MarketingExperiments Blog: Research-driven optimization, testing, and marketing ideas

As online markers and business owners, we have this self-imposed pressure to continually come up with tests to improve what we guess users might want. Yet, in talking with our Research Partners and other marketers, I find there is a real lack of direct communication with end users to help us actually know what visitors want.

We literally create test or optimization ideas in our cubicles, internal focus groups, and mass email conversations around the department. And then we leave out the most important people – our actual users or visitors.

If you just rely on your own analytical skills and creativity without consulting your users, two things are bound to happen. You will burn out by taxing your limited resources to constantly try to develop what to do or test next out of thin air. And eventually you will end up totally disconnected from your users.

When I bring up this topic, I often hear, “Now wait a minute…I have tons of metrics I can collect and pages to test. I can surely figure out my users.”

This is a valid point to a certain extent. We have our web metrics, click map reports, page testing, sales analysis, and the list goes on. However, even with all of those tools, what we are really trying to do is intelligently deduce what our visitors want.

Since it is the holiday season, I’ll use the example of the husband that gets the wife an iron for Christmas. He made the assumption (the keyword being assumption) that since she spends a lot of time ironing she would appreciate a gift that improves her ironing experience.

If Mr. Assumption had actually listened to his wife more about what she would like, maybe he would not have had to spend four days pulling sofa duty as a result. While we all laugh because we either connect to the mistake of giving or receiving such a gift, we have to ask ourselves if we are making the same assumption errors.

We often forget the tools that we have at our disposal to make a direct connection to users and learn what they need from us. Here are three often overlooked techniques:

Exit Surveys
While we all hate to add an extra step, some people do not mind giving information about their visit. This feedback can be invaluable in learning what areas of your website are worth testing to possibly lift conversions and ultimately provide a better experience for your users.

Remember, these exit surveys do not have to be just for people that are abandoning the process. Query visitors that have successfully made it through your signup process, lead capture form, ecommerce purchase…you get the idea. If you have never dabbled in the murky art of exit surveys, there are many companies out there that can help get you started.

A word of caution – don’t be greedy and ask 15 questions, especially all on the same page. People complain about exit surveys and completion rate, but I think the root of the problem is that they ask too many questions. Put yourself in the user’s shoes – would you take the survey? Perhaps test putting just one question per page and see if that leads to a higher completion rate or at least some partial data.  I have also seen some audiences respond better to a breadcrumb-type approach.

Cheat Sheet

• Engage your users through customer service response surveys
• Keep your questions to a minimum or completion rate will be abysmal
• Thank them for their feedback and let them know they are helping make a better future product
• Do not incentivize this area too much or data will not be accurate since people may be looking for a handout instead of trying to help

Actually talk to your users
OK, you may now get your brown bag out and hyperventilate a little before continuing.

Feeling better? Good. For many companies you already possess enough contact information to reach out to your customers. So pick up the phone or mouse, call or email those customers, and communicate directly with the people that make the Herman Miller chair you are now sitting in possible (or comfortable generic cloth one in my case).

If you are looking for feedback on a process, such as checkout, request it as close to the completed action as possible. This will ensure things are still fresh in their mind. And don’t just send a cold auto-responder email.  A personalized email or call will show how important you are taking this process and more likely elicit a better response.

Cheat Sheet

• Reach out to your users and ask them how to make things better
• Ask in a very personal, human fashion to avoid the “system-generated message” feel
• Thank them for their time and perhaps even reward them, but do not mention an incentive until the end to ensure you are not skewing the data
• Follow up as quickly as possible so users have the steps fresh in their mind

Interact with your sales and customer service team
The  teams that interface with users on a daily basis can give you insights to your customers’ kudos and complaints. They might compare your offering to a competing product, request an easier-to-read features page, or need a more robust FAQ section. You can learn a lot by taking the elevator down a floor and talking to the teams that have some of the best internal insight into your users.

Cheat Sheet

• Like Ziggy in the complaints department, front-line employees have unique insights into what your customers want and don’t want
• Take the elevator down a floor and ask probing questions about their customer interactions
• Then take the elevator up a floor and brag about how these insights have helped you deliver ever-increasing net profits

In conclusion, I present you with this challenge. Do not view your users as the sleeping dragon in the cave that you do not want to disturb. Ask, probe, explore, and create a working relationship to increase your site’s performance while delivering for your customers. This will ensure that you are actually creating processes and web experiences both usable and relevant to your users.  I know we only covered three ways to interface with your clients, so comment below on some methods you use to gather customer feedback.

How do you interact with your users and customers? What is your favorite metaphor for making assumptions about what your visitors want? Share your triumphs and ideas in the comments section of this post or start a conversation with your peers in the MarketingExperiments Optimization group.

During the Optimize your Email in Three Steps web clinic, I covered several measurement strategies to help you measure and prove the real value of your email campaigns. I was inundated with questions. Marketers are constantly in search of new “tricks” to find the perfect numbers that help them understand and tell the real story of their Internet marketing efforts.

While I was able to answer a few of these questions on Web Clinic Extra, I wanted to dive a little deeper today with some links and walkthroughs showing how to implement some of the metric items discussed. And please note, while these examples use Google Analytics, Omniture and many other companies have excellent tools with similar capabilities.

Tagging links within emails so you can measure email clicks within your Google Analytics

Requirements:

• Links tagged in email with Google Analytics tracking variables
• Destination Pages from the email with Google Analytics tracking code installed

Walkthrough:
First, with your emails, identify what links you want to track. For some people, just tracking CTA is enough, for others looking at additional navigational links (for example a supplied news article link or a support link) is also valuable data as well. Once you have compiled a list of links that you want to track, visit Google’s URL Builder Tool and start building your links. Please note that campaign source, medium, and name are minimum input requirements for this sort of tracking to work. You also have some remaining variables (name and content) you can use to insert segmentation data. In the example below, you will note that we inputted some demographic and business data:

Once you have built your links, insert them in the appropriate places in your email and hit the red button.

Please note, that using this tool is not necessary to build these links. Once you learn what variables are used, you can build a script that will automate this for you. You can then use internal databases of customer information to create dynamic and automated email tracking.

Also, once these emails go out, you can then create segments on these parameters and get targeted and segmented metrics for your email efforts:

As a final note, make sure you install Google Analytics on the page your audience will visit. This will be required to measure the clicks. Google Analytics tracking code is not required to be in the email, just the landing or website page they are landing on. The tracking script will read the URL variables that you put in your links in the email and recognize the data.

You can also apply these metrics to ecommerce and other reporting data within Google Analytics, giving you a further layer that attributes efforts to the bottom line.

How to incorporate form fields in goal reporting

Requirements:

• Adding the “onClick” markup JavaScript function in the form field you want to track
• Page must have Google Analytics tracking code installed

Walkthrough:
When I reviewed an example goal setup in the Optimize your Email in Three Steps web clinic, one of the steps I mentioned was a form field click as a goal step. In reviewing the clinic comments, I was stunned by the number of people that wanted to know how to do this and for me to explain further, so here we go.

First, as part of looking at email performance, many of us are sending users to pages that have form captures. For me, a great user experience or path to look at is users that click from the email, land on the target page, actually click into the form, and then submit/convert. So let’s look at a typical form code example, and how Google Analytics (GA) ties in:

Standard form input code example:

We can insert an onClick function to the form to capture when a user clicks into the field and complete the information. With this function we will be making a call to the GA tracking function: _trackPageview. What this function will do in our case is when a user clicks into the form field a page will be created in Google Analytics that we specify/create. For example purposes, with the page tracker function we will create the page /dec-email/form-field-email1.html.

After users have interacted with the form field, the /dec-email/form-field-email1.html will start to appear. Just to clarify, this page does not exist, but we have told GA to record clicks and interactions to the tagged form field to this mythological page we have made up. Also, if you are doing email testing, you could create a script that recognizes which email people are coming (e.g. URL variable) from and change this page dynamically as well. So instead of posting clicks to /dec-email/form-field-email1.html page, we use email2.html. Here is an example of Google Analytics markup on the form field:

Also, users tend to be quite “click happy” on forms, so make sure you look at unique visit data on these “pages,” rather than pageviews. Pageviews tend to be inflated because of this user behavior.

Lastly, once these “pages” are created in Google Analytics, you can insert them in goal funnels, just like other real pages. Your metrics will not skip a beat. Here is an example goal funnel that you could create in Google Analytics with the items we have covered:

**Make sure, if applicable, that you select the required step in your goal setup.

Leave a comment below and let me know the next measurement tricks you would like me to pull back the curtains on in future installments of the Magical Metrics Tour. Also, let me know if you find posts about custom or deeper metrics helpful.

For a more in-depth look at making email and social media deliver for your bottom line, check out Email Summit ’10 in Miami from January 20-22. PLUS, Dr. Flint McGlaughlin will teach a Pre-Summit Live Email Optimization Workshop to help you maximize your email capture rate and quality. Register by January 8 to receive an early bird discount of $200.

In Part 2 of our series on bounce rates, we examined how to ensure consistency between your site and the links and ads that drive users there. But once they understand the connection between where they came from and where they are, is there a clear path for what they need to do next?

Sometimes we get so caught up in creating great-looking pages, killer headlines, and excellent copy that we neglect making sure there is a clearly understood goal for what the user needs to do next. It is our job to make the next step in the flow logical and clear so we help them make the intended decision.

Some people arrive at pages and find everything attractive and desirable but there is no clear next step, so they leave. Think of it as having a great infomercial about a knife set. After watching for a few minutes, I’m convinced I need knife that can cut pennies. To top it off, if I act now I can receive a second knife for free. Score, I’m in! However, the infomercial ends there. No phone number or website. I’m excited, I want these amazing knives, but don’t know what to do next. Due diligence with graphics, copy, headlines and other aspects of web pages accomplishes nothing if you are leaving your visitors hanging.

Every marketer I’ve worked with thinks this is pretty basic and they’re doing it already. If you’re nodding your head in agreement, here is my Thanksgiving challenge to you. After the turkey has been devoured, let Uncle Saul and your niece with the pink hair sit down and try to navigate your pages. The results, and their hang-ups, might surprise you.

Now, to address a couple of excellent questions I received about the first two posts in this series…

How do I look at bounce rate for specific blog post pages that have descriptive links that direct traffic directly to the post? – @SquidleyRidley

On social media and other channels, there are many links to single posts on a blog. For example, tweets with a link to a blog post. This can cause, as Squid points out (can I call you Squid?), a situation where people might be more likely to bounce because of the very targeted nature of the visit.

However, as markers and content managers, it is our job to create an engaging experience. A good related post section at the end of each post, and perhaps in the navigation as well, can greatly improve engagement. Here is an example from GetElastic, a blog by Linda Bustos. Note how she engages users with related links at the end of her blog posts:

Just installing a plug-in to generate these related post links might return poor results, so this may require some manual work on your end. Automated systems have come a long way, but sometimes a human mind is needed in building a useful list.

Google is the most fantastic software on the planet. How else can we use Google Analytics to arrive at more intelligent bounce rate reporting? – Sergey, Larry and Eric

For the sake of space, I edited several questions I commonly hear about metrics into the above question and attributed it to three random people from Mountain View, California.

Avinash Kaushik and other metrics evangelists believe that a true measure of bounce rate is not the relationship with navigation, but rather time. While the exact number is different depending on the source, most believe that after a visitor is on the site longer than six to ten seconds they are no longer a bounce even if they never click to another page.

According to the Padicode blog, you can use event tracking in Google Analytics to generate a report that shows visitors that stay for less than ten seconds (or whatever time period you designate). The following code line should be added under the pageTracker._trackPageview(); line:

setTimeout(‘pageTracker._trackEvent(\’NoBounce\’, \’NoBounce\’, \’Over 10 seconds\’)',10000);

In the code, please note the 10000. This number is in milliseconds, so it equals ten seconds. Your report now shows an appended bounce rate of users that meet these criteria. This is not retroactive, so data will take a while.

Corey, you speak in generalities. We want something specific. What number should we be looking for? – Sergey, Larry, and Eric

It depends.

Thanks for reading this series on bounce rates. I hope all my American readers have a great Thanksgiving, and readers in the rest of the world enjoy a fantastic Thursday.

OK, I’m back. I can’t leave you hanging like that. I know you want a number, but I just can’t give it to you. I received many questions about this (and thanks to Sergey, Larry, and Eric for serving as a concrete representation of all those questioners), but there is no one right answer.

For certain pages/process, you have a tighter leash on traffic and should have a lower bounce rate than others. For example, homepages receive traffic from a vast array of sources, thus something in the 40-60% range would be acceptable. If you are using the above analytics tip and are thus measuring time spent, a much lower bounce rate should be the threshold.

For PPC landing pages, that number should be lower – in the 30-40% range. And ideally lower than that if you have a tight rein on your keywords and a healthy account structure with negative keywords. Again, if you’re using the above altered bounce rate reporting to measure time spent, even lower numbers should be expected.

Now that I’ve got you excited about bounce rates…

There is a point of diminishing returns with bounce rate, as there is with focusing too much on any one metric. In an attempt to squeeze incremental bounce rate improvements, you have to ask if your time spent is better spent elsewhere. For example, improving your product or offer, formulating strategic partnerships, observing changes in your competitive landscape, and so on. It is very easy to become a tad obsessive (I am currently recovering and a member of Web Analytics Anonymous as well) and we end up robbing precious brainpower that we can devote to other items.

I hope this three-part series on bounce rates proved useful and helps you gain a greater return from your pages. Make sure you perform due diligence to accurately measure bounce rates, understand why bounces happen, and test to lower that number by providing more engaging pages/process. Then let things cool for a bit and return at a later time with a fresh set of eyes.

What metric would you like us to address next? Leave a comment below about your measurement challenges and we will try to address them on future blog posts or in upcoming web clinics.

In Part 1 of our series on bounce rates, we explored how to drill down into your metrics to find the numbers that really matter. But that left us with an unsettling question. For the users that do bounce but shouldn’t…what is missing that would pull them into your site?

Look at where your traffic is coming from and where it is landing
Many people think that a high bounce rate means there is a problem strictly with the content on the page. While that can be the case in part, you should take a step back and look at where people are coming from and the messages they see before arriving to your site to fully diagnosis a high bounce rate. For example, let’s look at the following user interaction…I’ll be the guinea pig.

I’m thinking of buying a new turbocharger for the Subaru WRX I race on the weekends. So I search for turbochargers.

Then I click on a PPC ad that mentions the following items:

• Unbeatable prices
• Turbochargers in stock
• Free shipping

I think to myself, “Great! This is exactly that I am looking for.” I initiate a click and this is the page I am greeted with:

Where is free shipping?
Where are the in-stock items?
And most importantly… WHERE ARE THE TURBOCHARGERS?!

If you just looked at the content of the site in a vacuum, you would find it acceptable. But users being directed to this site from that PPC ad have expectations that this page isn’t fulfilling.

I have seen many people in this situation look at their site metrics and when they see the high-bounce rate, just keep radically changing the page without any real regard to the user’s thought sequence. They get frustrated when the page continues to underperform.

And remember, I am just using my quest to break the local time trial record in my tuned-up WRX as an example. These principles do not only apply to landing pages or companies running paid traffic.

Text links (on other sites directing traffic to your pages), emails, and newsletters set just as much expectation as paid search banners. For external links, use a research tool like Yahoo Site Explorer to investigate the links to your pages along with the messages being communicated. Then evaluate if your page connects with those messages (If you’re uncomfortable with how your page is presented, contact the owner of the page to edit the links. You will be surprised how willing people are to make those edits if you ask nicely.)

Of course, giving customers the information they need is only the beginning. If we really want to address the bounce rates of key segments we are concerned about, we must get them to act…

On Wednesday, Part 3 will examine how you get visitors to act by giving them a clear path for what to do next.

Have additional questions? Other metrics you’d like to look at? Use the comments section below or shoot me a tweet me at: @ctrentmarketing

Bounce rate is the metric that makes many marketers wake up in a cold sweat. Many Research Partners are consumed (dare I say haunted?) by this metric.

I don’t want to water down the power of bounce rates. But as with most metrics, it isn’t as simple as following a rule of thumb like “keep the number of bounces low.”

So in this series of posts, I’ll take a closer look at what your bounce rates are really telling you… and what is just an imaginary monster under the bed. Let’s look past the anxiety and hone in on what we can learn from these numbers. And, perhaps in the process, help you sleep a little better at night.

Analyze the action that a “typical” user will take on your page

Why does the visitor come to your site and what are they planning on doing there? Answering this question can help you determine if they’re leaving because the site isn’t delivering on some level, or if they simply got what they were looking for and moved on. To show you an example of what I’m talking about, let’s take a look at what may be happening on your blog.

I hear marketers worry about blog bounce rates quite often, so let’s think about how a typical reader views one. You probably post the most recent information right at the top where it is easily accessible – quite convenient for your readers. So the typical returning reader likely reads that fresh content and then heads elsewhere. This would be counted as a bounce since the user has not engaged in any navigation.

Also, readers who already like your blog and have added it to their feed aggregator tend to get in a click-happy mode and just read snippets of news. They might see your new post pop up, read it for a little bit, then move to the next article in their queue. These news reader programs allow users to sift through posts, find an interesting one, arrive at that page, and then leave. Again, if that is the only interaction they have with your site, it will be counted as a bounce.

If you are running a web program that lets you segment your visitors, and look at metrics like bounce rate, break up different sources of traffic into new users and returning users. Only then can you really learn from your bounce rates.

Here is a segment created in Google Analytics to look at new visitors and help us splice the data more deeply:

Now that you have segmented your visitors, what are they really telling you? I wouldn’t be too concerned if older users are bouncing. As we learned from getting in the mind of the typical return visitor, they likely just want to see your most recent post.

But if new users are bouncing, then you have more cause for alarm. Why aren’t your posts pulling them in to a deeper engagement with your blog? If you’re looking at a homepage, what is missing that would pull them into your site?

On Monday, Part 2 will take a closer look at where your traffic is coming from and where it is landing on your site to help you answer these questions.

Have additional questions? Other metrics you’d like to look at? Use the comments section below or shoot me a tweet me at: @ctrentmarketing

We’ve all got that one friend. If you went to high school with Paul Rudd, he’s personal friends with Paul McCartney. If you just got back from St. Louis, he’s bragging about his tan from St. Barts. I call this the 1up effect. And it is prevalent on web pages as well – where each element is trying to outdo the other.

Sure, we can cite many reasons for it. Designing a functional and attractive page can be quite a task. In most cases, we have many things to consider and people to please. The sales team wants us to collect lead information with a capture form. Legal wants visible terms and conditions. Corporate wants branded images. We want our call to actions, great copy, and killer headlines. And the list goes on.

So if you’re not careful, the elements on your page may be competing against each other instead of working in sync to guide your visitors to conversion. I recently commented on this 1up effect in the latter portion of the October 28 web clinic. It boils down to is a simple question – are elements on your page competing so much for attention that you are overwhelming users and losing sight of your main goals?

A great example of this is a TV commercial from Tide. A gentleman was interviewing for a job, but the hiring manager is literally focusing on a shouting stain on the poor applicant’s shirt. Because of this stain, the interviewer misses the conversation completely:

In the same way, strong elements of the page can shout so much that users have trouble using the page for its main goal/purpose. Below is an actual web page submitted by NASCAR.com that we reviewed on the October 28 web clinic. The page seeks to draw attention to a service for watching the races live:

As you can see, color is splashed throughout the page. If you make support items very colorful, you must then use even more color to 1up the support items and make your main goal stand out. However, as you can see in the picture, the main call to action (boxed in red) is totally lost on the page with all of the color and elements shouting for attention. I know that we are all in the hunt to have great-looking, impactful sites, but marketers take heed:

“In the pursuit of cutting-edge websites and pages, do not lose sight of primary page principles.”

A clear and easy to identify goal for the user on the page is a must. Consider the path some users take to arrive at your page. Traffic coming from search and PPC has just seen a list of alternatives (other company listings). If they arrive at your page and do not find a way to engage with your message (buy a product to fix their problem, learn more about the offer, etc), then they click the back button or close the tab and to your competitors they go.

Even if a user does not leave shortly after arriving, support elements that stand out too much can cause users’ eyes to bounce around the page. This can cause them to potentially miss the juicy parts of your copy, product features, testimonials, awards, value proposition, or a whole host of other things.
Plan out your pages with your main goal always in the forefront, then keep support items just as they are supposed to be – SUPPORT items. Use size, shape, color, position, and motion with caution to ensure you are not creating aspects of the page that are in a shouting match for the user’s attention and overwhelming them as a result.

Your challenge is to create a page with elements that work together as a team, not elements that continually try to 1up each other – leave that to your friend who just got back from St. Barts.

P.S. The Tide To Go Stick works pretty well.