. I've gone from one linkedin recruiter spam (I'm a developer) per month to one per day on average. Seriously. And most of them are from SF/ the valley (and I don't live in CA). Also, nothing has changed about my profile.

Everyone is getting funded. Ycombinator is getting one application PER MINUTE. I am confident that I could go out right now and get 750k funding for a startup idea within 6 weeks of coming up with the idea. My friends are all quitting their big company jobs and forming startups, it's a crazy time.

You're a little late though. Also, you want the bubble to peak around 4 years after launch to give you time to IPO

Note to other developers: if you aren't asking for a a significant raise, you are probably being underpaid in comparison to the market. The Google/Facebook war for developers plus a ton of funding sucked almost all of the free talent out of the ecosystem. The median pay at Google for a developer is over $200k now. This won't last forever - lock in a raise now.

good post sharpie

Ehh to be a bubble it has to be really out of whack with profits. Maybe its just the startups Ive run into but I dont see anyone getting funding completely disproportional to their idea and revenue but I havent looked that hard. Software is just a really damn good business model and its relatively easy (compared to the rest of businesses) to iterate and scale. Maybe Im out of touch with how crazy it is in the valley but people getting paid $200k makes sense to me when you can run Wikipedia type sites (5th in traffic on the internet) with <100 people. A top level, development savvy individual who can manage projects should be a worth a ****ton because the amount they can get done is huge.

Even if they arent worth $200k I would imagine they are $100k and 2x value just doesnt feel like a bubble to me.