Why Test?

As a consultant, I have a unique ability to work with different businesses on a regular basis. I am constantly exposed to different goals, initiatives and strategies. However, all companies SHOULD have something in common – Testing components of their website regularly. Whether you are a marketer, merchandiser, IT staff or manager, you have a need to test components of the site to improve overall performance. In all my experience, I have yet to come across the ‘perfect’ website that was without flaw. For this reason, I believe every website should be under constant evaluation and testing. This post is intended to encourage thought about testing and how it fits into your organization.

What to Test?

Test everything! Yes, that’s correct, test everything! Your goal should be to question the validity of ‘status quo’ when in the position to improve performance on a site. Ask questions like:

1. Would featuring product on the homepage increase shopping cart adds?
2. Would moving the call to action button above the fold improve checkouts?
3. Would featuring dog beds instead of leashes improve average order size?
4. Would a larger call to action button perform better?
5. Would less text decrease time on page and increase conversion?
6. Do more images lead to conversion or does video?
7. If we have 2 call to action buttons – one on the top of the page, and one on the bottom, what would happen to checkouts?
8. If we include a shopping cart widget that is visible on every page of the site, would conversion improve?
9. Do product reviews lead to increase in conversion?
10. Does the position of product reviews affect conversion?

All of these questions are: 1) Specific 2) Measurable 3) Actionable. If the test is not specific, it lacks the ability to implement. If the test is not measurable, then why test it? If you can’t change something on the site permanently, then you shouldn’t test it. All tests should fall within these parameters. Naturally, testing different elements of a website are also constrained by technical limitations and the feasibility of testing. There are many tools that exist to help with most technical limitations, but some test ideas are out of reach.

What should I do to start testing? The 3 B’s

#1 – Buy In

Without a managerial sponsor, you will get nowhere. Having a ‘C’ Level member of the company believe in testing is optimal but VP or Director approval will likely get you on the path to testing.

#2 – Budget

Testing will require investment into tools and people. You will need a budget for the tools and the time for people to test. People make the difference and investment should be more in the people than the tool they use.

#3 – Break the Ice

Once you have Buy-In and Budget, break the ice by running a very simple test on the site. The test should run for a week at a minimum, but no more than a month. The goal of the first test is more proof of concept that the tool you have invested in works properly and the results are easy to comprehend.

What Next?

Develop a long list of ideas and implement. Discussions to what types of tests can be run will be discussed in a later post. So, what are you waiting for??? Get out there and get Buy In so you can get started!!!

It doesn’t take long before one of your friends grabs a rock and hurls it toward one of the windows. He successfully shatters one of the perfectly intact panes, and now the act of breaking the windows is a matter of honor; a test of manliness if you will.

So, you reach down and grab a fistful of pebbles. You steady yourself, take a deep breath and heave your fistful of mayhem into the air. Your friends begin to chuckle as your multi-pronged assault bounces timidly off the window panes with “pings” and “pangs.” Your best friend, the one who began the melee, comfortingly pats you on the back and says, “The best way to break ‘em is to use one big rock, not a bunch of little ones.”

This metaphor represents a common mistake many American businesses make when executing a marketing plan. Most companies struggle to achieve market penetration or growth targets because they employ an undisciplined message about their products/services. Rather than pulling their sales, marketing, advertising, public relations and human resources together as a unified go-to-market machine, they treat these interrelated functions as unrelated cogs. In order to achieve optimal return on a sales/marketing investment, a firm must operate at peak efficiency in communicating the what, why, when and where of their products/services. This efficiency of communication can only be achieved when the aforementioned functions are working together toward the same end, in the same way and with the same rewards.

Until American businesses are willing to acknowledge and embrace the inherent efficiencies of this go-to-market model, they will continue to fumble their way through the new world economy and concede position as global business leaders.

Image Source: Mr Thinktank

This week Google introduced new search features that will change the landscape of real-time, mobile and social search forever.  This is definitely the next step in the evolution of information access.

Now, when conducting a search on Google, you will see live updates, news headlines and blog posts seconds after they’ve been published on the web.  Google will rank the relevant results to show the most current returns on the search page.  As Google strives to achieve comprehensive relevance with their search results, they are enabling users to discover information as it is happening.

The search can be refined by selecting “latest results” on the options menu, providing a full page of live web content.  Searches can be paired down even further by filtering results to view platform specific content, like micro-blogs only.  Further, Google has added a “hot topics” feature to Google Trends that shows the topics with the highest amount of online conversation.  All of these enhanced search features are available on mobile devices as well.

Mobile search, specifically, has a few exciting additions that include the ability to search beyond just text.  Users can now enter their queries via voice, location and sight.  This brings me to Google Goggles, a visual search application that lets you search for objects using images from your camera phone.  This is perhaps their most ambitious release, as image recognition technology still has significant strides to make before it is perfected.  But very exciting, nonetheless!  For more on these mobile enhancements, visit the Google Mobile Blog.

Google says that they are relying on “more than a dozen new search technologies” to provide these real-time results.  For more on the algorithm, read here.  In addition to these technologies, Google has announced an impressive lineup of new partners: Facebook, MySpace, FriendFeed, Jaiku, Identi.ca, and Twitter.  Though, even with all of these new partners, Google still has their work cut out for them if they plan to provide truly comprehensive real-time results.  This graph shows that there is still quite a bit of unaccounted real-time content to index (over 50%):

But what does this mean for marketers?  That depends.  Reputation management becomes a heightened issue as real-time content moves to the top of a results page, pushing corporate pages (home page, PR, newsletters etc.) to the bottom.  A company’s social persona now takes precedence, falling in line with what we’ve been saying since the dawn of social media: “your brand is not what you say it is, but what others say it is.”  This visibility empowers consumers, making brand conversation readily available by aggregating it across channels.  The outcome for brands could be good or bad depending on what is being said and the overall sentiment.  It is now up to marketers to find the most effective ways to mitigate negative conversation, in real-time!

Image Source: Manfrys, Spark Media Solutions

A case appeared in the Kansas City Star on Monday about a man who was tagged as a spammer by Facebook.  This shouldn’t be all that shocking or surprising; Facebook is heavily targeted by spammers.

This case, however, is different. This man wasn’t a spammer. He was just ‘not average.’ John VanPelt (the supposed spammer) was contacted by Facebook and told that “Facebook has limits in place to prevent behavior that others may find annoying or abusive” and that his account would be suspended from posting and/or sending messages.

Why is this important? Why am I even taking the time to write an entire blog post about it? Simple.

The bell curve.

While for some it brings back memories of high school tests, it seems to be popping up in everyday ways in analytics.

VanPelt was suspended for not being an ‘average’ user as defined by Facebook. But, what is an average user? And why do we not like those who fall outside of that average range?

An average user is one that (in an analytics, social media, or life sense) performs in a way we expect. This average user is the largest part of the bell curve and one that we can, with some certainty, predict.

In marketing, we target the bell curve. We talk about average page views, average time on site, average traffic path, and average conversion rate. We worry about the 80% of website traffic and attempt to influence that 80%.

We don’t like the non-average user because we don’t understand them. Sure, we can track them and interpret them, but they don’t fit our mold. We can’t answer the behavior questions with them because they don’t do the same things as our average users.

We target, follow and worry about the 80%.

But, what if we are all wrong? What if we start to worry, target and influence the other 20%?

Imagine if our 10% to the right of the main bell curve are our most dedicated consumers. They visit our site every day, sometimes several times, spending more time than our average user, buying more items and in general helping us succeed. The question, however, is what are doing to target them?

Imagine if Facebook called VanPelt and used him as a resource. Think about it. He is one of your most using users. He not only users your service but makes it a daily, integral part of his life. And yet, he isn’t seen as a resource, he is the annoying possible spammer.

Imagine that your top 10% users received special promotions, special website landing pages, targeted messages and in general were treated as un-average.

Imagine that the top 10% feel as though they are known to the company and are a valuable part of the sales process.

Imagine that we target the 10% and they become our focus.

To some this seems counter-intuitive. Why would you target the smallest group?

You target them because they are your greatest resource. They understand your site better than you do because they use it every day. They help move your products. They open your emails. They join the social communities. They comment on your Facebook page,  they follow you on Twitter, they see your videos on YouTube and they download your mobile app.  They are you advocates, proponents and fans of your product, service or company.

Ultimately, the 10% influence the 80%.

To friend or to "unfriend?"

“I can’t believe Sarah unfriended me!”

“Did you notice that Mikey unfriended Joe?”

“I can’t believe Amy forgot to blog yesterday.  I’m so unfriending her!”

Luckily, these aren’t really phrases I’ve heard around VML’s GateOne, but according to the New Oxford American Dictionary, this year the word “unfriend” became a legitimate part of the English language (See the Oxford University Press blog), and I’m sure we’ve all heard it at some point.  Yes, my friends, Facebook (and other social networking sites) has officially taken over our lives.  (“My friends?” – Who am I? John McCain?)

I’m sure that this is probably common knowledge at this point since the Oxford Dictionary’s word of the year was announced last month.  But in discussing “unfriend” with a friend – who I would never unfriend –  this morning, I started to think about the real-life social impact that Facebook has had outside of the time we spend on the site.  What does it mean for someone to unfriend me?  Can I be friends with that person outside of Facebook if they so rudely deleted me from their friend list?  Are my status updates that annoying?  Or even worse, what have I done to deserve such punishment? (See 8 Reasons to Unfriend Someone on Facebook.)

I took it upon myself to have an informal survey of nine people this afternoon about how they would feel and what they would do if a close friend unfriended them from Facebook.  Before answering, most people asked me “Why? Who deleted you?” under the assumption that I was asking for advice.  After assuring them that I was not, in fact, a victim of Facebook deletion, all nine people had similar responses.  First, they’d be offended.  And then they’d call the friend and ask why.  Facebook is a virtual world that mirrors our real lives.  The people I have at some point called “friends,” my personal interests, and brands I interact with are all displayed in a nice organized way for the world to see.  At the same time, deleting any of these things from my Facebook profile is a reflection of deleting them from my life.  It’s a simple, one-click process.  It’s one that, apparently, sometimes hurts (and that Burger King once tried to get us all to do).  It’s one that has become so important in society that it has made it to the black-and-white pages of the dictionary.  “Friend” is now just as much a verb as it is a noun.  I bet Mark Zuckerberg didn’t expect that when he founded Facebook.

Now this is normally the part of the blog where I would tie this back to how social media has had a tremendously large affect on everyone’s lives over the past few years (duh) and its importance in marketing.  I think those things have already been and will be covered over and over in the previous and future GateOne posts.  Instead, I’m going to go find out which of my 685 Facebook friends de-friended me this morning.

(And if you were wondering, I used the words “friend” or “unfriend” 19 times in this blog post.)

Image source: jeffkolar

NY Times had a great article Sunday that that talks about how the iPhone and Apple have changed the space of mobile and dominated the mobile app space – it’s worth a read:  http://www.nytimes.com/2009/12/06/technology/06apps.html

I am new to the iPhone world and for the longest time argued that marketers were missing the mark by primarily focusing on it – however after looking at the available data and mobile trends on our clients sites I stand corrected.  On average over 60-70% of all mobile traffic for our clients is iPhone based and for a phone that has less than 8% of the marketplace that is impressive!

Some interesting tidbits from the article:

• There are currently over 100,000 apps for the iPhone – the next closest is the Android at 14,000
• Developers keep 70% of all revenue generated from the App store – that $1 app can make a lot of money
• Apple says it receives more than 10,000 application submissions each week with most become available in the App Store within two weeks

Chances are you’re reading this on a single monitor – if you are this makes for interesting reading.

In a study conducted by the University of Utah under the auspices of NEC switching to a 24inch monitor over a single 17inch screen for a company with 250 employees could save that company 2.2 million dollars in increased productivity per year.

What this means for the individual information worker, is that they could be getting to complete their tasks faster with an investment of $200-$400

Now this I admit – that I don’t know the underlying methodology for extracting these figures – how much the monitor cost, aptitude of the computer user etc. But I challenge anyone who is using a single 17inch display to go to a 24inch monitor for a week or two and then willingly go back to using that single 17inch display.

View the study here: http://www.necdisplay.com/gowide/NEC_Productivity_Study_0208.pdf

View some graphs below from the study:

Update: Joe wanted me to post my current set up – please note that the main monitor is a Dell ST2410 24inch screen – just acquired – I will be adding the same monitor again to have a dual 24inch set-up. This is once I find a decent USB-VGA adapter.

Many eNewsletters are formatted in such a way that it is difficult to read on most mobile devices. Often times the eNewsletter is cluttered with links that are extremely long and difficult to look past, poorly rendered HTML code, and with graphics that appear as broken links.

With over 51m users (22% of 233m mobile users) using their mobile device to check their email in the US it is becoming increasingly important for marketers to consider the mobile space when developing their eNewsletter strategy and executing any program.

Frequency of Use:
- 13% (13mm) almost everyday
- 5% (12mm) at least once a week
- 4% (9mm) between 1 – 3 times a month

Personal & Work Email Use: (numbers are not exclusive from each other)
- 21% (48mm) use a mobile device to check personal email
- 11% (27mm) use a mobile device to check work email

BEST PRACTICES:
1. Provide Dual-MIME Encoding: This method, of embedding a text-only version of the content, drastically improves readability on many mobile (primarily Windows Mobile) devices and is easy to implement.

2. Provide Alternative View Links: At the top of the email provide a single link to an “Online /Mobile Optimized Version”. The link can than execute device detection and drive users to a version of the eNewsletter that is optimized for their viewing experience. If they are on a computer it can go to an online version and if they are on a mobile device it can go to a mobile web optimized version.

3. Simplify links to the online versions of the eNewsletter. Reduce the code and tracking links surrounding the “alternative view” and “having trouble reading this email?” links to a bare minimum. While removing the tracking links may reduce data captured from the email campaign, the benefit gained by reducing bad user experiences due to phones not being able to click the links is worth the sacrifice.

4. While this multi-faceted approach will not eliminate issues surrounding readability, on most phones—namely Blackberries, Treos and Windows Mobile devices—the readability should be drastically improved.

* All data 2009 ComScore

Last night Mark Zuckerberg posted an open letter on Facebook outlining new privacy changes.  Most notably, Facebook is removing regional networks completely, as some of these “regions” now include over 100 million people.  Facebook has determined, and I agree, that this is no longer the best situation from a privacy control perspective.  Regional networks made sense at Facebook’s inception, as users were specifically college students, and this allowed for a relevant sharing of information.  But as these networks have grown, users were no longer sharing content with just fellow classmates, but an entire city (or country for that matter, e.g. China, India.)

Facebook has developed a basic model for privacy management that allows users to choose between sharing information with friends, friends of friends, or everyone.  Totally simple.  Additionally, individual pieces of content can now be managed separately, allowing for ultimate privacy control.  Allegedly, over the next few weeks, users will be prompted to review and update their privacy settings.  Facebook will even make some recommendations based on current levels of privacy.

And despite the fact that the Facebook community tends to freak out when anything on the beloved platform changes, I think users will ultimately be grateful for enhanced privacy control.

You can view Mark’s entire letter and all of the user comments, who have now moved on to more important issues like the addition of a “dislike” button, here.

Image Source: Gauldo

Some uninitiated web analysts like to jump straight into the reports, which they expect to tell the story of how people are using their site. Since I’ve been on this pony ride more than once, I’ve learned not to open a single report until I get a lay of the land to see what my reports can tell me, and more importantly, what they can’t.

I don’t even need an analytics login to get started on my audit. Because practically everyone is using javascript for tracking, there’s nothing about a web analytics implementation that’s private. I flip on my web tape recorder (I use Charles Debugging Proxy, but Firebug is a free alternative), and just start clicking.

Neo: You could’ve just asked. Seraph: You do not truly know someone until you fight them.

Say Google rang my bell and asked how their Google Analytics Partners page is performing.  The first thing I’d do is take a screenshot of the page, and then start capturing what data is being sent with each click.

My big gray box on top shows the Partners page name as it will show up in the reports. The little gray boxes show the pathing report’s next page values. So before I even crack open a Google Analytics report, I’ll know what they will and won’t show me:

• I won’t be able to distinguish clicks on the Home button from clicks on the Google Analytics logo.
• I won’t be able to distinguish clicks on the left Nav’s Professional Services link from the Submit button in the lower right.
• The Blog clicks won’t even appear in this GA profile—I’ll have to go somewhere else to collect that data.
• The GAAC Support button in the left Nav has no tracking, and so those clickers are sneaking under my reports’ radar.
• No search terms or language options are being captured, so I can’t tell what people are looking for on this site.
• No Region or Sub-region data is collected, so I don’t know where to market the GAAC program.

I’m not trying to score irony points here—the truth is that there’s no such thing as a perfect analytics implementation. But I like to enter the game knowing what the reports will mean, because every site is tagged differently.