We've created Zuko, our next-generation form analytics platform. Explore Zuko Explore Zuko > Hide Message

Mistakes in forms – why do they happen, and how can you reduce them?

Form Optimisation

One of Formisimo’s most insightful reports is the Most Corrected Fields analysis. This gives you insight into where users have to change the data they have inputted. Some of these corrections may come when a user spots their own mistake, but many will come after a user tries to submit the form or continue the checkout, but is met with error messages. This blog shows you how to interpret Formisimo’s data, and give you some tips on why corrections might be occurring.

The corrections report

If you have Formisimo up and running (if you haven’t, you should), you’ll find it under Form Issues on the right hand navigation:

Screenshot 2014-04-22 13.52.53

Once you load up the report you’ll get a breakdown on which fields users have to correct most. You’ll also see the average corrections per visitor who starts to interact with the form:

Screenshot 2014-04-22 13.40.40

This is your benchmark, and something that you should aspire to reduce as much as possible. The fewer corrections a user has to make, the less friction your form causes.

Along the top of the table, you have the following columns:

Interactions – how many key presses and clicks have users made in this field?

Refocuses – how many times has a user left this field and had to come back to it to change or input more information?

Delete – how many backspaces and deletes have users had to press in this field?

Cursors – how many times has a user pressed a cursor key in this field?

Correction % – of all the interactions that take place in this field, how many are corrections?

Health – how does this field compare to the rest of my form and others in the same industry? Health scores use a percentage and traffic light system – the lower the percentage, the more problematic it is.

What causes users to make corrections?

So, now you can see where users are making corrections (and what type of corrections), but how do you interpret that, and what might be causing your visitors to make those errors?

  • Sloppy data input (by the user)
  • Unclear labels (including compulsory fields)
  • Unforgiving input requirements
  • Lack of good (inline) validation and error messages
  • It’s a captcha field

Sloppy data input

This is something that inevitably happens when users are required to enter certain, sometimes complicated information. Regardless of how well designed your form is, there will always been some mistakes made by users. Fields that will always have a good number of corrections will be email address, password and credit card fields. Even if you form is really well designed and very usable, a proportion of users will make mistakes and have to correct them.

The remedy?

You can minimise this through building a usable, clear and well designed form, but there will always be corrections. Aim to improve your current figures rather than focus on an absolute value.

Unclear labels

When users enter a new field, the label and any supporting text gives them the only information they have in order to complete their data entry successfully. Is your field label clear and understandable to every single user? Is there any ambiguity about what you’re asking for?

The Expedia example is a great one where users were confused about what they were being asked to enter. The field label read simply ‘Company’ but it appeared above their bank details. Users therefore thought they had to enter the name of their bank. Validation issues ensued, and users dropped out of the form. Expedia simply got rid of this field.

The result? A huge $12million in extra revenue in a year from the subsequent increase in conversions.

Another example from one of our users, who before the credit card section of their form, asked users to enter their cardholder name. The problem was that in order for the card payment to be processed, the name has to be the same as it appears on the credit or debit card. A high correction rate was the result:

Card holder name above had an 7% correction rate, meaning that 7% of all interactions with this field are what we class as corrections.

So the user’s name might be ‘Mrs Catherine Anne Wigglesworth’, but on her credit card it appears as ‘Mrs C A Wigglesworth’. If this user entered the former, her order would fail and she would have to correct it. The simple addition of ‘(as it appears on the card)’ reduces this problem substantially.

Another problem we often see in this category is when fields are not clearly indicated as optional or mandatory. Users will for the most part do as little as possible in order to successfully complete the process. If mandatory fields are not marked clearly, users may to try and skip over ones that they perceive to not be necessary to move forward. They’ll be greeted with error messages and have to spend more time than they’d like completing your form.

The remedy?

Clear, unambiguous labels and helpful support text where needed.

Unforgiving input requirements

If your form requires that information be entered in a certain way, you have to ensure that you do everything possible to let the user know this, or force them to enter it in a way that they cannot get wrong.

We mentioned this in reference to IKEA’s online checkout, in which you could not enter a UK postcode without any spaces.

Screenshot 2014-04-22 13.41.13

One user of Formisimo has a similar situation where a user could not enter a postcode without spaces. The result?  Lots of corrections, with a 12% corrections rate in that field (meaning that 12% of all interactions in this field were correcting previously entered inputs):

Another user of Formisimo did not allow users to enter their card number if they added spaces in between the groups of four numbers, like this: 1234 5678 1234 5678. It was the most corrected field in their form, with a very high number of re-focuses (where a user has to go back into the field to correct something) and deletes (probably to delete the spaces they entered originally).

If your error rate is high for one field, it may therefore be because you do not accept information in a way that many users expect to enter it. Let’s think of some of the possible permutations of the following information:

Telephone Number 0800 772 0982
(0)800 7720982
(0)8007720982
08007720982
+448007720982
+44 (0) 800 722 0192
+44 (0)8007220192
Date of Birth 01/01/1970, 1/1/1970
1/1/70
01/01/70
01-01-1970
Card Number 1234 5678 1234 5678
1234567812345678
Postcode M17 1DZ
M171DZ

Does your form accept all of the above as valid?

The remedy?

Be more forgiving with how data can be entered.

With a little bit of input from your developers, you can accept inputs in a variety of formats so your users with see less validation errors and be able to complete your form more easily.

Give clear instructions on data entry. 

Let users know before they start typing that data must be entered in a certain way. Yahoo’s password field in their sign up form is a good example of this:

Screenshot 2014-04-22 13.41.27

Force users to enter information in a particular way (that is not intrusive).

Gmail asks for your date of birth, but minimises potential errors by splitting the information into a drop down for month.

Yahoo’s form does the same for day, month and year of birth:

Screenshot 2014-04-22 13.41.38

Lack of good (inline) validation and error messages

Inline validation is a great way to highlight where users have gone wrong with entry and ensure they get it right second time. Even if you do not have inline validation, your error messages should be helpful and instructive (head to this article to see more on this).

If you only highlight errors but do not pinpoint exactly where the user has gone wrong and provide guidance, they may end up trying to enter information more than twice.

Example:

One Formisimo user had a Username field at the start of their sign up form. The problem was, the site did not let users know if the username they chose had already been taken until they completed the rest of the form and clicked submit. The message told the user that the username was already taken, but gave no guidance on ones that were available. Users would have to try several times to find one that would work. It was the most corrected field in the entire form.

The remedy?

Provide clear, unambiguous and instructive error messages, and if possible provide this as the user works through the form (inline validation).

NowTVs sign up form is a good example of providing useful feedback to make sure they get it right on the second attempt:

Screenshot 2014-04-22 17.23.36

As a user, I know instantly where I have made my error, and what I can do to correct it.

It’s a captcha field:

Captchas fields are without fail the most corrected field we ever see.

Screenshot 2014-04-22 13.41.49

The remedy?

Read this and consider alternatives.

Friction is bad.

Use Formisimo’s data to show you where users make mistakes, then head back to your form and find out why it might be happening. Then fix it. Corrections made Cause friction and frustration for your visitors and you’d hate for them to hate you whilst using your form.