Filling in an insurance form is a time-consuming task. On average it takes a user 246 seconds and information has to be entered into 56 fields across 6 steps. The information asked for is varied, not always easily accessible or remembered by the user and it creates a high cognitive load for them.
The optimisation opportunities for long forms are significant as there are a high number of elements that can be improved or removed. New optimisation opportunities become available as technology changes or new UX learnings are created, leading to an optimisation process that is continual rather than project based.
I carried out this analysis in April and May 2017. I created a persona that was used across all the insurers', meaning that the information I entered was standardised across all the quotation processes.
In our analysis I only counted required fields so optional fields were noted, but not included in our numbers.
For insurance aggregators (e.g. Confused.com) the payment process sits on the individual insurers' website so I took the average fields on a payment step (12 fields) and added this on to as a final step.
Finally, when an average number is noted in this report it is a median average unless stated otherwise.
I included almost every major UK insurer that offers a home insurance product, including aggregators
Our research is focused on every aspect of the form, from fields used, timing and mobile usability.
I scored each form based on critical factors that will increase user frustration and decrease conversions
Insurance forms could be half the length. The difference between the shortest and the longest forms over 100%. Reducing the number of fields has a direct impact on conversion rates.
There's a move to Radio Buttons: the highest scoring insurers' use fewer text boxes and drop downs. Radio buttons give customers immediate visibility on all the options and less interaction is required to select an answer, saving time and effort.
Mobile keyboards aren't being used effectively. 40% of insurers' aren't showing a mobile friendly keyboard, with both direct insurers' and aggregators missing out on this quick-win. Showing the right keyboard for the input type will save users time and get through the process faster.
There's a significant opportunity to improve inline messaging. Less than one in five insurers' responds in-line to the user when they make a mistake or enter information correctly. A recent Formisimo survey showed that 90% of consumers want inline error messaging but 76% also want in-line validation when they enter information correctly.
Required fields aren't marked, and that costs users time. Optional fields aren't marked by 55% of insurers' leaving the user to believe that every field has to be completed. If a field is optional then either mark it as such or analyse the field performance and remove it if it offers no value.
Each step in a home insurance form is a new page or new section for a potential customer to absorb. There's no magic number of steps that will lead to a higher conversion rate but there's likely to be drop off if the process gets too long. Legal and General's 15 steps are likely to put off customers, whereas MoneySupermarket's three step process is significantly shorter but the overload of questions on each step may cause drop-offs. There's a balance to be found by testing and monitoring; if an insurer has too many steps then try combining two steps and testing the response from the market.
The difference between the highest and lowest number of fields across all the insurers I analysed is significant, with Endsleigh requiring 92 fields to be completed and MoreThan needing 39. That's a large difference in time and effort required to complete the form, for what is ultimately the same process.
Aggregators asked for more information (76 fields on average, compared to an average of 53 for direct insurers) and this is likely because they have to cover off the questions for all potential insurers. A direct insurer has the opportunity to adjust their quotation process from end-to-end, and whilst it may require work beyond the form, it means that they can reduce their form size significantly.
For direct insurers this shows an exciting optimisation opportunity. Aggressively reducing the number of fields will have a direct impact on conersion rates, quotes and revenue.
Most insurers top-load their questions, asking more at the beginning of the process than at the end. Nationwide is an extreme example, with 73% of their questions in Step 1, whereas Barclays has less than 2% of their questions in the first step. Most insurers fit in around the 25% - 35% range, with a median of just over 30% of questions in step 1.
The question over top-loading (or not) is an important one. Ask too many questions in the first step and the user will be put off, whilst having too many steps will make the user believe that they're engaging with a virtually endless process.
A sensible starting point is a median across all insurers: topload in Steps 1 and 2, then trail off. The average number of fields in Step 1 is 16, followed by 12.5 in Step 2, then 7 in Step 3. Step 4 has an average of 4 fields, then 3 in Step 5. I talk below about how Aggregators split their fields between steps, and they take a different approach. There's a tendency to look to aggregators for best practice, but the reality is that the best approach is one that has been hypothesised, tested and then made permanent.
Considering the user's mindset is an important part of optimising your forms. A user faced with a single step with over 50 questions will feel unhappy with the process, and that's why splitting insurance forms into multiple steps is vital. Toploading is a trend I referenced above, and that gives the user a sense of relief as the process becomes easier as they complete each step, but the insurance aggregators do it slightly differently.
Direct insurers have 37% of their fields in Step 1, then 22% in Step 2, but aggregators keep Step 1 light with just 6% of their fields there, and then 31% of their fields in Step 3.
Keeping Step 1 light means more people are likely to start the process. The reduced number of fields makes the user believe that the process will be fast (even though aggregators ask for an average of 76 fields compared to the average of 53 for direct insurers) and a user that has already committed to start a process is more likely to complete it.
On average a home insurance form takes 246 seconds to complete, with aggregators taking an average of 290 seconds (vs 236 for direct insurers). There is a significant difference between the best and worst performing, with Aviva taking just 170 seconds and Endsleigh taking 388 seconds.
The main factor in form completion time is the number of fields, with a general trend of a lower completion time with fewer fields. A secondary factor is the type of fields used, with text boxes as the field type that takes up the most time, followed by drop downs, then radio buttons and finally tick boxes. Radio buttons take an average of 2.8 seconds to complete (from absorbing the question, to selecting the answer) compared to 3.6 seconds for drop downs. Not only does a higher number of fields mean more time within the field, but also time between the fields, as the user completes one question and moves to the next one.
Comparing the average time per field for GoCompare (second worst performer for the total time) to the Co-op (third best performer) shows the important of choosing the right field type. GoCompare has an average time in the field of 3.13 seconds, whereas the Co-op has 3.51 seconds. Whilst GoCompare has a higher number of fields they're using field types that are easier to complete.
Another secondary factor is the number of steps in the process, which are usually separate pages. Each step takes time to load, increasing the total length of time to complete the form. An example of performance improvement is Legal and General, where 13.47% of the total time to complete the form is related to page loads, whereas just 4.26% of GoCompare's total time is related to page load.
Radio buttons trump Drop Downs for two reasons. Firstly all of the potential options are visible and readable by the user on a radio button. A drop down requires the user to click and then read the options.
Secondly the user has less total interaction with a radio button than a drop down, a drop down requires a minimum of two clicks versus the single click for a radio. With a form that has a significant number of fields then saving the user fractions of a second will add up to a good time saving, and a greater likelihood of conversion.
From our aggregated timing data on forms (Formisimo tracks engagement with fields with millisecond accuracy) I can see that the time spent on a radio button (taken from the time of the interaction with the last field to the first engagement with the next field) is 2.8 seconds, whereas a drop down is 3.6 seconds.
Comparing the top five insurers' (ranked using our scoring at the base of the page) and the bottom five insurers the higher performing sites move away from text boxes and drop-downs and use radio buttons. Radio buttons (where the user has a choice between two or more items, presented as items that are clickable) are much easier to complete, they're not ideal for when there's a lot of options but when a user has to make a choice between two to four options it allows them all to be presented to the user without them having to click.
Comparing the UX of a text box (open, requires typing), drop downs (closed list, requires clicking and reviewing) and radio buttons (closed, all options visible) the radio button wins on total interaction required. A radio button requires a single click on the right option, whereas a drop down requires a minimum of two clicks, with an unfurling of the options that will then need to be read and absorbed.
From our research on insurance forms the radio button is much easier to restyle, to make it easier to absorb and also fit in with brand guidelines.
The prevalence of mobile usage presents new user experience challenges. Beyond the way that a quotation form is presented there's a second challenge of how the user enters data, and dealing with the complexity of data entry compared to entering information on a desktop.
The use of keyboard types on mobile devices makes it easier for potential customers to complete a form as it shows them the most relevant keyboard for the data input. As an example a number or telephone number field should show just the number keypad, rather than the full keyboard. A field that asks for an email address can show a keyboard that makes the "@" symbol visible without having to press the shift key. Adding this functionality to a field is relatively easy, and it generally has a positive impact on conversion rates.
Less than 60% of insurers use the "email" keyboard type, with half using the "telephone" keyboard type. There's a slightly higher take-up of the "number" keyboard type, at 53.13%. An insurer that isn't using keyboard types has a quick, easy to setup optimisation opportunity.
Inline validation is a good way to highlight issues as a user moves through a form, and when you have a long form with complex questions it's a necessity. Showing errors whilst the user is engaging with a question is better than having the user click submit and then they're then presented with a list of all their mistakes.
The uptake on insurance forms is low, with just 51.61% of brands using inline validation. However 100% of the aggregators use inline validation, whilst just 51.85% of non-aggregators do.
Website visitors have all experienced frustrating forms and when they start to engage with yours they are carrying baggage from the times when a form has been annoying. The length of an insurance form suggests that it will be a painful process for them.
A good way to keep users motivated, and to challenge their belief that the experience will be painful, is to use positive validation as they enter data. This is often in the form of a tick box next to a field, and is shown as they exit the field and move to the next one. This positive validation keeps the user happy, and helps them believe that when they click the "submit" or "next" button that they will be successful.
Showing errors in-line is equally important, as the user doesn't have to wait until the end of the step to see if they have made mistakes. A good inline validation process will not only show the user if they have entered incorrect information in a field (and so so as they move out of the field) but it will also explain what they need to do to correct it.
AXA use full inline validation, showing the user when information they have entered is correct and also when a user makes a mistake (along with a full explanation of what the mistake is). In the image above I have left the postcode field blank, and then moved to the next field.
Confirmation of your insurance is critical as without coverage you are exposed to significant losses. Receiving confirmation that you are insured is vital for peace of mind.
Only 20% of insurers' ask you to confirm your email a second time, either immediately after you enter it the first time or later in the process. Only 10% ask for confirmation of phone number, but 100% of insurers' have postcode lookup. Looking up the address via the users postcode not only saves time but improves the quality of the data.
To score each insurer I generated a ranking out of 70, which took into account the number of steps and fields (related to the median score for each) and scored insurers' based on the percentage they were above or below this. I also scored insurers' based on the removal of unnecessary obstacles, and the mobile friendliness of each form based on the usage of mobile keyboard attributes. Having a livechat and postcode lookup was also part of the scoring process, and using inline validation boosted scores further.
The aim of our scoring was to take into account attributes of the form rather than the look and feel. Some insurance forms were more attractive than others, but as design is subjective I chose not to apply any marks for forms that I felt looked better or worse.
"If you have forms, you need to analyze their performance. Formisimo does a very good job at providing in-depth information that really helps with optimization efforts. Useful data, simple interface. It is form analytics that actually works, especially when compared to some mouse tracking tools that have form analytics as one of the many features."ConversionXL Peep Laja