Return to Blog

[Case Study] Small changes lead to a 55% increase in conversions (on our own site)

Increasing conversion rates of Formisimo’s sign up page by 55%

We like to try and live by the principles that we encourage in our users. We’re constantly trying to test and improve our site to maximise conversions, so I’d like to share with you one example where we utilised our own tool to increase the completion rate of our sign up page.

The original

This was our original sign up form:


Relatively short, but there is always room to improve!

We noticed that over a two month period, our password repeat field (upassrpt below) was responsible for over a quarter of all the people that abandoned our sign up process:


Although not the most corrected field in our form, there were hundreds of corrections made by our visitors in Confirm Password, which were re-focuses (clicking back into the field after exiting it), plus deletes (deleting the whole password to re-enter it no doubt):


After reading a fantastic article by Jessica Enders at Formulate on double entry of information, we thought we’d try out a variation. We’d added in the second password field more out of  common practice than a deliberate decision (i.e. lots of our users were incorrectly entering their password and having to reset it later, which was not the case). So we decided to drop the second field.

Luke W has also written about the usability of password fields , so we decided to try out adding the ability to toggle making the password visible.

We went on to make some additional changes to the form:

  • Removing the number in the top left, as it implied there were multiple sign up steps, which there were not. The number ‘2’ simply displayed on the confirmation page
  • Changing the text at the top and replacing it with the the more powerful ‘Start tracking your forms’.
  • We also changed the colour of the call to action button to a brighter orange.

The new version

After the changes we had the following:


There were no other major changes the sign up page in terms of imagery and text, which was pretty minimal to start with.

In the two months leading up to the changes, our conversion rate of visits to the sign up page to completions was 15.8%. We also like to include to conversion rate of those that start the form (as an indicator of how easy it is to complete), which was 45%. Here’s a screenshot taken from Formisimo’s overview report for the previous sign up page for a two month period:


Some other important stats taken from Formisimo for this period:

  • It took users a median time of 43 seconds to complete the form
  • The median number of corrections per form starter was 10.21
  • The password request rate per new user was 10%

Here’s the overview graph from the updated sign up page for the subsequent two months:


For the new version:

Our overall conversion rate was 24.7% for the two month period. The form’s completion rate was now 61%.

  • The median number of corrections per start dropped to 7.76
  • The median time to complete remained at 43 seconds
  • The password reset request rate remained at 10%

The results

So, as a result of the changes we made, we experienced:

  • 14.3% more visitors starting the form after visiting the page
  • A 56.3% increase in overall conversions
  • A 35.5% increase in the proportion of those who start the form completing it
  • A 23.9% decrease in the  number of corrections made in the form

These changes also did not mean our users had to spend a significant amount of extra time in the form, or caused any increase in the password reset request rate.

It was great progress, but we’re not resting on our laurels. We’re split testing another version of the sign up page, so I will be sure to keep you posted on the results.

What we do know though, is that the insights that Formisimo gave us provided us with a starting point to improve our form, and then to measure the result with more depth and precision than we could have otherwise. We believe that the world needs better forms, and we want to provide the data to help people get there.


Keep reading

Digital Marketing and CRO in Political Campaigns

How do political campaigns use CRO and digital effectively? I take a look at leading campaigns and how parties are optimising their marketing and improving performance.

Read more

  • Paul Gregory

    Great article, and good to see some real statistics, but it’s not surprising that the captcha is the most ‘corrected’ field, as any other field needing correction would usually mean that a new captcha must be completed.

    Is ‘captcha is only corrected field’ usually higher than ‘captcha + one other corrected field’?

  • Pingback: 入力フォーム改善でCV率16倍!成果を上げる18のテクニック()

  • cfc

    Thanks for sharing this, it definitely provides some interesting food for thought in form UX design. However, I believe it overlooks a fairly major issue: Registration conversions do not directly correlate with new active users.

    Password reset numbers are quoted as if they accurately represent the number of users who may have mis-entered their passwords. I can’t help wondering: of the additional users who signed up due to the removal of one field, how many of those who mis-entered their passwords simply gave up when they next tried to login, rather than initiating a password reset? Users who are unwilling to fill out one extra form field seem unlikely to be willing to go through the entire email confirmation rigamarole of a typical password reset process.

    A much more valuable metric than registration conversions would be number of regularly active users. Are you able to provide any information on the growth rate of number of users regularly active on the site before and after the new form? It would be interesting to see whether that matches up with the conversion increase you saw.

    • Formisimo

      Hi and thanks for your detailed comment; definitely worth addressing the points you make.

      I think what you’re saying is although the number of users who sign up may increase, the number of active users may not, as those that entered their password incorrectly would be unlikely to be those potential active users that would a) gladly type in their password twice, and b) go through a password reset if needs be.

      I don’t currently have that data to hand but I can certainly explain our reasoning. We would rather people get a taste of our (hopefully wonderful) product before being put off (at the sign up stage). By their second login, they will have already explored the interface, and possibly started the set-up process, so the incentive to reset their password is higher than those who fall down before they see the product at all.

      Another point to make is that often sign up forms have to overcome the perception of friction almost as much as the actual effort required to complete it. Social logins buttons for example often take as long as speedily typing in an email and password, but the perception is that it’s a smoother process.

      We would, on the whole, always advocate the view that removing friction is a good thing and we shouldn’t try to let in only those willing to try hard enough to sign up!

  • jim birch

    Why no “login with f/g/t…” option? Personally, having to create and manage yet another password always makes me question the value of any service.

    • Formisimo

      That’s an interesting perspective. We didn’t add social login as this is very much a business product. Although it would be up to each account holder if they wanted to link their social profile with the tool, it’s not a request we’ve had before. Additionally, accounts are often accessed by different people in a team and so would not suit being linked to an individual’s social account.