When our software is installed, it installs some networking components. It may run fine immediately after the install, but chances are that some applications won't be able to talk to the network until after the reboot.

So we would REALLY like them to reboot, but we understand that some people might prefer to reboot later.

One of the developers feels strongly that the solution is:

  1. Ask them to reboot (recommended). If they answer "No" then,
  2. Tell them that they may have problems with network applications if they don't reboot. Ask them to reboot (recommended).

This is done via MessageBox.

I'm in favor of showing just the one message (#2 that is, the more specific message).

I see it as "nagging" the customer and simply being a pain in the ass.

OK, so here is the rationale as I believe I heard it from the person who's writing the double-message box approach:

  • Users don't read long text, we don't want to say too much on the first dialog.
  • But we really want them to reboot so that we don't get so many support calls, so it's worth asking them a second time even if it's a nag.

7 Answers 7


I'm going to give you a high level answer since everyone else is already tackling the "show one or two messages" part.

Don't be one of those apps

Instead, here are some counterquestions to possibly affect your design decisions:

  • What percentage of users is going to encounter problems if they don't reboot? Is it worth nagging 100% of your users if <5% has issues?
  • Is your application architecture able to support delaying the installation of the features depending on networking components until the user chooses to reboot?
  • Could you redesign your application so that users who will be using the networking features can install a specific networking-supported version, but everyone else installs one that doesn't require rebooting? Be as up front as possible; you could tell users this as early as the app's download page.

Use psychology to make the process less painful

If you must show messages, consider some behavioural psychology principles. If you're going to nag me, at least do it with a smile - that might soften the blow a bit. I remember this app that measures the time it takes you to read the terms of service and click "accept", and if you do it really quickly (because you didn't read them), it pops up an alert saying "Congratulions! You read the terms in 3 seconds! ...or did you?", which brings a sense of humour to an otherwise annoying step in a process.

Here are some options you could consider:

  • Offer the user something in return for rebooting. Seducing the user can be a quick & dirty technique. "We know it's a hassle to reboot your computer just to install your app, but after you come back, we'll have a surprise for you! Do you like free t-shirts?"
  • Invoke status and social proof. "93% of users reboot their computer immediately after installing for a fresh, clean experience. Do you like being fresh and clean?" ;)
  • Reframe the situation. "Did you know that rebooting your computer often helps prevent system degradation? Why not do it now while you're installing our software? Then you won't have to do it later! All thanks to us."
  • Combine various principles. "Only 7% of users choose not to reboot their computer now. Don't be one of them! If you reboot now, we'll give you a free t-shirt! (One-time offer.)"
  • Achievements. I went there. "You've gained the Rebooted During Installation achievement!"

You can kind of go on like this, but my point is that you could brainstorm with the team to find whether there are opportunities here beyond being yet another boring app that requires me to reboot in order to complete installation. Instead, choose to be a memorable experience. Applying any of the above would make you a million times more interesting than your competitors.

Don't look at it as a messaging problem. Look at it as a brand opportunity!

  • Awesome points! Just changing the whole nature of the "nag" can make it not seem like a nag at all! Well put! May 27, 2011 at 17:04
  • 3
    While I would shoot a team member that suggested any of those examples, the approach behind them is awesome. Nice one, Rahul.
    – gef05
    May 27, 2011 at 18:56
  • 1
    +1 for "don't be one of those apps" - remember one of the most loathed misfeatures of WinXP, the WUA Restart boxes? "Reboot now? No? Well how about now? And how about now? Oh, nevermind, I'll reboot anyway, its evident that you want to do that, otherwise you wouldn't be clicking Reboot later." I see that the OP is only suggesting two boxes, but the second one would be enough to question his sanity. ("Well you just asked me about the same thing, what are you, deaf or stupid?!?") May 29, 2011 at 17:16
  • I've chosen this as the accepted answer. Of course, some of the suggestions you make other respondents suggested as well, but this one seems to encompass those as well as make some great suggestions about the psychology involved in convincing them to reboot. (Although, we're not giving away t-shirts just for rebooting. ;-)
    – Fauxcuss
    Jun 2, 2011 at 16:25
  • 1
    @MarkS You wouldn't be giving away t-shirts for rebooting, but for installing your app. Add it to the cost of acquisition for that customer. Mailchimp does it as well (after sending your first newsletter with the service).
    – Rahul
    Jun 2, 2011 at 21:33

Your instinct is right. It will come across as nagging.

Your developer is right. There do need to be two dialogs. The first with Reboot (recommended) and Continue.. and very little else. The second giving some information on the consequences of having ignored the recommendation.

The solution is that instead of a Reboot button on the second screen you provide instead a <--Back button. This might seem to fly in the face of good user interface design. You want them to reboot, right, so why add an extra step and 'reduce conversions'? Wrong. Getting them to reboot is not the objective. You want them either to have rebooted, or you want them to continue on informed - with a higher probability of being aware what's going on when their Blackberry still won't sync over bluetooth.

Combine Them?

You could combine the two dialogs into one. If you do, then text which presumably you're both convinced is important is less likely to be read. Also it inconveniences the good citizens who do decide to just follow your recommendation.

What is the Second Dialog for?

The second dialog is an information dialog for a decision already made. It should not be a warning dialog asking them to reconsider. That's a very big difference. The <--Back button is just enough to convert what would have been seen as a nag dialog to what will now be perceived as a dialog intended to help and inform. Which is what it is. You're expected to continue on.


I'm not a fan of message/alert boxes in general, so to me a double message box (no matter how important) is more annoying than it is informative.

Once they dismiss the reboot later message box, users aren't likely to read the warning message box and just click whatever button makes the box go away.

What I would do is put some kind of alert/warning icon or user message ribbon somewhere in the app that will warn them of the potential issues without rebooting. This icon/message ribbon would only be displayed right after the initial install if the reboot hasn't happened yet. Once they've rebooted, the icon goes away and everyone is happy. :)


I'm with the idea of incorporating a message into the app. Could the installation not be considered "incomplete" until the reboot has happened? If the customer is clear from the start that they will need to go through certain steps like:

1 Open file | 2 Install | 3 Reboot

... then the reboot will not be a surprise or an annoyance, just part of the process. Sticking dialogue boxes in the way is a lazy patch, not a proper solution.

  • 1
    Now that you think of it, we could potentially not install the components that may/will cause problems until the next reboot, and we can set a flag that indicate that the installation isn't yet complete, so functionality related to that can be bypassed with a message in the app that tells the customer that we're waiting for a reboot to fully enable the product.
    – Fauxcuss
    Jun 2, 2011 at 16:20

One of the most important things in UX is, informing the user about things that will be done before they actually are being done. (E.g. giving the user the sense of being in control.) In this way, the user can think of the actions that will be taken or are needed to complete a process and decide to start or not.

This means that if the user is informed about the necessity to reboot before starting the installation, the user will carefully wager the urgency to install at this moment or wait until a more fitting time.

This prepares the user, makes the pop-ups unnecessary and creates an informative execution plan of the installation of the software.

Hope this helps!

  • 1
    It's a little more subtle than that. Where and what information you give is an art. You're correct in the case where actions cannot be undone. Often in UX giving less information, more of it will be read. So competing forces are at work. May 27, 2011 at 16:38
  • 1
    I agree with your response. However, when I've just installed the application and it informs me that I need to restart after I've installed it for it to work with 100% certainty, this feels like I don't have any option for the application to work then to reboot. Clicking not to reboot and continue and eventually not being able to work in a pleasing manner, does imply a situation where a previous action (the installation) can not be undone. Hence, my advise. :)
    – Jeffrey
    May 27, 2011 at 16:57
  • If no other applications are running, then this application that is in question probably will install and not really need the reboot. It's hard to say. But any applications with Winsock already loaded are likely not to function once this new software is installed. That's why we encourage people to reboot. We also encourage them to quit out of any applications before they install the software, but who ever does that when instructed. I know I don't myself. "Eh! It'll install just fine! <Click>"
    – Fauxcuss
    May 27, 2011 at 20:22
  • Ok, this sounds like there's another solution. How about scanning for usage of Winsock ports and inform the user that applications are using Winsock. To continue the installation, please quit the following programs: ... ? How about that? That would also solve it! :)
    – Jeffrey
    May 27, 2011 at 20:55
  • Yes, that might solve it. Though, unless we prevented them from installing the application, they might blow through the message and install anyways.
    – Fauxcuss
    May 29, 2011 at 15:45

Show just the one message.

The user has made up their mind about not doing what you said. Asking them again is quite rude, and rebooting is very disruptive. Remember that the user uses his computer for 100 things, only one of which is your software--no matter how important you might think it is.

However, that first dialog should make the consequences of their action CRYSTAL CLEAR. Perhaps you even have a check box that they have to tick, like those annoying click-through licenses.

Personally I don't think "stuff might not work" is dire enough even for that. (I guess it might lead to a lot of support calls.)


From personal experience, every time I am asked to reboot after installation i choose not to and prefer to wait until later. His statement about users not wanting to read long text is true. I would most likely not read to much of the text and assume it is the same old message box. Then when prompted a second time, I would take it more seriously and listen to it, as it is presented in a way reflects it's importance.

I do understand that this is annoying and invasive, but my concern is that no matter what the first box says - annoyed users and those who don't care to read will just click no. At least the second box would give them an opportunity to find out the important info.

Example: A lot of people speed while driving. They see the speed limit, but speed anyway. If they come across a cop doing radar, typically they slow down and pay attention. Same concept here.

  • What about people who read the first message box and understand it and click "No". Then they get nagged a second time? I get that popping it up a second time will increase the chance that we'll get the message across, but... seriously?
    – Fauxcuss
    May 27, 2011 at 15:39
  • @Mark, because by reading the first box, you would not know that the lack of a reboot would lead to network problems. May 27, 2011 at 16:02
  • 1
    For the cop analogy to work you can't have the same text twice. That's like having the speed limit sign twice. May 27, 2011 at 16:23
  • @James, not the same text:" 1. Ask them to reboot (recommended). If they answer "No" then, 2.Tell them that they may have problems with network applications if they don't reboot. Ask them to reboot (recommended)." May 27, 2011 at 16:26
  • 1
    @Matt, well, maybe I have to make it such that the first message box doesn't appear to be the "same old message box", so that they will pay attention the first time. What if the first speed limit sign had an image of a cop shooting radar along with the speed limit. The problem with the copy analogy is that if they're still speeding, it's too late; they have been caught. With our app, if don't don't heed the warning to reboot, they may end up experiencing the problem we are trying to avoid. Answer this: If 2 message improves things, why not three... or four? I'd rather have 1 effective msg.
    – Fauxcuss
    May 29, 2011 at 15:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.