If you're going to gray out something in the UI (and there are reasons why you might), here's a few things to remember.

1. Your users won't know why it's greyed out. Is it currently disabled? Is it not available yet? Have I broken something? Is it paid only? Is it launched?

Resolve this by explaining why it is grayed out on hover.

Anecdote: We all remember trying to click "Make Table" in Microsoft word only to find it was grayed out for some obscure reason (Scroll lock was on or something). Telling a user they can't do something but not explaining why is a shit experience.

2. Track clicks to it or hovers on it, it's a good indication of how much action it'll get, in my experience. If no one is going near the feature, then chances are it's either not worth building, or it's positioned in the wrong place on the screen.

3. Consider bringing it to a splash page where you post a teaser of whats to come and ask your users what job they're trying to do when they click "Merge Users" or "Share Reports" or whatever. Just to make sure you're delivering the right value.

Hope that helps.

Des

(Obligatory: Use Intercom (http://intercom.io) to understand user behaviour and talk to them, this helps avoid a lot of second guessing)

I have recently read a lot of blogs on this exact issue. After all the points of view and insight, what I would suggest is show a link to your new feature, but make it an active link do not grey it out. The key with the active link, you can measure how many of your users actually click the link and would be interested in your new feature without having actually spending the time/money/effort creating it. However, as you mentioned the draw back is by activating the link to a feature that is not there you will have to post an under construction page or the mm/dd/yyyy the feature will go live.

I think it really depends on how engaged your users are with your product and who you're targeting.

If you have a customer base that is really into your product and the page is question is aimed at current users, then offering information about an upcoming feature will be something that's engaging to them.

Conversely you expect the page to help you acquire new customers then it will just serve as a type of "excess" information.

Sounds like a good idea to me.

Tripadvisor does something similar: http://techcrunch.com/2011/12/22/founder-stories-tripadvisor...

It seems that data is a core element of the 404-test/ 404-link-test approach. Would you ever do it without data? to teach a user about a future feature?

Two answers:

- Theoretical discussions about UX are often meaningless. Create wireframes and UX test them.

- Look in UX pattern databases to help you decide what to test. Sometimes your design choices will have previous significance to your users that are different from what you intended.

As with most things I don't think there is a black or white answer to this. I think it's OK to have some features that are unavailable if tastefully done, but it will make your product have an unfinished look to it.

what requirements would you have to implement a grey-ed out feature?

No.

Because you may change your mind about that feature before you launch it. UNLESS you're doing a test to measure how many people click on this feature before building it, but that's a different story altogether.

so data is a requirement for you too?

teaching a user about a pattern is not a good enough reason?

No because you're just teaching them something doesn't work (yet).

It's a real commitment to a feature you haven't made with little reward. The only reason to do it is if customers are asking for it. If they aren't, you promise a feature and then it doesn't go in for whatever reason, this sucks.

If you've got a good customer base and you want to let them know you're working on stuff. I like Freeagent's approach. http://depot.freeagent.com/

Meta-question: should Ask HN questions (the primary piece of content on the page) be greyed out?

That's always bugged me and I never understood why they're greyed. The original comment is very important in the context of the discussion below.

We use greyed out buttons in two ways: 1. To gather data around interest in the feature. After clicking tell your user "thanks for the vote you just cast for this feature, we will notify you when its ready." 2. To indicate paid account only features

Why put in anything that doesn't work? That's bad UX practice in my experience. It offers no immediate functionality and just serves to confuse the user.

My gut reaction agrees, but I have seen times, especially at 'telly' where we were able to do a quick test and prove that users were interested in a feature.

However in that case we were both watching data and only running the test for a few hours.

I'm wondering when there are best practices to do this, and maybe when not to do it.