A list to run through to help maximize your form or checkout conversion
Knowing where to start with your form optimization project can be difficult. We often get asked how to plan and go about identifying the issues on forms and fixing them. To help with this we've pulled together a checklist to guide you through the processes we use. You can download a pdf version of the form optimization checklist or read all the points covered below:
Whilst you can run form optimization programmes using heuristics and “proven” patterns, it’s much better to make changes based on data and insight. To this end, you’ll want to make sure that your data systems are set up and working properly.
- Check form is live on your website.
- Test submissions are being delivered to the back end.
- Ensure form analytics software is installed.
- Create tracking attributes for relevant audience segmentations (device, operating system, browser, product, traffic source, etc).
- Let the form tracking run for the length of time needed to collect enough data to make statistically significant conclusions (this varies by traffic volumes but you can use one month as a rule of thumb).
Analysis and Diagnosis
Once you’ve gathered enough data you can use it to identify where your users are struggling and make inferences about why.
- Check data on total abandonments - where are the biggest volume of users dropping out of your form?
- Examine abandonment rates - are there fields with low interactions that have a high rate of dropoff?
- Which error messages are triggering the most? Are they associated with sessions where the user abandons the form or ones where they successfully complete it?
- Which fields do users have to return to in order to make corrections?
- Which fields do people spend the most time completing?
- Segment all your data (field returns, time spent, etc) between sessions that abandon and those that successfully complete. If there is a significant difference between these cohorts for a particular field it is an indication there is a UX issue with that question.
- What do users do after a failed submission? If they jump back to certain fields, it’s a good indicator that they struggle with them.
- Segment abandonment rate data on a field level. Look at audience groups like browser, device type, operating system, traffic source, etc. If a field has a significantly higher abandonment rate for a particular audience you’ve identified an issue for that user group.
Break the Form
Data goes a long way to showing you where the issues are but you’ll also want to have some “hands-on” time with your form to try and identify and replicate any UX problems.
- Pull up your form in the relevant format. Eyeball it to see if there are any obvious issues that are apparent.
- If you’ve got session replay, watch sessions for the relevant audience group that include an interaction with the problem field.
- Submit the form blank. What message does the user get? Is it clear what they’ve done wrong? Are they told how to fix it?
- Miss out certain fields and try to submit. What happens? Are all errors shown clearly? What do the errors show? Are they easy to understand?
- If your form is a multi-step journey, go forwards and backwards multiple times to see what happens. Is data wiped or kept? Does a customer have to repeat information they’ve previously entered?
- Leave the form for 20 minutes. Are you timed out? Are users warned about this at any stage?
- If there is a number field, try to enter letters and symbols. Does the form accept or reject? How clear is the error message?
- Do the same with numbers in a text field.
- Test the address finder. Enter 5 different addresses. Do they appear as expected?
- Use spaces and brackets in phone numbers. Does this cause validation issues?
- Deliberately misformat emails, phone numbers, credit card numbers, etc. Enter too few and too many characters to see how the form handles it.
- Password fields - put too few letters, only letters, basically ignore any best practices and see what the error messages are (if any).
- Buttons - is the call to action clear? Do they work all the time? Do inactive buttons look inactive or do they seem like you can click them and get a response?
- Repeat the above points on different devices (mobile, desktop, tablet, etc).
Create Improvement Hypotheses
Data analysis will show you where the UX problem may be. You’ll now need to bring up your form alongside the data to hypothesise what the exact issue is. This list gives you a set of common issues you may see.
- Is the form in a single column (rather than multiple ones)?
- If the form is longer than 4-5 questions is it broken into multiple stages?
- Are there any fields that are unnecessary at this stage and can be removed?
- If you are asking for information that the user won’t have to hand (such as SSN or passport number), do you let them know this upfront?
- Do you have a progress bar that accurately shows the user how far they are along in the process?
- Are you asking for sensitive or overly personal data that the user might not want to share?
- Is the form visually complex or intimidating?
- If the form is long is there a ‘save’ button?
- Are your questions in an optimal order? Could changing the order or removing some be beneficial for the user experience?
- Are you using dropdowns for fields with 5 or less options (use radio buttons instead)
- Is it clear which fields are optional and which are compulsory?
- Are your dropdowns searchable?
- Are text boxes sized according to the information needed?
- Are field HTML types set to the appropriate format to ensure the correct mobile keyboard is displayed?
- Are you utilising mobile’s native features where it would be beneficial? (camera, voice, location services, biometrics)
- Do you have a nuke/reset button that deletes all completion progress?
- Have you ordered the fields so that the “easy” questions come before the more difficult ones?
- If it is an insurance form, are you asking for personal details too early (consider asking after they have the quote)?
- Have you audited your form for accessibility?: Use tabbable navigation | Label your fields well | Don’t use icon fonts | Avoid dropdowns if possible | Use a vision preview tool
- Password: Are your requirements too restrictive or not explained clearly? Are you requiring password confirmation? Do you allow password unmasking?
ZUKO RECOMMENDATION - Have a minimum character requirement (usually 8) but no maximum. Special characters and numbers should be allowed but not mandated. Do not force password confirmation. Ensure that the password can be unmasked during input.
- Phone Number: are you forcing the user to add spaces, brackets, and international dialling codes? Or are you preventing them from doing so?
ZUKO RECOMMENDATION - Use a text box with a minimum of 8 characters. Let the user enter the number how they want and leave it for your phone staff to interpret.
- Zipcode / Postcode - Are you requiring or blocking certain elements they would expect (spaces, letters as well as digits).
ZUKO RECOMMENDATION - Use an automated address lookup.
- Email - Do you block unusual domain names? Or prevent the user of Gmail?
- Date of birth - Are you using dropdowns forcing users to scroll to find their birth year?
ZUKO RECOMMENDATION - Use a three text box format.
- Employment details - Is your list of options too narrow meaning users can’t find a satisfactory match?
ZUKO RECOMMENDATION - Use a free text box with a smart search function.
- Income / wealth - Does the user have to research and calculate an exact figure?
ZUKO RECOMMENDATION - Use radio buttons with income / wealth brackets rather than exact figures.
- How long have you lived at this address - Are you forcing the user to remember their history?
ZUKO RECOMMENDATION - Only ask for month of moving and auto-calculate the length of time for them.
- CAPTCHA - Are you using outdated technology that increases friction?
Errors and Validation
- Are error messages displayed next to the relevant field or does the user have to search them out?
- Are error messages specific about what the issue is?
- Are the messages helpful - do they tell the user what to do to fix things?
- Are you using accusatory rather than neutral language in errors?
- Are you using industry jargon rather than simple language?
- Are error messages being triggered too early (before the user has completed what they want to enter)?
- Are errors being shown inline after entry or only on submit?
- Are you using positive (green) validation when the user successfully completes a field?
- Could you use microcopy to explain how to answer questions so errors are avoided altogether?
- Could you use smart defaults to prevent errors?
- Are you using a discount code field that is driving users away to search? Could you hide it or make it less prominent?
- Are you making it clear which payment methods are accepted?
- Are you asking for unnecessary fields? (card type / start dates)
- Are you sending errors to users who include spaces or dashes in their card number?
- Are you telling people why you want their data (for example, explaining why you need their phone number)?
- Could you add reassuring microcopy around the privacy notice to make it clear that you will never use personal information except for delivering what they signed up for?
- Could you use social proof to increase trust on your form?
Determine Changes and Test
You’ve got your data driven hypotheses on what may be causing form abandonment. Now is the time to make the change and test whether it improves things.
- Set up an A/B test using an experimentation platform.
- Pick the one variable / change you would like to test (only test one at a time). This will most likely be the UX issue causing the most abandonment on your form.
- Be clear on the KPI you are testing - form completion, abandonment on the field, etc.
- Create one “challenger” version of the form to compare against the “control” status quo.
- Set sample size and randomisation rules - your testing tool should be able to set the correct level of traffic.
- Put the test live.
- Don’t “peek” until you have reached the required traffic levels (except to confirm that the experiment is delivering the correct sample ratios).
- Review the test results to see if there is a statistically significant difference between your control and challenger versions.
Rinse and Repeat
You’ve determined whether the changes were effective. Now is the time to either implement them or abandon them. After that you can start the process again.
- If your test variant delivered significantly better results then roll it out permanently across the form.
- If the test is significantly worse than the control then do not implement it.
If the test and control variants are not significantly different then you can:
- Make no changes (there is no evidence there will be an improvement).
- Make the change permanent (not generally recommended but sometimes there are other reasons to make the change and there is no evidence that there will be a negative impact on performance).
- Evolve the test and re-run. Change the proposed solution to fix the identified UX issue.
Run the next test. Either:
- An evolved version of the previous test.
- A fix for the next highest priority UX issue.
Download the pdf version of this checklist here