How to Use KISSmetrics to Improve an E-commerce Checkout Funnel

Source http://feedproxy.google.com/~r/KISSmetrics/~3/qBd_lGhbbo8/

One of the great things about KISSmetrics is that it helps you continually optimize your business. From revenue reports that tell you which ad campaigns bring the most customers to cohort reports that track critical metrics like login retention over time to many other insight-driven reports, KISSmetrics delivers pretty awesome tools for any marketer or manager.

We couldn’t possibly cover all the benefits of KISSmetrics in one post. We would have to write a book.

So, in this post, we’ll concentrate on one key feature of KISSmetrics – the funnel report.

Specifically, we’ll focus on improving your e-commerce funnel. We’ll begin by viewing a funnel report and finding areas that need improvement. Then, we’ll form a hypothesis setting out changes that might improve conversions. We’ll run a test. And, then, we’ll close by viewing the hypothetical results.

1. View the Data

Our company, Acme Widgets, sells home goods. We’ve grown steadily over the years, but the bottom of our funnel (sales) hasn’t done very well when compared with our traffic. We want to get into the nitty-gritty to understand how each step performs and find the bottlenecks we can improve.

We’ve installed the JavaScript tracking code and created our events. Now, we need to set up our funnel. To do this, we go to the Reports section of KISSmetrics, click on Create New Report and select Funnel Report:

We’ll name our report “Checkout Funnel”:

We want to track a month’s worth of data, so we changed our date range to June 1 – 30, 2014.

We’ll start with people who have put at least one item in their cart. The name for this event is “Cart Page”:

The next step of our checkout funnel asks visitors to enter their shipping address, along with their name and email address.

This event is called “Address Page.” We click on the + sign in our funnel, add this event, and hit enter:

Once we enter the Address Page event, KISSmetrics runs the report. We see that 62% of people who placed an item in their cart went on to the Address Page. That means that a good percentage of people who place an item in their cart have the intention to buy (or at least want to see what the next step is).

The next step of our checkout asks users to enter their billing information. This event is called “Billing Page.” We add this event to the funnel:

We see that there is a drop-off of people. We’ll continue creating our funnel to see if it’s our major bottleneck.

After our billing page, there is a confirmation page. It’s a summary of the order and a few forms for ordering, as well as a gift wrapping option. We’ve named this event “Confirmation Page.” We add it to our funnel and see our results:

The last step of our checkout funnel is the “Placed Order” page. After the user confirms their order, they are taken to a webpage thanking them for their order. As the last step of our funnel, we add this event:

We see that we have about a 91% conversion rate from Confirmation Page to Placed Order. When people get this far in the funnel, few exit. The biggest drop-off in our funnel is our Billing Page. People see the page where they enter their payment information, get cold feet, and bounce.

Overall, our conversion rate from Cart Page to Placed Order is only a little over 5%. We think there is room for improvement.

We also want to see the big picture and get our overall site conversion, so we’ll set up a report for that as well. We’ll track people who visited our site, made it to cart page (triggered when a product was put in their cart), and placed their order:

Out of the 6,545 people who visited our site in June, 1.2% converted to purchasing a product. We think we can do better than this. We’ll start by focusing on improving our checkout funnel.

2. Form a Hypothesis

When conducting a funnel analysis, we need to pay special attention to those who decided not to purchase from us. Learning why people chose not to make a purchase can provide us with actionable insight for overcoming those barriers to purchasing.

We decide to run a survey to get a better understanding of these barriers. We’ll use Qualaroo and ask our visitors, “If you did not make a purchase today, can you tell us why not?” It’s an open question. There are no pre-set multiple-choice answers.

The survey displays after the visitor has been viewing a page for 20 seconds.

Many people stated that they didn’t feel comfortable purchasing. They had doubts about the security and lack of general information about the company.

The first thing we’ll need to address is the doubts around security. To make visitors feel more comfortable with purchasing on our site, we’ll be adding a few security elements throughout our website and on our cart page.

Our Hypothesis: Adding Security Elements Will Improve Checkout Conversions

Each page of the website will have the “Norton Secured” symbol in the bottom footer. This will provide visitors with reassurance that their data is safe because the website is protected by a known internet security company.

We’ve also recently become a Google Trusted Store and will be embedding that symbol in our bottom footer.

We’ll have both symbols on our cart page, along with a message stating that we’ve shipped over 5,000 orders since our founding just 3 years ago. Also, the following message will appear on the cart page:

“Our website uses a 256-bit AES encryption to protect all your information. Your credit card will be processed through PayPal, which has handled well over $1 billion in transactions. Click here to learn more about our security practices.”

The link takes visitors to a newly created webpage that outlines the security measures we’ve taken on our website.

We believe adding these elements will provide reassurances to customers that our site is secure and our business is reputable.

3. Run the Test and View the Results

Our test will go live for every person who visits our site. We’ll let the test run for a month, and afterward we’ll see if our hypothesis proves correct.

When testing changes to your website, you can either eyeball your data or run an A/B test.

When you eyeball your data, you make changes to your website, all visitors to your site get the change, and you view your data to see if the change increases conversions.

There are a few cases where you want to eyeball your data:

  • If you’re a young company that doesn’t have a lot of traffic. In this case, it’ll take too much time to run an A/B test. Keep your focus at the top of the funnel and earning more traffic before you A/B test. Until then, eyeball top-of-the-funnel conversions, such as sign-up rate.
  • If you’re a high price / low volume sales company. If you get only a few sales annually, keep to eyeballing your data.
  • If you are, as in our case, responding to overwhelming feedback from visitors. For us, they’ve made it clear that they do not trust making a purchase from us.

In some cases, you’ll want to set up an A/B test. The following situations generally warrant an A/B test:

  • If you have enough traffic to run an A/B test. More traffic will allow you to get much more reliable results.
  • If you’re making small changes that may not result in big conversion changes.
  • If you’re expecting less than 30% improvement. Small changes need an A/B test in order to see an impact. Otherwise, the small improvement just gets lost in the noise of the data’s natural fluctuations.
  • If you want to find wins more quickly. You can detect smaller wins.

If you fit the A/B testing criteria, you’ll want to sign up for a service like Optimizely or Visual Website Optimizer. KISSmetrics has an A/B test report that integrates with these services, and you’ll be able to view those results in KISSmetrics. The benefit of using the A/B test report is that you can view actual people who did or didn’t convert in an experiment.

Going with our eyeball test, we come back one month later, and we view our funnel report. We select July 2014 for our date range. Our report loads, and here’s how it looks:

Here are a few initial impressions:

  • The number of people who placed items in their cart was lower than in June. This is because of less traffic to our site, which we’ll soon see in our main acquisition funnel. In the future, we may want to add the “Visited Site” event to our checkout funnel. This will show us how our checkout funnel performs relative to our traffic.
  • Every step of our checkout funnel had an improvement in the percentage of conversions compared with last month, especially in the last 2 steps.
  • Despite lower traffic, we still had many more customers place orders with us.
  • Our conversion rate quadrupled. About one in every five people who placed an item in their cart completed their purchase.

This is awesome news. It appears that our hypothesis was correct and the changes we made to our site instilled confidence in visitors, who were then more likely to buy.

Now, we’ll run the report for our overall site conversion. Remember, in this one we’re tracking total people to our site, then people who put an item in their cart, and then those who proceed with their purchase.

At the top of our funnel, we can see that we had less traffic in July than in June. But, a larger percentage of the visitors put items in their cart and placed orders. We got more customers on less traffic! Throughout our funnel, larger percentages of people converted to the next event step, especially in the last 2 steps.

We have the essence of our main conversion funnel. It gives us a bird’s eye view of how well our website converts. If we want to further understand what’s driving it, we can segment these people in this funnel by properties. Properties are various characteristics about the people.

A common example of a property is “returning,” which tracks whether the person is a new visitor or returning. Also, if someone came to your site via a search engine, the property “search engine” will display which search engine the person came from.

Segmenting the Data

For this funnel report, we’ll segment our visitors by the property “channel.” It categorizes traffic into referrer segments. Here’s how it looks:

For more information about the channel property, check out our article on Channel Definitions.

This report shows us how each channel performs, not just in bringing visitors, but throughout the funnel.

We see that one of our strongest, top-of-the-funnel channels is organic. A lot of people find us by search, and a fair number of them convert to place an item in the cart and then purchase.

A high-converting channel is Referral. These include sites that link to us. Since it’s successful for us, the more links we can get, the better.

We see that our paid efforts bring us visitors and some of them place an item in their cart, but none of them place an order. We may want to consider some remarketing efforts for the people who abandon their cart. If we don’t want to add to our costs, we may want to cut off paid acquisition. But before we do, we’ll have to view our cohort report to see if some of these visitors make purchases later than those from other channels. We’ll get into the cohort report a little later.

Email appears to be a strong channel for us. We may want to promote our email newsletter more. Here are some ideas for how to do that:

  • Offer 5% off for new visitors who sign up for our email newsletter
  • Place a modal on our homepage offering our newsletter
  • Once customers place an order, have a message suggesting that they may want to sign up for the newsletter

Throughout all messages, we need to explain the value of our newsletter, as well as how often subscribers are emailed. We could mention the tips we provide, exclusive deals, and new product arrivals.

Cohort Report

Up until now, we’ve been using the funnel report to tell us how our website converts for a specific date range. Moving forward, we’ll want to set up a cohort report. A cohort is a group of people who share a common characteristic or experience within a particular time span.

Here’s how a cohort report looks in KISSmetrics:

This report ties together two events: Signed up and Logged in. We first track people who signed up, and then we see the rate at which those people logged back in for the first time over a span of 12 weeks. This report keeps people separated, based on the time they signed up. This allows us to track how each group’s behavior changes over time.

We can set up this report for any event. For our example of tracking conversion rates over time, we would want to tie the Visited Site event with the Placed Order event. Here’s how we would set that up:

The first step in creating a cohort report is to select the two events you want to track. We want to track the people who came to our site and ended up placing an order, so we are using Visited site and Placed Order as our events.

Under Advanced options, we can split people into groups by minute, hour, day, week, or month. We’ll set up weekly buckets. We’ll track people by when they visited the site, and we’ll group people by week.

We run our report and get our data. It appears that the majority of visitors place their orders within 2 weeks. After 2 weeks, there is a low chance visitors will place an order. As time goes by and we get more data, we’ll split people into monthly buckets.

Most cohorts are broken down by time (month, week, day, etc.) In a KISSmetrics cohort report, you also have the ability to segment your cohort by any KISSmetrics property. In our case, we’ll want to view people by channel. This will show us if any of our previously unconverted paid visitors ended up placing an order. To set this up, we track the same events, but this time group people by channel property:

And we get our data:

Grouping people by channel is useful for tracking how these channels perform at converting visitors over an extended period of time. We aren’t looking at a simple funnel report that sticks with a date range. We’re viewing our visitor behavior and how well our site performs at converting visitors.

In our example cohort report, we see that nearly every channel, with the exception of paid, brings us the bulk of our customers within 2 weeks. Our paid customers are more hesitant to purchase within the first week, deciding instead to hold off and complete their purchase at least 1 week after visiting our site. The number of paid customers is still small compared with other channels, so we’ll have to check our marketing expense to see if it’s profitable for us to keep our paid acquisition running.

As time goes on, we’ll continue to rely on these cohort reports to track how our conversion rates change over time and how each channel performs at converting visitors into customers.

Before we go any further or conduct any more tests, we first need to move on to the next and final step of our test process.

4. Repeat

After we’ve run the test and gotten our results, we’ll move forward with our next test. We can pat ourselves on the back for this test knowing that we found the main bottlenecks in our funnel and made changes that positively impacted our conversion rates. Going forward, we’ll want to sign up for a service like Optimizely or Visual Website Optimizer to run our A/B tests.

Finding the Variables That Impact Your Audience

You probably have hundreds of ideas for tests you can run on your website. The bad news is that most tests you run won’t move the needle much. In light of that, focus on running tests to identify the variables that do impact your audience.

To get started using people data, login or sign up for a KISSmetrics account now.

Some of the links in the post above are “affiliate links.” This means if you click on the link and purchase the item, we will receive an affiliate commission. Regardless, we only recommend products or services we believe will add value to our readers.

Share and Enjoy

  • Facebook
  • Twitter
  • Delicious
  • LinkedIn
  • StumbleUpon
  • Add to favorites
  • Email
  • RSS

Leave a Reply

Your email address will not be published. Required fields are marked *

CommentLuv badge