Take the 2-minute tour ×
User Experience Stack Exchange is a question and answer site for user experience researchers and experts. It's 100% free, no registration required.

The company I work for rents out vacation properties, with an alphanumeric ID for each property. Such as http://www.example-domain.tld/abc123456

There are SEO recommendations for using a slug / permalink with 3-5 keywords in them instead. Such as http://www.example-domain.tld/vacation-property-in-italy/

The internal (company) reason cited to keep the ID as-is, is for better user experience, with the explanation that if users edit the URL, they will encounter an error page. However, this got me thinking as to what percentage of users actually do (and know how to) change the URL of a specific page. I was unable to find an answer, searching online, and figured that this would be an excellent forum for it.

Is there any statistical information on whether users manually change the url in the browser? Are there any user tests that have been published, with information about whether people even know how to change the URL?

EDIT: A number of comments relate to people's own practice in modifying URLs - I appreciate the feedback on that, however I am looking for information (statistics, studies or similar) that support either the claim that users do or do not (or don't know how to) modify the URL of the page they're on. With the stats in hand, improving the user experience with regards to the URLs is easier.

share|improve this question
1  
I'd love to see that info too, but I can't think of any way to actually collect it. AFAIK Browsers don't record when a user interacts with the url bar, so there's no way for us to collect the data. –  Racheet Feb 26 at 12:34
    
No, there's no way AFAIK to programmatically collect this information. However, it would be very interesting to see any references to studies or other data that highlight user behaviour when it comes to the interaction in the address bar of the browser. –  Calle Feb 26 at 13:31
3  
I'm unclear about the rationale here - how is the user editing a jumble of characters going to work out better than them editing a set of words? Either way, they end up with an error page unless they know exactly what they are doing. –  Michael Kohne Feb 26 at 13:41
    
In this modern era, I'm more inclined to edit an url that I can understand than one I can't. If I see "vacation-property-in-italy" I may edit and try "vacation-property-in-spain", but if I see "abc123456", I'm not going to try alternatives because I assume that the ID is generated, kind of, randomly and whatever I try is not going to take me anywhere. Also, if I can read/understand something, I know what to change/try, but if Idon't understand what I see, I don't know what to try. –  PatomaS Feb 26 at 13:43
1  
One thing worth noting is the people who are likely to edit URLs are generally going to be savvy enough about what they are doing to not be surprised or terribly disappointed by an occasional 404 when doing so. –  aslum Feb 26 at 14:34

4 Answers 4

up vote 5 down vote accepted

I don’t know any general studies about it, and I guess it would be very hard to come up with a sensible one. I think it’s safe to assume that this very, very much depends on a) target group and b) URL design.

So even if users regularly manipulate URLs, this doesn’t mean that they do it on every site, because often the URL design is not good, i.e., the URLs are not browsable at all.

Why? (Motivations)

I think there are three main motivations for manipulating the URL:

  • Navigation: Users know (or assume) how the URL design works and want to reach a specific resource.
  • Error correction: Users try to fix a URL that leads to a 404 error page, because they think the page should exist and they assume that the URL might contain an error (e.g., missing a part, mistyped something, etc.)
  • Curiosity: Users try to find (possibly "hidden") content, but they don’t know if it exists at all.

Other motivations might be: trying to make the URL shorter (removing unnecessary parts); removing referrer parts (e.g., from affiliate links); removing identifying parts (e.g., session or tracking IDs); ….

How? (URI parts)

Host / domain

Users might try to remove or add subdomains, second-level, and top-level domains.

Examples:

  • Switch from the English to the German Wikipedia by replacing the subdomain en. with de.
    1. http://en.wikipedia.org/
    2. http://de.wikipedia.org/
  • Switch from Amazon France to Amazon Germany by replacing the top-level domain .fr with .de
    1. http://www.amazon.fr/
    2. http://www.amazon.de/

Path

The slash / is probably the most used delimiter for path segments.

As the URI standard says (emphasis mine):

The path component contains data, usually organized in hierarchical form, […]

So for navigational purposes, it can expected that users

  • remove path segments (delimited by /) from right to left
  • add path segments from left to right, delimited by /
  • replace strings between /

Examples

  • Switch from the "php" to the "python" tag by replacing the corresponding strings:
    1. http://stackoverflow.com/questions/tagged/php
    2. http://stackoverflow.com/questions/tagged/python
  • Remove the whole path to reach the home page:
    1. http://en.wikipedia.org/wiki/Stack_Overflow_%28website%29
    2. http://en.wikipedia.org/
  • Append a segment to reach the Edit page:
    1. http://any-wikkawiki.example.com/FAQ
    2. http://any-wikkawiki.example.com/FAQ/edit

Query

Typically more complex than the path.

It’s likely that only users that know the typical format (each key-value of the formkey=value, the first one introduced by prefixing ?, additional ones by adding & as a delimiter) add/remove whole key-value pairs.

Others would probably only replace existing values.

Examples

  • Go from page 2 to 5 by replacing the value of the page key:
    1. http://example.com/products?page=1
    2. http://example.com/products?page=5
  • Switch to French translation by adding the whole query component:
    1. https://support.google.com/webmasters/answer/47334
    2. https://support.google.com/webmasters/answer/47334?hl=fr

Fragment

Unless the site uses JavaScript to show different pages when changing the fragment, users can manipulate it typically without "breaking" something. They might try to add the whole fragment component (prefixed by #), replace the value of an existing fragment, or remove the whole fragment.

Examples:

  • Reach the top of the page by removing the whole fragment:
    1. http://tools.ietf.org/html/std66#section-3.5
    2. http://tools.ietf.org/html/std66
  • Replace a value to jump to a specific section:
    1. http://tools.ietf.org/html/std66#section-3.5
    2. http://tools.ietf.org/html/std66#section-4
  • Guess that a page allows to jump to a specific section by adding a fragment:
    1. http://de.wikipedia.org/wiki/GeSHi
    2. http://de.wikipedia.org/wiki/GeSHi#Weblinks

Your example

For your example case, only the URI path matters, so it’s /abc123456 vs /vacation-property-in-italy/.

If the motivation is Navigation:

  • The cryptic path doesn’t allow anything except removing it completly to reach the home page.
  • The readable path additionally might allow users to switch to similar pages, if they assume/know that they exist (!), by replacing values, e.g., by replacing italy with france.
    • I don’t think that users would assume to find something when removing only parts of the dash-delimited path segment (e.g., from /vacation-property-in-italy/ to /vacation-property-in/ to /vacation-property). This would be very uncommon as almost no site allows this (as this is typically only possible with /)

If the motivation is Error correction:

  • The cryptic path doesn’t allow anything except "obvious" cases like removing whitespace or trailing punctuation.
  • The readable path allows various kinds of corrections, e.g., when users see /vacation-property-in-italyy/ they might try to fix it by replacing italyy with italy.

If the motivation is Curiosity:

  • The cryptic path invites users to try to find other pages by changing the ID, maybe with the hope to find hidden/private content, especially when it looks like the number is incremental.
  • The readable path doesn’t convey that there would be something users wouldn’t be "allowed" fiddle with (like it’s the case when cryptic IDs are used), so there is probably no motivation to find private content. So the only thing users might try would probably be replacing some words (similar to the Navigation case).

Conclusion:

  • In the Curiosity case, anything goes, no matter how the path looks like. Users probably never expect that their manipulation works at all.
  • In the Navigation case, the readable path wins if there are such related pages. If not, it doesn’t matter (then it would fall into the domain of the Curiosity case anyway).
  • In the Error correction case, the readable path wins.

So the company is right, the readable path probably leads to more users trying to manipulate it. But the company is wrong in assuming that this leads to a worse user experience. If the URL/link contains an error (for example, because some characters were stripped in a mail client, or because it was dictated via phone, or because someone mistyped it), then users have a chance to correct it. In all other manipulation cases, users know that they might and expect they will stumble upon error pages.

(Apart from SEO and manipulation, readable/descriptive URLs offer, of course, various other benefits.)

share|improve this answer
    
Thanks, @unor. Even though not citing studies or statistics, this outlines the underlying issue, with various considerations. Marking as best answer. –  Calle Feb 27 at 12:20
    
@Calle: Thanks, but please feel free to snatch it from me, e.g., when someone comes up with study citations. –  unor Feb 27 at 12:29
    
absolutely, I will keep track of the question and evaluate any new reply along the lines of the initial intention with the question. If a better answer is provided, it will be marked as such. :-) –  Calle Feb 27 at 12:30

I don't imagine many "Edit" the URL but most people do copy and paste, or click URL's. When a URL is abc123456 it's easy to be mislead or have to click on every link until you find the one which is actually the one you want. URL's are made to be read by humans, they aren't for the computer, they're for US. Having a human readable URL is a long established usability benefit.

http://en.wikipedia.org/wiki/Clean_URL

share|improve this answer

The URL should ideally be tied to the information on the page in a way that the user can know what information a page contains prior to clicking on it if the link is visible somewhere or if they save it.

If users modifying the url is a concern, try the following, which is a combination of both:

  • www.example-domain.tld/Property-abc123456-vacation-property-in-italy

This gives a unique URL for each property available, but still allows for descriptive text in the URL and page title.

With that method, I would assume the page title would be similar to "Property abc123456: Vacation Property in Italy

share|improve this answer

Do it like stackexchange does and combine the ID with a friendly URL:

 http://www.example-domain.tld/abc123456/vacation-property-in-italy/ 

When the user inputs an URL where ID and title don't match, the ID should have priority - redirect to the page with the same ID and the corrected title. That way you keep the ability to navigate by ID number and avoid name conflicts, but still have URLs which are human-readable and search-engine-readable.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

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