Add Funnel Steps to Improve Conversion

In my last post I discussed how an understanding of consumer psychology can be extremely helpful for understanding how to design funnels that convert. In this post I’m going to give a concrete example of how I did this at Lookcraft to drive 18% conversion through a funnel with 18 discrete questions and 4 major steps. I got 18% of users who hit our homepage to enter their email and password, information about their fashion taste, and full sizing information. I’ve used these same tactics across multiple different products to similar effect with other companies I’ve worked with or consulted for and hope they will be useful for you as well.


Discuss this post on Hacker News.

About Lookcraft

Lookcraft made it easier for guys to buy clothes by giving them a style and fit quiz, recommending them clothes to buy, and shipping them those clothes for free to try at home. Guys would pay for only the clothes they decided to keep, with free shipping both ways. We achieved a high (around $400) customer lifetime value (LTV) in a difficult market. In spite of high LTV (which did factor in variable costs) the business was quite capital inefficient which created financing issues leading us to eventually shut it down to focus on other things.

Step One – Email Address

The goals of your landing page should be to provide exactly as much information as needed to clearly communicate your value proposition, overcome any user anxiety about your ability to provide it, and minimize friction for the user to provide you with their information.

Many companies feel the need to fully explain all the details of their product on the landing page. Its not uncommon to see pages fully loaded with user testimonials, press hits, demo videos, feature descriptions, links to a FAQ, an about page, and more. This is counter-productive. When you load up your page with every little detail about your product you’re actually giving users more reasons to say no, and you haven’t gotten any information about them yet. Some mystery is good, its your hook. If you have a users email you can – and should – educate them more about your product over email if they do not convert.

Here is the page we ran:

9zar

This converted at around 35% to email addresses. This ranged from upwards of 60% for very qualified traffic (e.g. direct) to as low as 20% for less qualified traffic. There are a couple key takeaways here.

  • We made the page very simple, it gives the user enough information to make a decision to type in their email or not, and no more.
  • We had the page only asks for email address. Choosing a password is high friction and at this point the user is not very engaged, asking for it here would cause further drop-off (I have A/B tested this extensively in the past and shown it to be true multiple times).

Step Two – Style Quiz

Quizzes are one of the highest converting process flows on the internet. They appeal to users’ desire to learn things about themselves. At Lookcraft, we needed to get data on two important things in the onboarding process, users’ clothing size information and users’ fashion taste. We deliberately structured this with fashion taste first because we wanted to hook the user with a quiz that purported to tell them something about themselves – their style – while also being low friction to fill out. There were two steps to the style quiz we ended up running. The first page asked users to select the images they liked from a set of 3 images 3 times, hitting next after each step. Users then saw a page where they selected all the images they liked from a set of images.

Here are the pages we ran:

a0~p

ftul

This converted at around 80% to completion. Some key takeaways here:

  • We told the user we were going to teach them something about themselves.
  • We started with something low friction. Its really easy to answer questions about what you like and don’t like.
  • We showed the user their progress so they would understand where they were in the process.

Step Three – Fit Quiz

We needed to collect information about how clothes fit the user so we could later show them items that were in their size and also likely to fit them well. We assumed in the beginning that this would be the most difficult step as there is no easy way to collect this info without just asking the user a lot of multiple choice questions. We intentionally put this step last so we could engage the user with easy steps to first get buy-in. Users who got to this point had already given us their email address and spent a few minutes with the product.

When constructing the funnel this way we took into account that people are prone to loss aversion. This is a concept that applies to time costs as well as actual costs. By getting users to spend time with our product in simple, low-friction ways we built up a time commitment. As we get further down the funnel we can ask users to do harder things, and they are more likely to do them to avoid sunk cost of time. This step had 12 multiple choice or drop down questions which included: height, weight, body type, shirt size, shirt fit preference, preferred jean brands, jean size, jean inseam, jean fit preference, preferred jean brands, shoe size, and preferred shoe brands.

Here is the page we ran, with more questions continuing on the same page below the fold:

j9ng

This converted at around 77% to completion. Some key takeaways here:

  • We built up time commitment to the product before asking harder questions, taking into account human loss aversion tendencies.
  • We let users know they were almost done with the process to remove anxiety about future time commitment after an already difficult step.

Step Four – Style Results

At this step of the funnel we were not yet done collecting all the information we needed about a user, but we felt like we had asked them many questions by this point and had messaged to them that after the previous step they would be done answering questions, so we wanted to reward them for the time invested in a quiz that purported to tell them something about themselves. We did this by building a style results page that gave them a ‘payoff’ on their effort so far. It was deliberately simple with only one call to action at both the top and the bottom of the page to go to the next step in the process.

Here is the page we ran:

A55590RCAAA2ZpI

This page converted at 98% to the following ‘View Recommendations’ step. Some key takeaways here:

  • After asking a lot of questions we gave the user a bit of something back, we delivered on our promise to tell them more about their style.
  • We made this page have one very clear call to action, there is nothing else to do on the page but go to the next step.

Step Four – Name & Password

We did not collect the users’ name or password in the first step because we intuited that those questions would be too high friction for most users to answer at that point. However, we still needed this information in order to complete the users’ account. We decided to put this step after we had asked users the hard questions and given them some payback (telling them their style), but before they got to see the full payback (product recommendations). We did not want to make this its own screen because we felt like asking for two pieces of information on its own page would look awkward. We also felt like we needed more of a hook in order to get the user to enter in their name and create a password. We had just given them some value back, but they needed to see what else they were getting to get past this step.

This is the page we ran:

54p2

This page converted at 83% to the following ‘View Recommendations’ step. Some key takeaways here:

  • We put the hardest step at the very end after users had already committed a lot of time to enter in earlier information, and had gotten value back by learning their quiz results.
  • We entice the user by letting them see what they are about to have access to if they just get past one last simple step.

Results

Here you can see the numbers straight out of Mixpanel for the above funnel.

Screenshot 2013-11-16 21.41.51

Screenshot 2013-11-16 21.35.52

Multi-step funnels can be very helpful for driving conversions because they leverage people’s propensity to fall prey to the sunk cost fallacy of time investment. Loss aversion is a strong and versatile psychological tool that is very useful when designing product. It is the driving force behind funnels that convert better when you add more steps to them. Try it out for yourself!

Did you like this post? Share now!
Email this to someoneTweet about this on TwitterShare on FacebookShare on LinkedInBuffer this page

I write 1-2 times a month about Product Marketing and Growth. Sign up now to be notified of new posts!