The general opinion here is definitely negative. Thanks for all your feedback on this issue. Since I already have the floating sidebar implemented, I'll perform some live tests to see if it even makes a difference in users sharing. In the end I'll probably go with a static share bar on the bottom.

Mind if I ask why you hate them so much? This is assuming they don't cover the content at all (I wouldn't do that). Even if they're beside the content, do you find them too distracting? Too difficult to focus on the content?

Easier to find, more likely for users to share (at least in theory).

Useless is relative here. It may be useless to many users, but it is important for the website. If there's one thing I've learned as a web developer it's "if you build it they will come" doesn't apply.

I agree. No screen real estate will be sacrificed for these buttons. The content already has a max width, and the floating buttons will only appear when there's enough space beside the content. Otherwise they will appear below the content.

I've been learning about A/B testing, and am using it to test the color scheme on my website.

I've always struggled with the design aspect, and I hope A/B testing will allow me to objectively improve the design.

2mb web pages!? And to think I made the effort to reduce these pages from 60kb to 40kb! Now it just feels silly.

No offense taken, and thank you for the suggestions.

I see what you're saying. The light blue on blue for the main content (or cards as I've taken to calling them) doesn't have enough contrast. Something closer to white (like your picture) will probably look better. I'll have to play around with the colors more, and maybe put up a survey for user feedback.

I should be able to shrink the footer on wider screens. It's a fairly new addition that still needs work. People actually use the social buttons a lot, but I'll see about making them blend in more.

Also, I changed the predictor list so predictors that need responses won't link directly to the submit response page. This way, users won't immediately see a registration prompt, and hopefully it will be more clear.

What about the website makes it look dated? I'm the only one developing the website, and I have no graphic design skills. So user feedback is very useful for me.

Also, see my response to Grendel about registration.

Registration is only required to contribute data to a predictor, which makes it more accurate. Predictors need a minimum of responses before they can predict, which is indicated with the "needs responses" tag on the predictors list.

I've been playing around with directing users right to the page for submitting a response when a predictor needs more responses. In your opinion, is this too confusing? For reference, the predictor page for the pets survey is: http://www.cleversurveys.com/survey/5709198289534976_5629499534213120/what-pet-should-i-get

Absolutely. It's an interesting aspect of human psychology.

The internet has definitely lowered the barrier to entry. My website runs on cloud infrastructure, and I pay based on the traffic I'm getting. This makes it easy to start with low operating costs.

I was referring to the marketing aspect. However, I don't think it's a result of a fundamental shift, as much as it is maturation. When the internet was new, every website was on an even playing field. Now, people have developed preferences and biases for websites. It's difficult to break those preferences, and change someone's browsing habits.

As someone who recently launched a website. I have noticed a strong bias towards established brands. If people don't already know about you, they'll be more critical. If they haven't seen you all over the internet, they won't share, because they don't think people care. It's a catch 22. You need recognition to gain recognition.

I'm developing a new feature for my website, www.cleversurveys.com. It will allow for more accurate predictions on all predictors, and automatically disable questions that don't correlate well with predictions.

It's a system similar to feature selection (for anyone familiar with machine learning), but more fine grained.