User Experience Stack Exchange is a question and answer site for user experience researchers and experts. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

This is purely a curiosity question but one I thought would be a bit interesting. Is anyone aware of placebos in use on popular web sites or applications? Either intentionally or unintentionally?

Some examples of physical world placebo buttons:


locked by JonW May 9 '12 at 18:01

This question exists because it has historical significance, but it is not considered a good, on-topic question for this site, so please do not use it as evidence that you can ask similar questions here. This question and its answers are frozen and cannot be changed. More info: help center.

Lovely question, and thanks for alerting me to placebo buttons (I preferred the alternate description in Wiki....'idiot buttons'. My first thought was 'OSX is one big idiot button' ;) – Nick Fine Jul 22 '10 at 8:10
It seems that many people have misunderstood DA's question. A placebo button is one wich does exactly NOTHING when clicked. That's not the same as an unnecesary button (with an actual effect or function). – Luis Parker Jul 23 '10 at 19:36
Ever seen teenagers play DotA? click click click click click click attack – pate Jul 4 '11 at 7:46
I can't think of a single one - mind you that elevators in this country doesn't have a close button (none that I've seen during my life) and zebra crossings that automatically "press" the button will show the button actually in a "pressed" state when it does that and will then not respond to physical presses - like someone pressed it for you. – Oskar Duveborn Oct 26 '11 at 18:03
If there really were a placebo button, I would go any lengths to have one! I could always tell my boss when he wants the results yesterday: "Oh yes, Sir, I already pressed this button you see, and the results must be here any moment now. I am just waiting as anxiously as you are!" And, wait I would... (Then again, that would make the button a no placebo anymore then?) – Kris Dec 6 '11 at 4:16

29 Answers 29

Dan Lockton's carried out some interesting research into this area.

Here's a real-world example that illustrates the point well:

elevator close button

Often, pushing this button does nothing to accelerate the door close mechanism, but it does let the user feel they're in control.

Interesting, all the door buttons I've ever pushed in elevators always makes the doors open again if they're closing (and resets the close timer if they're opened, delaying the closing) - I always reach for that button if I see someone sprinting towards the elevator I'm in. Elevator buttons is an annoyance in general though - if someone hits a floor button by mistake it cannot be undone... – Oskar Duveborn Oct 24 '10 at 9:31
In some elevators, a double-tap will turn off a floor button. Worth a try next time you get in an elevator to find every floor lit up. – Bevan Jun 2 '11 at 20:56
I've never had a close button do anything, but the open button typically works. – Nick Bedford Jul 4 '11 at 4:43
A more in depth discussion about the close button in elevators: – htanata Jul 12 '11 at 10:30
the elevator close button in my former college actually worked (the doors would take a long time to close if you didn't press it). – João Portela Sep 22 '11 at 12:27

On the ecommerce package I work on we have a placebo button in the form of the update quantity button in the shopping basket.

To work around a technical limitation (to do with submitting quantity adjustments to multiple order lines at once) we needed to submit the form and update the quantity when the qty field lost focus. The update button was simply there as a target for the user to click to then lose focus on the qty field.

That's an interesting one! – DA01 Jul 12 '11 at 13:56
So it serves a purpose, doesn't it? Other than making the user mighty pleased, that is. – Kris Dec 6 '11 at 4:12
a workaround this is to trigger after a delay, that's what autocompletes use. – teebot Dec 16 '11 at 10:08

The Print button on webpages is sometimes a placebo. Many sites use a print stylesheet so that you can simply File/Print the page and a nicely formatted version will print. But many (most?) users expect the printout to be pretty much the same as the screen view.

So to manage people's expectations, some sites have a Print button that acts as a Print Preview command. I call this a placebo because it's not a real print preview -- its purpose is simply to reassure the person using it that the printout will be nicely formatted. (The browser provides a real print preview that takes paper size and so on into account.)

+1, good example – Zoltán Gócza Jul 23 '10 at 12:10
this is a good example – Phelios Jun 29 '11 at 3:34
+1 I do this as well – wildpeaks Jul 1 '11 at 13:31
I'm not convinced that this is a placebo. A placebo is something that doesn't actually do anything but has a (psychological) effect because people believe that it is efficacious. Your example reassures users by providing them information. – Splog Jul 8 '11 at 9:12
@Splog A fine distinction indeed. :) – Bennett McElwee Jul 10 '11 at 21:50

A Save button in many services (like Gmail or Google Docs for example). Since there is autosaving, the button works mostly like a placebo.

Believe it or not but I actually use the save button because I find that sometimes the gap in between autosaves is too long e.g. I'm typing a mail and halfway through I need to do something else in Gmail. Instead of clicking the "Open in New Window" button, I click save and then come back to the message later. – AndrewJacksonZA Jul 23 '10 at 8:35
But that's not a true placebo any more than the save button in MS Word is a placebo... sure, most of the time, what you're working on is auto-saved, but only at predetermined intervals. If I need to save a draft email RIGHT NOW, I need the save button (otherwise Gmail throws up a warning about "navigating away from this unsaved draft". – Dan Newman Jul 23 '10 at 14:30
I agree with Dan and Andrew here. The "Save" button is still useful for saving "right now" instead of waiting for the auto-save to kick-in. It would've been a placebo if those services immediately saved your changes. Even then I think it would be better to have a "Save" button. Personally I feel safer when I save my work myself for some reason. It's that false sense of being in control, but it works. – Mashhoor Aug 2 '10 at 3:11
Even if Gmail / MS Word autosaved after every single character change, my betting is there would still be a 'save' button and people would still use it. – Bobby Jack Aug 3 '10 at 23:11
The save button does WORK though, so it;s not a placebo per se. It also functions as a save indicator, which is invaluable in case the autosave fails (in case of a timeout ect) – Ben Brocka Sep 9 '11 at 18:27

A few years back I worked on a form for a website. At the start, the form included a simple straight up promise statement that the company involved would do X and wouldn't do Y. We changed that statement to then include a checkbox, so the user could selectively acknowledge the statement. It didn't matter at all what they selected, the company wasn't going to do Y and would always do X.

The customers felt better, completion rates increased, abandonment rates dropped.

Was it a placebo though?

Similarly, shrink-wrap EULAs UIs are more or less placebos as (I've heard) they often hold no water in court. – Merlyn Morgan-Graham Dec 14 '11 at 7:17

We use placebo input, but not explicitly a button. It's the same idea, though; the user is prompted to input extra data that isn't essential to the final result, but without this extra bit of input, the entire process would seem incomplete to the user and [s]he might question the legitimacy of the result.

The non-essential data acts as a placebo to give the user the feeling that all the entered data was taken into account (thus the feeling of a more accurate result), vs. the relatively small subset of everything they entered.

Interesting. Can you share some more details? What unnecessary field was added? – Patrick McElhaney Nov 4 '11 at 13:30
Hypothetical: imagine you are applying for our very selective program through our website. The application wizard walks you through 3 steps in which you are prompted for gender, followed by zip, and finally birth date. Say we only accept females from 90210 born post-1985, so we could immediately reject a user for choosing “Male” in step 1, but to protect our selection criteria, and especially to create the affect that we consider everyone, we have them complete each step so it feels like all their information was taken into account when they receive their acceptance or rejection. – Tim Nov 4 '11 at 18:55

Recently we were designing a form in a workshop with a client and we discovered a design issue that demonstrated some of these competing demands, and we ended up with perhaps an unusual solution. The form, like many others, required some disclosure of the financial position of the customer. Part of this was their employment details. The form we had sketched had fields for 'employer' and 'job title'. Our client explained that actually they need the 'job type' and not the 'job title'.

'Job titles' often carry a sense of identity and can infer status (there's more Senior and Principal UX professionals than Standard ones) whereas 'job types' are averaging. [...] We felt that swapping 'job title' for 'job type' would take too much shine off the emotional well-being of our customer during this particular engagement. We decided to leave in 'job title' so people could tell us something special about their work, and then collect 'job type' straight after.

Job title: Queen of the Cubicles, Job type: Office/Clerical


We're having a hard time coming up with a perfect example. Even if they're out there, it would be hard to notice them. If you find a button that doesn't do anything, you think it's broken. If you don't realize it's not doing anything, you're not going to think of it here.

So, here's another imperfect example. When I log in to Wells Fargo, it always nags me to open a brokerage account before I can look at my accounts. The choices are "remind me later," "not this time," and "apply now." I don't know what the difference is between "remind me later" and "not this time." From what I've observed, the main difference is the button label. So, it's not so much a placebo button as a false choice (not listed as an option: "no, thank you"). The intent still appears to be to create an illusion of control where there is none.

Nags? You mean 'upsell' right? ;) – DA01 Oct 21 '10 at 14:13
Would 'remind me later' not remind you during your session at some point, or trigger 'reminder' emails? That's what I assume the difference between remind me later and not this time meant. – Anonymous Sep 25 '11 at 10:47

Where Google Instant is being used, the search button is kind of a placebo: you don't need to click it after updating the field. The button might be retained in case the browser doesn't communicate back to the server properly, but otherwise it's not really good for anything else.


In the UK, where there is a pedestrian crossing next to a traffic light-controlled junction across a one-way street (with traffic approaching junction), the green man will usually show when the traffic has been stopped at a red light, whether the button has been pressed or not. Pressing the button does nothing to speed up the traffic being stopped. But not having the button would be confusing for pedestrians as they are used to having a button to press.

In the image below, pedestrians will press the yellow button on the crossing even though the traffic on the one way street will soon be stopped to let the two-way traffic proceed, and the green man will be shown whether the button has been pressed or not.

This is the case for most fixed-cycle traffic lights; note that some modern traffic lights are variable-cycle: that is, the phases can be shortened, prolonged, inserted or dropped depending on the traffic detected (e.g. an extra phase for light rail only needs to be inserted when there is actually something approaching on the rails). Contrary to vehicular traffic, detecting pedestrians is tricky, and giving them an explicit "signup" button works reasonably well. – Piskvor Mar 23 '12 at 10:10

The one instance of obvious placebo buttons that comes to mind is in fake dialogues that appear in all kinds of spam banners, like "you won $1,000,000" with the buttons "claim" or "cancel". The entire banner is a link, but the buttons are just "painted on", they aren't real.

Not a placebo button. It does have an effect when clicked, namely taking you to a spam website. – Mechanical snail Aug 23 '11 at 7:26
No, the button doesn't even exist. The whole banner is clicked, the button is just "painted" on, there's no button there. – Vitaly Mijiritsky Aug 23 '11 at 9:49
Those buttons would be better described as Red Herrings or misleading elements. – Todd Sieling Sep 9 '11 at 18:10
@ToddSieling I'd say that all placebo buttons are misleading elements :) – Vitaly Mijiritsky Oct 17 '11 at 7:24
The most annoying of those banners vibrate as if they were about to explode. – kinokijuf Dec 8 '11 at 19:48

On some Android apps there's an "exit" placebo button.

Some people can't stand the fact that you don't close apps in Android, so some developers put a button or menu item in their interfaces that doesn't actually exit, or close the app, it's just like pressing "back" button. The app will still stay in memory as much as the OS likes.

I hate how people equate more memory usage to a slower experience. It's quite the opposite. Just because the app occupies memory when you hit the home button, doesn't mean it's doing crazy calculations 24/7. It could, but only a few apps need to do background processing. Pandora, for example, needs to continue playing music in the background. For all other apps, keeping it in memory helps make it restart faster because RAM is 1,000 times faster than the harddrive. – JoJo Nov 19 '11 at 17:11
@JoJo I guess it's an idiosyncrasy due to when there wasn't much RAM, and avoiding swapping memory could save a lot of pain. – bigstones Nov 19 '11 at 17:50

Office thermostats are, more often than not, placebos.

As for the "close door button on elevators discussion" above, the same source states that they

[…] won't work unless you're a fireman or an elevator operator with special access to the system. The rest of the time, in deference to various building codes, it's deactivated, according to engineers at Otis Elevator.


Internet veterans may remember: "THE REALLY BIG BUTTON THAT DOESN'T DO ANYTHING"

It is pure placebo at its finest.

Definitely. I still can't resist to click it to figure out if something actually happens. Don't know if I could call it "placebo" as it is quite obvious it does nothing... but still. – Alpha Dec 14 '11 at 0:26

I've never thought of it specifically as a 'placebo button' but some parts of a check-out process do this TWICE - once as buttons/elements and once as fields. When asking for 'Credit Card Type' and 'State and City' (when asking for locations in the USA.)

[Note: I went searching for quick examples but it may take a while to find some of these and screenshots of e-commerce sites is an awkward thing, even with enough coffee.]

Almost every e-commerce platform can identify your card type from the card number, selecting or typing 'MasterCard' or 'Visa' is superfluous. Early on most sites would ask for each but over the years more and more just ask for the number and fill the card type in after the fact. These sites will often display the card logos (the set of MasterCard, Visa, American Express and Discover Card) in connection with the field. In some cases they are aligned with the form fields and appear clickable, even having hover states.

When asking for address in the USA once you have the zip code you can auto-fill the city and state. More and more forms are using the zip to pull values from a database and ensure the address info will match the US Postal service. You can put whatever you want in the "City" and "State" field but the system will use the Zip field and replace or ask to replace those values, rendering those fields unnecessary and purely for placebo purposes.

If the card type you entered doesn't match the type of number, some sites/engines will reject the entry. It's more like a confirmation than any new information in those cases. – Merlyn Morgan-Graham Dec 14 '11 at 7:22

In many sites.

But the placebo effect is not for users, but for clients (site owners). All the time I have to implement buttons and functionality that I know are not going to be used. Think for instance a language selector in a site that preselects it based on the browser setting. I know (because I tested it) that it is never used, but it is still there, because it gives site owners a feeling that "every single case is controlled".

Another example. I am really tired of crappy global site search systems that (let's face it) are not up to par with Google's. Some of them deliver results so bad that people use it once, then fall back to menus. I never use a search box in certain kind of sites.

We still put them there, because it seems reassuring for users and for us as UX designers (how not to put it in?)

This isn't really an example of a placebo button, just a bad call when the time for cutting out features comes. If it's never going to be used, don't put it in there. Placebo buttons are there to reassure users that common actions (like saving) - which are performed automatically - are indeed being carried out. – Bruno Abrantes Jul 23 '10 at 9:16
Take the meaning of "placebo" as something that does not really work but having it feels reassuring. My examples fit it. – luna1999 Jul 23 '10 at 13:41
It's not a "placebo button" because it works. It is (mostly) a placebo feature, in that it is not necessary, but not a placebo button, in that it is still functional. – Daniel Jul 11 '11 at 20:18

The 'If this download doesn't start, click here" link could be a placebo. The download takes time to start, but it gives users a sense of control. Even if the user 'clicks here', most of the time we still have to wait a few seconds anyway.

I think this is mainly for those people who switched off JavaScript, where the download really does not start automatically (and will start if you click the button). – Paŭlo Ebermann Jun 26 '11 at 20:31
JavaScript is unnecessary; a delayed redirect works, too. – Mike L. Nov 19 '11 at 18:02
There's one of this cases in a popular filesharing site (although I cannot remember which one). The "click here" link would link to the same page so it loads again. Kind of like not doing anything if the download didn't start on the first try. – Alpha Dec 14 '11 at 0:29
@Mike L.: Both JS and a delayed redirect (e.g. with a meta tag) might be blocked by browser extensions such as NoScript; in that case, I'd like to have a link which actually points at the requested file and not an UI obfuscation - least surprise and whatnot. – Piskvor Mar 23 '12 at 10:15
Take a look at, for example, (after having the license accepted). There a meta redirect works fine for me without adding a rule in NoScript. – Mike L. Mar 23 '12 at 17:02

Might be a bit of a stretch, but how about the requirement to create a username on sites that then require you to log in using some other piece of information, such as your email address? Flickr effectively does this, IIRC. You get to give yourself a handle (everyone likes having and/or creating a username, right? It's all about establishing the identity you want for yourself), but one that's not really functionally useful. Drives me insane, incidentally, because I can never remember which credentials I should be logging in with on Flickr.


Similar to @Senor Swinstead answer, I once had to change the implementation of a Windows desktop option dialog so that the Apply button actually did nothing.

Turns out that those would work how you would expect:

  • Ok button saves changes and closes the dialog
  • Apply button saves changes, leaves the dialog open
  • Cancel button discards changes and closes the dialog

Turns out that users wanted the "Cancel" button to revert changes if they were already applied by the "Apply" button, and the changes were not visible without closing the dialog at all (those were configurations that drove the application logic).

My change was to make the Ok button work as normal, Cancel button work as normal and Apply do nothing. We actually kept the Apply button because users would feel reassured that their changes were saved, but they would make the final decision on hitting "Ok" or "Cancel".

You violated the Microsoft UI guidelines. Now your app won't get the Windows-compatible logo! – kinokijuf Dec 25 '11 at 9:44
@kinokijuf – Alpha Dec 25 '11 at 23:28

I once had to create a sort of Placebo view for Senior Management to view network operations status. We had a variety of monitoring systems in place to keep on eye on all sorts of things, but management couldn't be bothered to look at all of them. So, they declared that a magic super monitoring system would be created that would have all information in it in an easy to digest view. Only "important" alerts were to propagate to this UI, and senior management would check it every morning to verify that systems were ready for the start of business.

Actually integrating data from disparate monitoring systems, and presenting it in a way that a senior manager could usefully interpret with no training in a harmonious way like they wanted was flatly impossible. So, I created a static web page and a private web UI for updating it. It was a shiny green circle. If something bad happened, somebody from Operations could change the color of the circle to be yellow or red, and add a status message. It was functionally exactly equivalent to the NOC monkey sending an email, except that it had to be explicitly checked. But, it gave the execs a sense of false control over the flow of information because they could check it whenever they wanted.

It was declared something like the Strategic Enterprise Multimonitoring Infrastructure Status Tool.

If they had a shortcut on their Blackberry to the web UI of a green circle, the desktop shortcut would indeed have been in a sense a placebo button. (Any information on the status display would get sent out as an email before the NOC monkey bothered updating the web page, and the email would have gone directly to said blackberry. If it was an email outage, anybody with access to look at the green circle was on the phone tree and would have already been called. It could, by definition, never actually provide new information.)

More 'active' placebo buttons are hard compared to placebo information displays. (See also, "Has the LHC destroyed the world yet?") Something like a hospital call button works because it involves a person. The button could work, even if the person takes a long time to show up. We expect proof of action when we are using a computer. If we click save, we expect to be able to find the file some place. I think you can get away with it with something like an "optimise" button with nebulously defined methods. There is actually a class of scamware which takes money, claims to make your computer faster, and accomplishes nothing beneficial. But, some users will believe it worked because the placebo effect demands that they must have gotten something for their money. In some cases, repeated defragging of unfragmented volumes will also have exactly this effect.

Another example might be in "unsubscribe" forms, and "complaint" forms. abuse@ and postmaster@ email addresses. Anywhere that you want annoying messages to get dumped and sent to null so that nobody is bothered with them. A user still feels like they accomplished something by the contact.


We have a soup machine at our school, on which you can choose how strong you'd like it.
We tested it out tough, but putting the 'flavour-meter' on strong doesn't change anything compared to the less flavoured option.

Must say that it does give the idea of control.


Probably the most famous example of a placebo in software UI is the Diablo chat gem.

Even after it was confirmed not to do anything whatsoever in-game by the developers, players kept pressing it repeatedly before starting game sessions based on the belief that it increased the chance for more treasure (and specifically, gems) to appear.

This was compounded by the fact that pressing the gem had a very high chance to display "Gem Activated/Deactivated" and a very low chance to display "Perfect Gem Activated" or "Moooo". Players, of course, attributed special properties to the different printed messages, when they were confirmed not to have any.

A decade after Diablo II was released, the myth persists that activating a "Perfect Gem" increases loot/gem drops and there are still players who will not enter a game until they've managed to activate it.


I always thought that the "Click here if the page is taking too long to load" buttons were placebo buttons.

Not a placebo, surely. These links tend to do something useful. For example, "Your download should start automatically. Click here if it does not start with 10 seconds". – Bennett McElwee Jul 23 '10 at 1:09
Yes you're probably right. What I mean is that it sometimes feels like they don't do anything. It reminds me of that old myth that a certain brand of elevator installed the "Close Doors" button as a placebo button for those impatient people when in fact it did nothing. – jaslr Jul 23 '10 at 3:17
@jaslr: It does plenty if the page is actually wired to wait a few seconds (via Javascript) before proceeding. Just like the elevator doors, it seems like a case-by-case. – doppelgreener Jun 26 '11 at 11:33
I think in this case, sometimes the intermediate page itself is the placebo and there's no real need to show it other than to let the user feel like something is happening. – Wesley Murch Nov 5 '11 at 16:32
@Wesley Murch: Actually, sometimes such page is not useful to the user, but it is to the site: having a n-second wait before the download starts a) gives the user an option to cancel the download before hitting the server, and 2) alleviates (and/or distributes in time) the server load somewhat. – Piskvor Mar 23 '12 at 10:18

Here's a fun one: the "Search" button when Google instant is on. Now, it's not actually ignored, and google asks you to actually press it when you're searching for something that isn't safe search friendly, but most of the time, hitting the button doesn't actually do anything.


I strongly suspect the X button in iTunes' header is a placebo button, it never stops ongoing requests as far as I can tell:

enter image description here

It is actually there for dismissing the message, but yes, it is confusing. Can't tell what it does unless you have read all of the Apple UI guide – edgerunner Oct 26 '11 at 9:07

The Google Search is now effectively a placebo input box. Indeed, even when the input box has lost focus and you start typing, the window still listens to keyboard events and Instant Search kicks in straight away with the appropriate search.


We had a scenario where we needed a placebo option. It was in a form that had a list of actions that the user might take. Submitting the form without choosing an option performed a default action.

But we needed to add this default action to the list as a placebo so users weren't confused by it not being in the list. If you've got a list of actions it's easier to just choose one from that list rather than understand that the option you want is the default action of the form.

Choosing this option had no effect on what happens when the form is submitted. It tooks some explaining to the dev team that we wanted an option that did nothing :-)

I'm confused. How is this different from a set of radio buttons where one is selected by default? – Patrick McElhaney Nov 19 '11 at 13:23

Interesting question!

The closest I can think of is the delete button in some web apps I've worked on where it doesn't actually delete the selected data from the server, it just hides it from the user. This is done so the operation can still be easily 'undone'.

But of course the button still does something - it hides the data - so it's not quite what you're looking for.

Well, to define "deleted" isn't really about physically removing the data from the harddrive - which would require more than just normal operating system deletion of the files and data. It's deleted as far as the user knows. – Henrik Ekblom Jun 28 '11 at 7:50

Many Windows dialog boxes have Apply and OK buttons. Apply doesn't do anything that isn't done when you press OK anyway so Apply is completely unnecessary

This isn't necessarily so - I can't say for sure in current Windows implementations, but historically it has been inconsistent. 'Apply' actually writes the GUI actions whereas 'OK' doesn't in those cases. I used to be an MCSE for NT4 and many times you had to 'Apply' not 'OK' if you wanted the changes to be effective. – Nick Fine Jul 22 '10 at 8:08
Apply is like a Preview - do the action without closing the window. If you don't like the result you may change something and press Apply once again. – Kostya Jul 22 '10 at 15:20
Yes, Apply means "carry out the changes but leave the dialog box open", and OK means "carry out the changes and close the dialog box". So Apply isn't a placebo -- sometimes you really do need to click Apply instead of OK. – Bennett McElwee Jul 23 '10 at 1:12
Apply isn't really a preview in Windows. If you click Apply and Cancel in succession, the changes still get applied as if you pressed OK – edgerunner Oct 26 '11 at 9:04
-1 Being unnecessary has nothing to do with being a placebo. If it has an effect, then it's not a placebo. Being redundant (or not) is an entirely different conversation. – Phong May 9 '12 at 17:44

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