jondyer.net

If your cart abandonment rate is, say, 58% – where do you start to fix the problem?

Many marketers would jump in with optimizing the checkout process by reducing steps, slicing and dicing form fields, changing button colors and adding security badges.

But what if most checkouts were not abandoned because of anything “confusing” or broken in the checkout process?

According to Forrester Research, only 11% of consumers report their last abandoned cart was due to a long or confusing checkout. Only 12% believed the site was asking too much information, only 14% were unwilling to register with a site.*

The most common reasons customers bailed boiled down to “sticker shock” (due to high shipping charges, taxes or other fees, or a high product price), the desire to comparison shop, and simply not being ready to check out at that moment. (Good ol’ FUDs).

No matter how pretty your cart button, how short your checkout process or how clear and usable your form fields, you can’t save these sales. But that doesn’t mean you can’t optimize your site for non-usability factors of shopping cart abandonment.

Let’s take a look at 5 top reasons why customers abandon, and how you can address each Fear, Uncertainty and Doubt.

Shipping costs too high

1. Make free shipping or shipping discount applications very prominent

• Pre-checkout shipping estimate tool helps the customer judge the shipping charge

I was not ready to purchase the product

1. Create urgency

• Include value proposition of “owning it today” (this will depend on your product and the purchase context)
• Highlight any limited time offers/discounts – don’t bury them below the fold or in graphic elements that may suffer from “banner blindness”
• Highlight financing options like $X/month (when applicable)
• Show stock availability / mark products with high sellout risk

2. Allow customers to log in to save their carts

• 30% clear cookies daily, a saved cart safeguards against wipeout
• Enables registered customers to retrieve their carts across devices
• Abandoned carts can be remarketed to (triggered email program) when you have an email address

3. Persistent cart

• To support those who will not choose to save the cart

I wanted to compare prices on other sites

1. Reinforce your unique value proposition

• Value props are not just for home pages and landing pages – the cart is perhaps the most important place to convince the customer to buy from you and nobody else

2. Suppress your coupon box

• The presence of a coupon box may send your customer huntin’ for deals from your affiliates, costing your margin and commissions. This is a form of comparison shopping – where can one find the best deal? Handle this issue by hiding the coupon box for customers who have not been referred by an email or affiliate with a discount (using cookies), or auto-apply any discounts these shoppers may receive

Product price was higher than I was willing to pay

1. Financing, if you offer it, could make the price more digestible

2. Boldly highlight (in red, green or orange) any auto-applied savings or sale prices

Just wanted to save products in my cart for later consideration

1. Use a persistent cart and enable saved carts, save-to-wishlist

• Track % of “save-to-wishlist” actions and deduct from your cart abandonment rate

While these tips won’t take care of your cart abandonment completely, addressing these issues before getting into design and usability will get you further than UX alone. Stay tuned, next post we’ll look at the design/usability things that make a shopping cart page effective.

* Forrester Research has also reported that 23% of customers would abandon carts when asked to register. Because many sites do offer guest checkout, 14% abandonment due to required registration does not mean that customers care less about site registration, rather, less customers encountered them.

Looking for help with A/B and multivariate testing? Contact the Elastic Path consulting team at consulting@elasticpath.com to learn how our conversion optimization services can improve your business results.

Just a quick reminder that our Google Analytics Seminars for Success will be in Boston, May 18-20.  You can do one, two or three days, depending on your level:

• May 18, 2011  — Google Analytics 101– for beginner analysts and marketers
• May 19, 2011  — Google Analytics 201 — for analysts and marketers who want to “take it to the next level”
• May 20, 2011 — Google Analytics 301 — Google Analytics for Techies.

Jonathan Weber says that GA 101 is like driver’s ed. GA 201 is like learning to drive a race car. And GA 301 is like learning to be a mechanic.

Each day is $499, with discounts for multiple days, breakfast and lunch, handouts and jumpdrives, plus wifi for all (so bring your laptop if you can.) It will be at the John Hancock in downtown Boston.  Most important of all, bring questions, We go out of our way to answer questions about your special problems, so that by the time you leave, you can really do  a lot with your new-found knowledge.

You can get more information about the courses, see the FAQs wrt the Boston Google Analytics training, or just register now.

Robbin

Google Analytics Training in Boston: Seminars for Success is a post from: Google Analytics, SEO, Social Media and PPC blog

Related posts:

1. Columbus, OH: Google Analytics Seminars for Success
2. Last Call: Google Analytics Training in Boston
3. New Google Analytics Training: October 3 in Washington, DC

If you caught our last post on tips for testing across a large enterprise (based on an interview with Symantec’s director of optimization and web analytics, James Niehaus) you’ll recall the number 1 item on Niehaus’ testing wishlist is executive sponsorship. James calls it the “single most critical piece to any organization big or small. If your leadership is on board, they can make things happen for you.”

James is not alone. I asked some of the top conversion optimization gurus what, in their experience, are the biggest challenges for enterprise testing initiatives. There was unanimity in their responses – the number one hurdle is gaining support from the higher-ups.

“WhichTestWon.com readers have told us it’s the largest enterprises that have the hardest time testing, not because they don’t have ample traffic or internal resources, but because of office politics. The bigger the organization, often the bigger the political hoops you have to jump through to get something as high-profile as a web page or landing page tested.”

- Anne Holland, founder, Marketing Sherpa and Anne Holland Ventures / WhichTestWon.com

“In our experience, the biggest challenge for testing in large enterprises comes down to culture. Most enterprise organizations’ cultures are still risk-averse, opinion-driven and brand-centric. Testing introduces risk that many large organizations are not prepared to culturally embrace – employees are simply not allowed to ‘fail.’ Testing is only successful in an environment where risk is acceptable.”

- Brooks Bell, Brooks Bell Interactive

“Organizational challenges present the biggest roadblocks. Most organizations don’t have the same level of structure or clearly defined process for making design decisions as they do for enforcing HTML coding standards (even though better design/UX could significantly impact business results far more underlying code structure.) The reality is that most design decisions in most large companies are made either by committee or via the HiPPO method (Highest Paid Person’s Opinion).”

- Lance Loveday, Closed Loop Marketing and co-author of Web Design for ROI

Testing in large organizations may be more difficult because there are more layers of approvals, less tolerance for risk, a lack of well-designed processes for design decisions and a tendency to rely on experience/gut-feel rather than data. If your organization does not come with a testing culture baked in, you may have your work cut out for you in winning support. But there are strategies for winning over your HiPPOs.

HiPPO who?

“HiPPO” as defined above has also been defined as the “highest paid person in the organization.” Thus, it’s important you sell your case for testing to the HiPPO. However, you’ll find in large organizations, there are many HiPPOs that you need to win over, in many departments and in different levels of your organization. They are not necessarily the highest paid, but they may be anyone High in Position, with Potential Opposition to testing. This could be your boss, the brand manager, the IT department, the finance department, sales or even human resources.

Convincing Your HiPPO

Winning over your HiPPO depends on knowing your HiPPO and seeing things from his/her perspective. What keeps your HiPPO up at night, and how can testing make life easier? What are the fears/uncertainties and doubts (FUD) he/she may have about testing, and how can you educate and motivate?

Let’s look at these concerns from the perspective of various HiPPOs.

C-level executives, Finance

FUD: Testing is a waste of employee time and money

Though testing is typically “free” aside from the cost of enterprise testing tools (and potentially consultants), there’s always an opportunity cost when you assign talent to one activity at the expense of another. Testing often involves multiple departments, so there has to be a compelling reason to back up your proposal.

Your best bet is to literally “show them the money.” Use visual aids to illustrate the increase in bottom line even a conservative improvement in conversion can have vs. “buying more traffic” through email campaigns, paid search, display advertising, SEO campaigns, etc.

A great example from Lace Loveday and Sandra Niehaus’ Web Design for ROI (chapter 2) compares the potential ROI of investing $100K in buying more traffic vs. an equal investment in conversion optimization. (I’ve adapted it into the little table below).

Here you can see even a 2.5% increase in conversion rate results in a 400% increase in ROI vs. the traffic buy. Remix this into a fancy chart and you have a powerful argument!

Strengthen your case with examples of other sites that have benefited from testing (especially your competition), and examples of KPI benchmarks that fall behind industry averages. Great resources for test examples are WhichTestWon.com, ABTests.com and the Marketing Experiments Quarterly Journal.

FUD: You’re challenging my opinion

By suggesting there’s a better way than the status quo, you may be stepping on HiPPO toes. If your goal is to bring decision making authority under your own domain, you always want to be sensitive that a) HiPPOs have feelings and b) the HiPPO is an intelligent person with valuable opinions, often based on experience.

Never attack the existing site or decisions, rather approach it from the perspective that not testing leaves other good ideas on the table, and testing gives you the ability to truly understand what works best for your site and your customers. Offer to involve the HiPPO in the creative process.

In reality, your ideas may not be any more effective than the HiPPO – you never really know until ideas are tested against each other. Don’t give the impression that you or your team has better ideas or that you want to “fix” the existing site. Rather, you want to test various things to learn more about your customers, their behavior and the changes’ impact on usability, conversion and profitability.

FUD: This looks risky…

Educate your HiPPO on the benefits of testing. Fist, demonstrate it’s too risky NOT to test (provide examples from other sites/competitors). Second, reassure the HiPPO that you can reduce risk by testing smaller samples (for example, 20% of traffic rather than 50%) for the first test or two, until you can demonstrate some wins and gain the confidence to run tests with larger splits.

Branding / Legal

FUD: Making changes will undo years of brand building
FUD: Some elements on the site are untouchable for legal reasons

Both of the above FUDs are valid. Legal is easier to win over – just gain approval before running any test. Branding, on the other hand, is a bit more challenging.

“Brand-building is a long-term investment and depends on static, consistent messages. Testing takes the polar opposite approach – rapidly changing, sometimes
wildly different messages. Brand-centric cultures are frequently uncomfortable with the idea of presenting inconsistent experiences to their customers. Brand, legal,
and IT teams often have political sway, and can easily undermine commitment to a data-driven testing methodology.” – Brooks Bell.

Many organizations, irrationally, will protect the brand at the expense of driving sales. Here’s where executive sponsorship can help you tremendously. If you have the support of the highest of the higher-ups, to whom branding is accountable, you’re more likely to get co-operation.

Working with brand managers to gain a better understanding of where they are coming from can also help. If you can understand why branding decisions are made, it may change your opinions of what to test, and branding will be more trusting of the optimization group when it knows their concerns are considered.

In the end, your hands may be tied when it comes to testing copy, colors, slogans or value propositions, but it doesn’t mean you can’t test anything on your site.

IT Leaders

FUD: This is extra work, and will pull resources from more “important” projects

“If it usually takes 2-3 months for a simple design change to be pushed live, the timeline probably doubles when you’re looking to test multiple iterations. And if you’re implementing a new testing platform at the same time you can probably double it again due to all the approvals required, paranoia about 3rd party code on the site, troubleshooting tracking issues, integrating with analytics, etc. So if the normal web development process is like herding cats, testing is like getting the cats to stand up and do a synchronized kick line. Not easy.”

- Lance Loveday

Furthermore, IT and marketing are not always BFFs. Often marketing views IT as a roadblock to progress, and IT views marketing as fickle and unaccountable. As with the branding team, it’s important to get to know what keeps your IT leaders up at night, and understand their processes and existing projects. Don’t take them for granted or take advantage of them, or expect IT to be at your beck-and-call. In order to make testing an integral part of your company, you’ll need to work very closely with IT.

James Niehaus from Symantec suggests learning to speak the language of IT. (This involves learning about their business processes and pain points!) “It’s not about getting your own way, but understanding how testing will support the activities on the other side.”

UX Team

FUD: We’re experts. Are you challenging our expertise?

While UX teams are the experts in web usability and design, they can also be the most likely to make gut-feel decisions based on generally accepted design principles and “best practice.” For example, usability experts may perform heuristic evaluations and recommend changes based on experience rather than customer behavior. While observed usability tests can be helpful to identify problems on a page, they are no substitute for quantitative data on what customers with real purchase intent do on your site.

Because of the UX team’s influence on design and functionality, it’s KEY that they drink the testing kool-aid. Help them understand that site testing is not undermining their expertise. Rather, their expertise is critical in developing and evaluating tests, and that testing allows them to actually expand their skill set through experience of what really works and doesn’t, as opposed to theory, what Jakob Nielsen says or what worked for “site XYZ.”

In Sum

You may encounter other HiPPOs, but these are the main ones that could hinder your efforts to launch a testing program. No matter who you need to convince in your organization, make sure you understand what their fears, uncertainties and doubts are when pitching testing. (Not unlike optimizing your ecommerce site, is it?) Be armed with a strong value proposition for testing (how it will improve their lives – not just yours), and be willing to work co-operatively with them (which may include their inputs along the way until confidence is built).

Note: Google Analytics has let us know that the feature called Multi-Channel Funnels discussed in this blog post is in limited pilot. That means that Google Analytics is testing the feature and its usefulness to a small group of trusted testers, and have not made any plans or a timeline for a full launch.

Google Analytics is finally getting an upgrade to its outdated attribution model in the new Multi-Channel Funnels. Announced at Ad:Tech San Francisco last week, this new model will show how all of your various traffic sources are working together to create conversions on your site. This is without a doubt one of the coolest features to come down the Google Analytics product pipeline, and here’s why.

Most of us are used to dealing with last-click attribution. When a visitor to your site converts, your analytics tool gives credit to the last channel that brought them there. With Multi-Channel Funnels, we get a bigger picture. It’s similar to the Search Funnels features that AdWords launched last year, except now we get to see how all traffic–search, referral, whatever–interacts as your visitors move in and out of your site.

So join me as I take a look at some of the features of Google Analytics’ new Multi-Channel Funnels.

Are You Missing Part of the Story?

Bear in mind that these new reports are only available within the new Google Analytics interface. The good news is that everyone should have access to v5 by now. The bad news? You probably don’t have access to Multi-Channel Funnels. They’re still awesome and once you do have them, they’re gonna change your life. So bear with me.

Multi-Channel Funnels will show up as an option under the Conversions menu. The first report you’ll see is, of course, the Overview. I can almost guarantee that it’ll bring a smile to your face, because it’ll tell you something like:

Sure, it doesn’t really give me a lot of information, but it does tell me that, up until now, I’ve been missing out on important intelligence. Those Assisted Conversions? They tell me that I’ve got traffic hitting the site with one channel and then converting through another. That seems useful, right?

I gotta admit, I’m a sucker for Venn diagrams. The Multi-Channel Mix report on the Overview is really fun for seeing how your various channels interact with one another. Check off the sources to the left and see how they overlap in the diagram:

Move Beyond Last-Click Attribution

For the meat and potatoes, we’ll dig down into the Path Length report. Here you can see how many “interactions” with your site it takes before visitors convert.

Above, we can see that 40% of our conversions occurred after more than one channel interaction. Even cooler, we can see the percentage of conversion value. Over 60% of our conversion cash is attributed to two or more channel interactions. Is it important that we spend some time analyzing ways to get first-time visitors back to our site via other channels? The numbers don’t lie.

But how do we see which channels are best at up-front conversions and which excel at last-click? Well, you’d use the Assisted Conversions report:

The report defaults to the “Assist Interaction Analysis,” which shows us a list of our channels with the number of times they appeared in the conversion path but were not the final conversion interaction. This, compared with the number of times they were the final conversion interaction, hints at the channel’s role. Does it serve primarily as a way to eduacte your visitors about your product, or does it push them to convert? To use the obligatory basketball analogy, is it John Stockton or Michael Jordan?

Know Your Channels’ Roles

This new metric is currently called “Assisted/Last Interaction Conversions,” which, frankly, sucks. I like what Justin Cutroni calls it: the “Exposer to Closer Ratio”. It is, in a word, awesome. If there’s any better indicator of the contextual efficiency of your marketing channels, I don’t know what it is. Previously, if your social media referrals were showing a low conversion rate, you’d probably stop trying, right? Cancel the Twitter account and stop all the Facebooking? Now you can see that they’re just creating prospects who come back to the site via other means and ultimately convert.

If you change the view from Source/Medium to Default Traffic Groups, you get to see something that a lot of marketers have been clamoring for: Channel Grouping. I had a call this afternoon, in fact, where a client asked me how he could better organize his channels in Google Analytics. I told him the same thing I’ve told all my clients for the past five years: use filters. And then I sighed heavily into the phone.

Thankfully, it looks like those days will soon be over. By default, Google groups your channels into various buckets for you:

But here’s the best part: you can make your own! If you want to define your own groups, it’s as easy as setting up a new segment. It even uses a similar interface:

This is big news. One of the biggest complaints I’ve heard from others users is the inability to modify how Google Analytics presents its data after it’s been collected and run through filters. With customized groups, users are now free to change the information architecture of their reports even more than before. This is a very, very good thing.

For more information on the other Multi-Channel Funnels reports, check out the official help page from Google and the videos below:

Multi-Channel Funnels in Google Analytics is a post from: Google Analytics, SEO, Social Media and PPC blog

Related posts:

1. Funnel Problems in Google Analytics
2. Answers to your Top 10 Google Analytics Questions
3. How they landed that 6-digit sale with Google Analytics

With many business units, web domains, and campaigns, large companies face far more challenges performing conversion testing than smaller companies.

I recently sat down with James Niehaus, Director of Optimization and Web Analytics at Symantec to talk shop about the challenges of testing in a large organization, and how Symantec has established a thriving testing program across multiple business units and functions. Symantec currently runs a handful of tests each month, ranging from simple headline experiments to complex merchandising strategy tests, and is driving towards having over one hundred tests per month.

Says Niehaus: “Conversion testing is the most critical strategic lever we have. If properly implemented, it allows you to optimize all your other optimization tactics (like SEO, paid search etc.). Conversion testing should be part of every customer touch point or business function.”

This post is a redux of 10 practical takeaways from that interview that you can apply to your optimization process. (Though most applicable to large companies, many of the tips are helpful to businesses of any size).

Organizational Structure and Support

1. Ensure you have executive support

Niehaus believes executive or revenue owner buy-in is the “single most critical piece to any organization big or small. If your leadership is on board, they can make things happen for you.”

Unfortunately, for many organizations, convincing the HiPPO (highest paid person in the organization) to back testing is an uphill battle. Stay tuned, in our next post we’ll explore strategies for gaining buy-in from key decision makers.

2. Create a conversion center of excellence

A center of excellence is “a formally appointed, and informally accepted, body of knowledge and experience on the subject area. It is a place where the highest standards of achievement are aimed for in a particular sphere of activity” (thank you http://en.wiktionary.org/wiki/center_of_excellence Wiktionary, I couldn’t have said it better myself).

Although Symantec has always believed in the importance of optimization, it has not always been a core competency of the company. Previously, its ecommerce operations were outsourced. As Symantec has brought those activities in-house, a renewed focus on making optimization testing part of the DNA of the company emerged. Symantec’s center of excellence is a 7-person team with a diverse skill set, weighted heavily towards optimization and web analytics, but also including technical support (e.g. experience with APIs and JavaScript).

According to Niehaus, it helps to speak the language of IT if you want your voice heard.

“You must be able to speak in terms of the IT folks to get your code on your site. Once it’s on, you can do a lot yourself. That’s the power of the tool. But it can’t happen until you get the code implemented.”

The center of excellence is responsible for test design, execution, and measurement. In order to deliver effective tests in an efficient fashion this team must thoroughly understand every unit of the business and ‘how they deliver’. An end-to-end view is critical says Niehaus because “I have to propose tests and ideas that can get done.”

The center of excellence also works to help others in the company apply best practices around conversion optimization to their particular focus.

Centralize or decentralize?

Having led optimization for four companies, Niehaus believes that businesses should not all operate in the same way.

“Developing the structure that’s right for you is dependent on your organizational structure as a whole. My division is a reflection of my company.”

Symantec started off with a very centralized conversion program and has now started decentralizing.

“To centralize or decentralize comes down to two things: organizational resources and your approach to global. What is the organizational strategy around globally supporting the business – to centralize or decentralize? Your optimization program will align with that.

If your business lines are standalone or siloed, then it makes sense to have an optimization expert take the lead in each business unit. It depends on how much autonomy each group has and how much their activities need to be coordinated with those of other groups.”

3. Get cross-functional buy-in

By necessity, multiple departments are involved in any conversion testing program. In addition to executive sponsorship, Marketing provides strategic input and business guidance. User Experience designs the concepts. QA validates the code. Development, Merchandising, IT, Program Management, and the platform team are also all involved.

“Until optimization is understood by most business units and embraced, they will see it as more work. And it is more work. If they have not bought into the vision, they will present challenges. Whoever drives the optimization project has to paint the vision and appreciate the challenges of each business unit, otherwise, they’ll be caught up in roadblocks and red tape.”

Technology

4. Aim for platform flexibility

You can only move as fast as your business moves. The larger the organization, the more technological barriers will pose a challenge. Ideally, optimization would be built into the actual platform but a close second is a flexible platform. And the longer the lead cycles, the more difficult testing becomes because of the processes and structures in place. Conversion testing relies on rapid release cycles (daily, not yearly) to roll out tests in a timely manner.

Process

5. Use an enterprise testing tool

Symantec’s tool of choice is Test&Target, part of the Adobe Omniture suite, a market leader in the enterprise space. While free or low cost tools like Google Website Optimizer are sufficient for many small and medium sized businesses, larger organizations that are serious about testing and have a need for advanced functionality should consider enterprise tools.

6. Crawl before you walk

When you begin conversion testing, don’t jump head first into complex multivariate tests. Year one is about finding low hanging fruit and showing value. Year two is about refining execution and doing more complex tests.

“You don’t do too much segmentation or profiling at first. As your program matures, it becomes more sophisticated, and you start doing more advanced tests to find those hidden gems and constantly innovate.”

7. Consider both sides of the optimization coin

Before you launch a test, it’s important to consider the potential upside and downside – there is always the risk that a test might perform worse than your control. Understand your traffic, conversion rate etc. before the test.

“You plug in the numbers to get a realistic understanding of the impact and potential risk. You need to provide four or five scenarios to the business so they understand in advance and are prepared for the unexpected. Experience helps here and you want to watch closely so you can take advantage of good and bad results quickly.”

8. Prioritize tests appropriately

“Test ideas come from everywhere. Each group has different goals and a different view of the business that’s valuable.”

With many sites and pages to test, how do you determine where to start?

Niehaus prioritizes tests based on their potential value to the business in terms of bookings and end units, the level of effort needed to execute a given test, and what areas are available to test. Since you need to wait for test results materialize, you can’t launch a new test on the same page or flow while another test is running.

9. Gain insight from tests methodically

Symantec’s methodology for picking a winner varies depending on test complexity. Single page tests are different from multi-page tests. A/B and multivariate tests are different. Price tests require modeling to predict optimal pricing.

Test&Target is leveraged to get a real-time view of the results, and depending on the test they examine KPIs, average dollar value, product mix and/or how purchase behavior is impacted. Decisions are made based on high-level metrics but take into account the user experience.

Because Symantec sells subscription software, they look at multi-year behavior to understand how changes might affect future business. Consider the behavioral changes you’re generating when you modify the purchase process, messaging, pricing, etc.

10. Roll out intelligently

We’ve written before on Get Elastic about the benefits of rolling out a test winner slowly, rather than flipping a switch suddenly.

Symantec tests regularly across their e-stores in 10 to 15 countries, but mostly in the US and UK as these regions generate the lion’s share of revenue for the business. Once they are confident they have a winner, Symantec switches on the variation in Test&Target first and then will work with the business units to implement the change as a permanent part of the platform.

However, when they find a statistically significant winner they don’t simply roll it out globally. A winner in one channel or region doesn’t necessarily imply a winner in the other channels or regions.

Big thanks to James for sharing his secret sauce for site optimization with us. Stay tuned for our next post, where we’ll tackle the challenge of getting executive buy-in for your testing program.

Looking for help with A/B and multivariate testing? Contact the Elastic Path consulting team at consulting@elasticpath.com to learn how our conversion optimization services can improve your business results.