Ask YC: How well does your sign up page convert? | Hacker News Reader

Ask YC: How well does your sign up page convert?

18 pointswebwright19 comments

What percentage of people who click on your "sign up" or "create an account" button actually fill out and submit the form?

Note that I am NOT asking how many people who come to your site click on the "signup" or "create an account" button (but if you want to answer that question, too, I think it would be interesting)

View on HN

This is a common question, and I'm not sure there's a "golden number" for how well any page ought to convert; it depends a lot on where the traffic is coming from. One of your best tools in improving your page, however, is to test multiple versions at once and see which one converts better.

Moreover, anything you can do to reduce the number of steps and requirements for the user will also improve conversion. For example, does your product really need them to confirm their password if they can have it emailed to them when they sign up or at any future time? Try removing unnecessary form fields, or deferring them until after the user has signed up.

For reference, my most recent site (still in development) has the sign-up form on the front page, and we're able to convert 10-15% of random visitors from AdWords. I feel like that's pretty good. On another site, I struggled to get 8-10% overall conversion. And, on a very simple project (http://file.io) that received a good number of hits from del.icio.us/popular (very targeted/interested traffic), I was converting greater than 25%.

Just keep making it simpler, and keep testing, and you can improve your page.

A/B testing can help. In a recent project I increased conversion from ~6% to ~13% between six possible layouts.

http://flickr.com/photos/wehriam/428832337/in/set-7215759419...

I don't have any stats of my own to share yet (soon, very soon I'll have some from our site once we launch) but here's a page from a marketing report that shows you 'industry average' of many forms of conversions.

http://img267.imageshack.us/img267/2910/conversionrateshm0.j...

Enjoy!

PS: Ping me off line for more info.

The sign-up procedure we have on http://gtalk2voip.com is very simple and quite unique: no new login names, no passwords, not even an email address, but a working IM account. All user has to type in is his/her GTalk/MSN/Yahoo/AIM or ICQ id. Upon submission our system instantly sends back an IM message right to user's messenger with a unique link to his newly created personal account page. Since we don't have authorization per se (it's done on the messenger side when user signs-in into IM account), the percentage is very high, I believe 50% or even higher. One of the advantages is there's no need to remember another, most likely useless, login name and password. ;-)

The other question I would arise is what is the percentage of active users you have. I have a feeling that most sites and services keep on in their DBs up to 95% of dead souls -- those who once tried and never returned.

On my yet-to-be-launched site, I only have 4 form fields on the signup page:

I think anything more (ex: name, age, address) is too much for most sites... Lots of people will enter fake information anyway, or they'll outright turn their backs. It doesn't even help against multiple account creation on sites that forbid it. The best way not to get sensitive information stolen is not to have it in the first place...

Why not just use the email as a login name and make it 3 fields instead of 4?

I think it's best to offer a two stage user signup. On the first page you ask for two things: email and password. On the next page they are asked for addition optional information to build out their profile (or whatever) with things like name, location, etc....

This way you make signing up as easy as possible (2 pieces of info) and still give them the opportunity to give more info if they want.

It's better not to use just email because people hesitate to give out their addresses. I give them a choice:

https://ourdoings.com/reg.html

On a social site where you display usernames in public, I'd hesitate to make them unable to use a more private handle. Tipjoy uses an email for a "lite" signup, but certain features require activation. Things like dialing up a tip to give more.

I just figured people would prefer the convenience of typing a short username instead of a long email address, when the time comes to log in (which unlike signup, happens repeatedly).

Also, my site is strongly related to another site (it's some sort of fansite), and that one uses a login name too.

For my MMORPG, Star Sonata, I get very close to 50% conversion (downloading the free trial) on my landing page: http://www.starsonata.com/welcome.html

The weird thing (to me) is that out of the people who download the trial, only 1 out of every 3.66 people actually play it. To play, you have to choose a user name, password, and enter an email address, but nothing more.

Funnels on google analytics is really easy to setup. Unfortunately, this is hard to measure for tipjoy, because you can signup in the button, and you aren't taken to a different page that G.Analytics can detect. I'm recording page views and uniques for the button, and the rate of signing up. Luckily the only errors possible are filling out an improperly formatted email or using an email already in the system.

Google Analytics actually just added support for AJAX/Flash actions...look up the documentation, you can have each step of that process appear as a "page".

cool!

If it benefits the user to enter real information (or plausible) then you won't have problems with bogus data. Example; social network. If however the email is only the key into the site, expect a lot of temp emails (with confirmation) and bogus (without).

If the service is free I've seen as much as 25%, lowest being maybe 8%. If it's paid you hope for 1% and dream of 3-5%.

I got 1:11 visitors to buy a pallet jack when I started (only 20-30 visits/day back then) now it has dropped more but less targeted traffic but greater sales.

Just wanted to tell you that you are on the right path, track everything, every stage, optimize the pitfalls.

E-commerce averages 3.16%. Forms average 2-10% depending on how targeted the referring source is.

10%

1% of which seems to just like to send in fake emails

I give out fake emails to anyone who forces me to register to try out a service.

However, I'm perfectly willing to give my e-mail address to a service I've already verified that I like and want to use.

Try before you sign up is really important for getting good info.

Reply on news.ycombinator.com