• #357 on Amazon’s Bestsellers
• #1 in Computer Science
• #1 in Website Analytics
• #1 in Web 2.0
• … (the list continues.)

Obviously the book has gained tremendous ground. It has a host of 5 star reviews and 10/10 from many, many sources. I was even on a plane last week and saw 2 (!) people reading it. I’ve seen that happen with Seth Godin or Malcom Gladwell but that’s even surprising for Avinash!

So Web Analytics 2.0 is hot. I don’t need to write a review to validate that. I have however been surprised by the lack of adoption amongst some of my advanced peers, so below are three good reasons for even the most knowledgeable analytics guru to give it a read.

Learn How to Choose an Analytics Provider

The popular thing to say when it comes to choosing an analytics provider is either “just use Google analytics” (<25, small to mid market) or “just use Omniture” (manager or above, works for IR 100) but there is a lot more to it than that. Web Analytics 2.0 dives thoughtfully into solving the “who” question and provides a solid model for moving forward.

Anyone who has ever been through an analytics RFP will particularly enjoy the questions to ask sales folks…

See Multitouch/Attribution Explained Well

I would say that I’ve probably seen a hundred presentations or talks on multitouch campaign / attribution analysis. Of those, I’ve seen a handful of good ones. Of those, I’ve seen one or two that you didn’t need years of experience or an advanced degree or brilliant mind to truly get.

The book has one of the best multitouch overviews that I have ever seen.

Analysts: you need to know this.

Consultants: you need to stay ahead of the analysts.

Managers: you need to be able to explain this to executive management.

Executives: you can’t afford to get this wrong.

Gain Analytics Budget

I currently work with around 80 of the biggest brands online, and out of those, only a few appear to be adequately funding their analytics teams. Of those, I’m sure that the managers on those teams may feel otherwise. Mark Twain said “it is wiser to find out than to suppose.” Isn’t business intelligence amongst a company’s greatest assets? In a perfect would, analytics efforts would be infinitely funded.

In only three or four pages, Avinash manages to lay out a great set of strategies for getting the funds you need to provide world class analytics. In other parts of the book, strategies for finding the right people, and setting up a strong team. There is an entire section on the analytics career.

I could come up with way more than three reasons. This is by far the best over book I have read on analytics and I highly recommend picking up a copy. Let me know what you think!

There’s a pretty telling quote at the end, “if you run an e-commerce site, don’t do it like this.”

The checkout in question is that of Navicat (they also has some canonical and meta issues on the SEO side, but I’ll leave that for others.) Navicat is actually a pretty awesome product for database administration, so it is surprising to see them put up such a disjointed ordering experience on their site.

Checkout Redesign Tips

They are probably missing out on a big chunk of revenue, and we hate to see that happen, so here are some tips for making it back:

1. It starts with the click. The buy now button from the software should go to the appropriate buy now page for that version, the user shouldn’t have to navigate through three levels to find it.
2. I was actually surprised to see Google Analytics on the site. If fallout reports aren’t set up, set them up!
3. Simplify the options for the product selection.
4. Simplify the options for checking out.
5. Navicat is constrained by the payment processor in this case.  Upgrade!

Just a little something to get the day started… let me know what you think!

Before getting to things, I am not claiming the following:

• That coupon codes decrease revenue
• That retailers should remove coupon code support from checkout
• That retailers should stop using coupon codes

In fact, the retailer that I reference below recently had their best week ever through offsite marketing channels thanks to a promotional effort that relied on coupon codes. Further, they actually saw their margin increase while discounting products.

Now with that out of the way…

One Problem with Coupon Codes

I don’t have millions of dollars of research to support this, but I hypothesize that one problem with coupon codes is simply that they can and do distract users from completing their transactions in a linear fashion. A coupon code entry field exposed during e-commerce checkout is analogous to a sign at the register in a brick and mortar store that directs a customer that if they leave and look around a bit, they may be able to save money. At that point, the store is basically guiding the customer off course and then hoping that they finish what they started. Not ideal.

Online, this problem can be compounded for many reasons. My top 3:

1. It’s very easy for buyers to get distracted.
2. Smart competitors are good at stealing customers during checkout.
3. Smart affiliates are incredible at stealing customers during checkout.

So? All three of these scenarios results in lost revenue. The second scenario results in lost revenue and gained revenue for a competitor, which is an even worse result.

(I have been dinged a few times for treating affiliates unfairly. I once was a very successful affiliate. The majority of very successful affiliates are crafty folks that find gaps in a company’s online marketing campaigns and fill those gaps. That is an incredibly valuable service, but if  as a retailer you aren’t policing your affiliates and learning from them, then you are unnecessarily allowing your cost of sale to increase. It’s that easy. E-mail me for more thoughts on that.)

Coupon Codes Are More Prolific than Ever

This all matters because shoppers are not only using coupon codes more, they are searching for coupon codes more. This provides competitors and affiliates the opportunity to reroute them and either drive revenue loss or increased cost of sale.

Coupon/Promo/Discount Code Search Volume

For  a real life example – the retailer I eluded to earlier. This retailer is one of the largest online and has multiple stores that are all experiencing the same thing. Searches like “[brand] + coupon code” are on the rise in a big way. These searches are already up 40% year-over-year and that number is expected to take off in November.

We know how coupon codes can drive revenue loss, and why we care more now than ever. What are we going to do about it?

Stop Losing Revenue to Coupon Codes

The main challenge is that the coupon code field can drive users away from your site. To solve that challenge, you have to break it down to the why’s. This list isn’t exhaustive… please share ideas in the comments section and I will add to the section below.

1. Everyone wants to pay the lowest price.
2. You (the retailer) aren’t providing coupon codes when relevant.
3. You aren’t actively routing limbo customers back to your site effectively.

And here is where it gets fun. Ideas for reducing the revenue lost from coupon codes. (Note that if you try these ideas, you need to be responsible and test them!)

1. Hide the coupon code field if you don’t have active coupons or if a cart does not qualify for a promotion.
2. Automatically provide relevant coupon codes to customers.
3. Promote your promo codes. (GetElastic via MarketingExperiments)
4. Optimize a page on your site (SEO) for “[your brand] + coupon code” and variants.
5. Increase coverage on “[your brand] + coupon code” keywords and variants via PPC. Make sure that you give users a great landing experience!
6. Restrict affiliates/partners from bidding on “[your brand] + coupon code” keywords and variants.
7. Provide an explanation of coupon codes via a tool tip or help box.

If you have any more ideas or experiences here, please do share! There were a few thoughtful posts/articles that inspired me to share my thoughts here. Thanks to:

• Gaby Diaz at Marketing Experiments
• Linda Bustros at Get Elastic
• Why Users Abandon their Carts at eMarketer
• Search Results like This
• How Much is Your Coupon Code Box Costing You?

The Challenge

Most challenges you will face when optimizing dynamic checkouts are rooted in data. Bad data can cause you to misinterpret, misreport, and miss what’s really going on. As an example, when taking a high level look at your checkout conversion rate, you’ll probably create a report that measures drop off in each step of your checkout. Chances are that there is a problem already, because most people ignore the fact that not all users move through a purchase path the same way.

How would users move through a checkout path differently? Any time a user has the chance to skip a step or a page, there is a potential for anlaytics data to misrepresent that user’s actions. The most common cause of this would be a quick checkout feature, but also could be due to purchasing different product types, gift options, alternate payment methods… etc.

On most sites, returning and new users experience different paths through the same checkout.

Reporting by Segments

The solution to the challenge of reporting on dynamic checkouts is to have multiple stats. Your site conversion rate is an important number, but probably not as important as the conversion rates of your top marketing channels. When analyzing what’s really happening, you dig into the the numbers for search, feeds, etc. Doing the same with your checkout conversion rate will allow you to diagnose what’s really going on and where to focus your efforts.

Getting Started

1. Define and map out all paths through or branches in your checkout process.
2. Focus on the paths that are most used, but don’t ignore the ones that aren’t.
3. Setup reports via your analytics platform to track each path separately.

Now, when asked about the checkout conversion or shopping cart abandonment rate, you’ll be able to communicate where the opportunity really is.

I’m a big fan of the folks and the content at Marketing Experiments. In the past they have elduded to their own test protocol, and they actually go over it in one of their latest webinars. It is definitely worth reviewing. Here’s a rough overview:

• What question is being asked?
• What are the variables?
• What are the metrics?
• How are the results being validated?
• What were the results?
• What is the interpretation?
• What is being tested next?

I can’t speak highly enough of documentation around testing. Test protocol is one of a few methodologies to drive the documentation that will in turn drive long lasting incremental results and a successful optimization campaign.

[Average Order Value] * [Abandoned Carts]

This of course can be broken down further. For instance, this is the recommended formula for calculating abandoned revenue with Omniture’s calculated metrics:

([Revenue]/[Orders]) * ([Cart Adds] – [Orders])

Setting up a Calculated Metric in Omniture

To view this data in Omniture, create a calculated metric by clicking “Manage Calculated Metrics…”, and then clicking “Define a New Metric.” Then paste the above formula in the text box that appears. When you save the new metric, you’ll be able to select it and view the results instantly, which is a cool feature.

If you’re a Google Analytics user, calculate abandoned revenue using your AOV and the data from your Abandoned Funnels report in the goals section.

(If you know of a slick way to accomplish this in Coremetrics or any other top tier analytics providers, please let me know via a comment!)

On the Numbers

When you look at these numbers or – more importantly – make any decisions based off of them, it’s important to remember a few things.

1. This metric is based on a very important average; average order value.
2. You’ll want to capture ALL abandoned revenue, but it’s not realistic.
3. If you don’t set a realistic target, you will stumble.

Startled by your abandoned revenue? Let me add some perspective. Abandoned revenue is often 5-10x (or more) daily revenue. I have multiple clients with millions of dollars of abandoned revenue per day. Don’t freak out… and don’t go tell a VP (yet).

One of my points above is that it is not realistic to capture all of your site’s abandoned revenue. I am all about expectations. I want to be honest and the truth is that it’s impossible to eliminate abandoned revenue and abandoned carts.

Of course, we would not have jobs if it was not possible to capture some of that revenue.

[Header image source.]

The Overview

There are two different ways to look at shopping cart abandon rates (or the reverse, cart conversion rates,) and it is important to make a distinction and always measure apples to apples. At different points in a campaign, one methodology will fit better than the other. Note that I’ve completely made up the names for these methodologies below… but maybe they’ll catch on and I’ll be famous.

Pure cart methodology. In “pure cart,” abandonment is measured as the % of shopping carts initiated that do not result in an order. A cart is typically initiated when a user adds a product to the cart.

Pro’s:

1. Gives an optimistic view of opportunity
2. Allows you to track interaction with the “checkout now” button
3. Can expose more users to testing, driving more revenue

Con’s:

1. Doesn’t necessarily represent actual revenue opportunity
2. Some users are just shoppers, not buyers
3. Can pull focus from true detractors of conversion

True checkout methodology. In “true checkout,” abandonment is measured as the % of users that click through the shopping cart (via a checkout now button, place order link, etc.) and that do not complete an order.

Pro’s:

1. Gives a conservative view of revenue opportunity
   
2. Forces attention on forms and high value elements
   
3. Drives an understanding of users that want become buyers

Con’s:

1. Can cause shopping cart usability issues to be ignored
2. Discounts potential revenue opportunity higher in the funnel
3. Ignores all actions before the checkout click

I mentioned that each methodology is more or less appropriate given the situation, and this takes a little bit of feel. Look at the your objective, and apply both methodologies. Document what you gain and lose from each one, and choose the best option. Or, if you can manage it, measure both pure cart and true checkout abandonment.

Measuring Shopping Cart Abandonment Rate

The major analytics providers support this metric to varying degrees; from out of the box to requiring significant configuration changes. The best thing to do is to consult the documentation or your account manager to ensure that you are properly set up to measure abandonment. See “Appendix A” for further notes on specific analytics providers.

Regardless of the analytics platform used, you must take care to ensure you are set up for success. In the analytics world, this means make sure that you are tracking user activity properly and accurately.

Abandonment tracking tips:

1. Ensure each step in checkout is tagged and tracked separately through a unique page or event name.
2. QA your data before beginning a project. If you setup your funnel and see more visitors in Step 2 than Step 1, you have a problem.
3. Ideally you should track every state change, including error handling, and be able to report back on changes on even the most granular of activities.
4. Any time there is a change to any element in your checkout, re-confirm that your tracking is accurate.

After you confirm that you are setup to track users and report back on their activity accurately, create a baseline for shopping cart abandonment. Throw away any historical numbers your business has. Educate your peers and any relevant stakeholders on the current running abandonment rate and make sure that once the metric is tracked correctly, it stays correct.

As a final note on tracking: be smart when comparing cart abandonment rate over time frames. This is a dynamic metric and subject to many outside influences. Seasonality, marketing programs, sales, promotions, and economic conditions can all have drastic effects on the metric. When testing and optimizing your shopping cart to reduce abandonment, ensure that you are running inline A/B tests and comparing apples to apples.

Reduce Shopping Cart Abandonment Rate

Before we jump into tips, ideas, and concepts for reducing abandonment rate, I want to be very clear; there are no silver bullets. Best practices are a great starting point but you should always hypothesize, test out, prove, and document optimizations to the shopping cart if it is an option. This is especially important in large organizations, where often you have to start small to prove your case. When you feel confident you are ready to get started, brainstorm with the questions below.

• Do we have too many steps?
• Are we being clear about costs like shipping and tax?
• Do users have good reason to feel secure?
• Are we asking only for relevant information?
• Do our forms follow good usability practices?
• Are we communicating to our users with step-guides and helpful tips?
• Do we display supporters to conversion, like guarantees?
• Are we focusing user attention where it needs to be?
• When an error occurs, is it clear where it is and how to fix it?
• Are coupons positively or negatively affecting our business?
• Can we follow up with users to drive more conversions?
• If a user exists to browse, is the path back to the cart clear?

As you dig deeper into your own rates and performance, you will be able to expose the bottlenecks specific to your business. Brainstorming on these bottlenecks will expose more questions and lead to more actionable insights. Just work the process! Analyze, hypothesize, test, confirm, repeat.

Hopefully this guide has given you a good overview of cart abandonment, how to track it, and you have found value in the tips and suggestions presented. If you have any questions or further thoughts on cart abandon rates, please leave a comment and I’ll be happy to update the guide above.

[Header image source.]