I thought landing pages were super easy, until I messed up. When I put up the initial landing page for my latest startup ( http://questcompanions.com/ -- finally launching next week!), the 'Enlist' button said 'Start' instead. Due to the non-traditional input box and the text on the button, people would just click 'Start'. Due to a lack of error handling in the frontend code, the empty email would go up to my server, go out to Mailchimp, and they'd hand back an error. All the user would see is "Invalid or duplicate email", then they'd move away.
When I integrated Mixpanel, I saw why immediately and was able to fix it, but by that point I'd already wasted an easy $10-20 on Facebook ad clicks leading to the page. Tiny sum of money, but those were good clicks. Implementing error handling, better button text, and slightly different text on the input box caused the conversion rate to go way up, and the rate of people clicking the button without filling in an email dropped to nearly zero.
Long story short, measure everything and be very, very smart about your messaging.
I built my first web site/blog a couple of weeks ago, and I've been experimenting with advertising on Google, Bing, Facebook, etc, and I have to say it's been a really interesting experience.
Basically the most important thing I've learned is that I have to understand why my users are coming to my site, and from there tailor my site around that.
I have a very high bounce rate of around 75%, but most of that is due to smartphone users and the fact my site is pretty hard to use for them. Creating a simplified page is my #1 goal over the next couple of weeks.
Looking at logs, I also realized that my landing page was hard to use for a significant percentage of my new visitors. I could see from the logs that they were getting confused, trying it out a couple of times and bouncing. I was working on this last week, and tried a couple of GUI changes, but decided that having unintrusive error boxes would probably help better. Hopefully that will help those users having trouble.
But from the ones that stick around, I do have a few pretty regular users, and that leads to what I think is a kind of nice ad clickthru rate. I've paid a good chunk of all my costs in the first two weeks (of course, my costs are only around ~$30).
Thanks for the pointer to Mixpanel. I was looking for a better type of analytics than Google Analytics which doesn't offer me a great deal of information. The only hesitation I have is whether or not having all this javascript client-side tracking will cause users to question their privacy. I guess I'll cross that bridge when I get to it.
When I integrated Mixpanel, I saw why immediately and was able to fix it, but by that point I'd already wasted an easy $10-20 on Facebook ad clicks leading to the page. Tiny sum of money, but those were good clicks. Implementing error handling, better button text, and slightly different text on the input box caused the conversion rate to go way up, and the rate of people clicking the button without filling in an email dropped to nearly zero.
Long story short, measure everything and be very, very smart about your messaging.