Intuition is seductive. No matter how often it disappoints us or leads us astray, we always seem to return to it with open arms. Then it betrays us again. And again.

In a session this morning at the Selling Online Subscriptions Summit, Flint McGlaughlin demonstrated the fallibility of marketing intuition with three recent case studies. Before presenting the results from our optimization tests, he asked the audience to pick the winners. Each time, the majority of the crowd selected the page that underperformed. In one example, only two attendees chose the page that improved results.

This isn't meant to embarrass this group of marketers. On the contrary, these are seasoned, savvy professionals who know the Web and their industries. It simply reaffirms what MarketingExperiments has found time and again with optimization: more often than not, intuition is wrong.

That's why we test, get results, and test again, ad infinitum. This basic but essential mantra is what gets the double-digit gains we saw in the three case studies (see them here: one, two, three). It also keeps our learning process on track when control pages outperform certain changes, or a good idea falls through the cracks.

We should be learning constantly, from our successes as well as our mistakes. That's one of the best ways to train ourselves to outsmart our marketing intuition.

If you manage a subscription website and are looking for ways to boost results, here's your chance to get insights from your peers and experts in the industry.

Next week, I'll be blogging from the Selling Online Subscriptions Summit 2008 in New York.

This year's event will cover several topics, including: offer strategies, building subscriber bases through segmentation and usability, re-purposing content, and proven ideas for getting paid in a slow economy.

Please email me or comment on this post with your questions about online subscriptions. I'll try to get answers from our Summit speakers and attendees, and post the best Q&A of the bunch throughout the week.

And look for us to continue the theme this summer with a MarketingExperiments Web Clinic on optimizing subscription paths.

The same way the telephone replaced the telegraph and the Internet surpassed snail mail, email is running out of gas when comes to communicating with Gen Y.

Texting is increasingly associated with convenience, immediate gratification, instant results, friends, and fun. Email is associated with responsibility, work, relentless spam, and long-winded missives from boomer parents.

So I'm going to make a leap: If your business model falls into the convenience, immediate gratification, instant results, or fun categories (or all four of them), and you haven't yet added texting to your marketing mix, it's time.

Let's look at some hard numbers to bring it home.

Papa John's earned $400 million in online sales in 2007, and in November of last year rolled out a text-ordering service. Today, more than 20% of all sales come from online and text messaging, and profit from those channels is projected to grow by 50% a year.

WJBQ (Portland, Maine) just had its second annual WJBQ "Q Baby Idol" contest. According to a recent Mobile Insider article by Steve Smith, the contest drew 400,000 emails and 231,000 text votes last year. This year it saw almost a million texts and just 250,000 emails.

Mainstream marketers are also forging ahead with texting services as a primary way to connect with their customers.

Hearst Magazines has provided a texting option for its Gen Y CosmoGirl! readers for over two years, but recently teamed up again with ShopText.com to roll out text-based coupons, free samples, and contest entries to their Good Housekeeping, O,The Oprah Magazine, Redbook, and Seventeen readers as well.

Amazon recently launched its TextBuyIt service, which allows customers to enter UPCs and product names in their phones, compare prices, and buy immediately if they like the Amazon offer best.

Email obviously isn't in danger of extinction any time soon, but a recent study by The Yankee Group is projecting 1.7 billion global active messaging users by 2009. Why not start communicating with your future customers now via the channel they respond to best?

As Paul Golding said so eloquently: "Email is like placing a letter in someone's in-tray, whereas texting is like tapping them on the shoulder and saying look at this. . . ."

And if you need a translation of this entry's title, check out the links below, courtesy of lingo2word.com:

f UR nt txtN UR lEvN $ on d table

I hope you were able to join us for yesterday's Web Clinic on optimizing eCommerce websites. It was a lively, actionable session and we were sifting through attendee comments all morning.

(NOTE: Subscribe to our free MarketingExperiments Journal and you'll be notified by email when our Web Clinic content is posted online.)

Our optimization experts -- Flint McGlaughlin, Jimmy Ellis and Aaron Rosenthal -- plowed through a series of research-tested concepts, best practices and pitfalls to avoid with eCommerce sites and landing pages. They also performed a rapid-fire review and critique of five eCommerce websites submitted by our Journal subscribers.

We'll break out some of those sites and the specific recommendations in future blog posts. And if those sites apply the ideas, perhaps we'll get some new success stories.

Meanwhile, one key takeaway was the idea that eCommerce site visitors fall into two main categories: Hunters and Browsers.

Hunters already know what they want, and are looking for a quick, safe transaction, while Browsers need more convincing and a different approach to prompt them to make a purchase.

To be most effective, an eCommerce site must address the motivations and thought processes of both audiences, and take them both into account when developing pages, site paths and conversion funnels. That's the foundation that should be in place before the real fun starts with page elements and design using weighted objectives.

Keeping those two audiences in mind, here are five questions that will help to frame optimization efforts for eCommerce site pages:

1. Which type of visitor -- Hunter or Browser -- is this page trying to serve? (Consider the channels and traffic sources.)
   
2. What are the weighted objectives of this page?
   
3. How does this page stop the visitor and connect with them?
   
4. Does this page instantly communicate my Business Value Proposition to visitors?
   
5. How does this page attract my visitor deeper into my product mix as it relates to the weighted objectives?

When a site or page has these fundamentals locked in, it's much easier to determine a baseline for success, test changes to increase conversions -- and measure the results. Try these questions with your own eCommerce page and let us know what you think.

Want to have your website or landing pages optimized by our experts? Just sign up for the MarketingExperiments Journal. You'll stay apprised of all our upcoming Web Clinics and have the chance to submit your site for a live optimization.

We received some great feedback today from Eric Stevenson, the editor of co-brandnews.com. Eric increased his site's conversion rate by 69% after he implemented the recommendations from our recent Web Clinic.

(See optimized version, right, and earlier version, below.)

"Giving your suggestions a chance to show results, I waited sixty days since rebuilding the site following your webinar participants' helpful comments," Eric said. "Conversion rate rose from 3.9% to 6.6% (30-day results)."

"I should also point out that we took the opportunity to target our paid-click advertising on those keywords which were more relevant -- and cut out those which were not productive. That reduced our ad spend by 60% yet increased conversion 200%.

"In conclusion, design and delivery of the message is foremost and many websites would benefit from your work -- I highly recommend you for that."

You can click here to read our brief containing the recommendations Eric received. It also includes the extensive guidance five other sites received at the same clinic.

You're also welcome to join our next free Web Clinic on May 7. Our optimization experts will be reviewing eCommerce websites, making specific recommendations, and answering audience questions. If you haven't participated in one of our live optimization clinics yet, what are you waiting for?

You don't want to pass up the chance for a double-digit increase in conversions, do you?

Optimizing your transaction pages is one of the best investments you can make in your website. All too often, these are the pages that stop qualified prospects in their tracks.

But while copywriters are focusing on snappy offer language, and designers are worrying about typefaces and buttons, information graphics (like a comparison matrix) can get lost in the shuffle. So does the ROI that these page elements can help produce.

Our sister company, MarketingSherpa, recently reaffirmed this with a test.

By adding a comparison matrix (see image) to underscore the benefits of membership, Sherpa increased free trial subscriptions by 76%.

Testimonials to the right of the new chart and below the call-to-action also reinforced the facts, demonstrated the value of a membership, and helped relieve anxiety.

Why did a simple matrix table get such a dramatic response - especially when its length increased the amount of friction on the page? Because the eyes and mind process the comparison much faster than if the information was written out in copy.

Scan the matrix and the thought process goes something like: "OK, non-members get this. Members get all that. Wow, that's a lot more good stuff for members. Seems worth it to me. And this is a free trial? Let me get my credit card. . . ."

When potential customers are in a hurry, weighing their options and facing a decision, the best thing your transaction pages can do is make their choice easy, comfortable, and fast.

Tony Vacarcel, Marketing Optimization Manager for MECLABS, contributed to this blog post.

What if every time you visited your favorite store, everything looked exactly the same? Or it was so crammed with stuff you could barely move down the aisles? Would you keep going back?

Probably not. So why is this problem so common with eCommerce websites?

Instead of staying lean, they grow larger and larger . . . and things get ugly. The sites end up with dozens or even hundreds of bloated pages with no eyepath; they get overrun by ever-smaller fonts, graphics, ads, and photos; obsolete information and broken links sprout like weeds.

A friend of mine works in a brick-and-mortar store where the owner believes that if you can't turn around without knocking something over, people will think he's going out of business. "The shop looks empty," he says when my friend tries to change things up, knock off the dust, and discount stuff that hasn't sold in years.

The reality is that even the conservative display changes and occasional culls my friend gets away with allow customers to "discover" products that have been there awhile. "The store looks great," they often say. "Did you get some new things?"

On the other end of the design spectrum are my friends Brad, Lew, and Gregory, who own a contemporary home furnishings store in Phoenix. They believe in frequent, major revamps, mashing older and newer things together in dramatic ways that keep their regular customers guessing (and excited).

For most retail websites, the sweet spot falls somewhere in between. And they have an advantage over brick-and-mortars, because eCommerce sites can test a radical redesign idea and see what happens before rolling it out. Multivariable testing gives online businesses the opportunity for "breakthrough thinking" and much higher conversion rates if executed properly.

As our research brief on the subject says:

"When you can test only one change at a time, you are under pressure to think of a 'good' change . . . something you think has a high likelihood of delivering improved results. This can lead to cautious thinking. However, with multivariable testing you can test as many changes as you like. This takes the pressure off and gives you enormous creative latitude, opening the door to breakthrough ideas you might otherwise never have tested."

The caveat is whether a site will get enough traffic in the time allotted for a test. That determines whether the tested changes are statistically valid and significant. In other words, if showing your redesign to only 10% of your traffic means it will take years to get to 95% confidence in your findings, you might want to up the ante.

So go on, move some things around. Try a whole new approach. The payoff could be huge.

And if you're looking for ideas to test, come to our 5/7 Web Clinic on optimizing eCommerce websites.

One of my favorite “Flintisms” is a warning against “unsupervised thinking.”

In essence, it means that when a visitor gets to your landing page, it should be easy to find what they really want. Make sure they know they’re on the right site, and don’t obscure what they came for. Think Alice, always keeping that Brady house in order.

Simple, right?

Not so fast.

Our TSS team was recently brainstorming ways to help a partner with a very cluttered landing page, “featuring” at least twelve different, competing products, plus an extensive left nav list for a hundred product categories, a deal alert sign-up competing with a search field, warranty purchase options, shipping account logins, shopping cart item counts.

The page looked like a Moroccan bazaar.

“You have a shotgun approach on this page. It takes you everywhere,” Flint said.

Now, some folks enjoy strolling through the Internet equivalent of a Moroccan bazaar, nav’ing and clicking through pages and pages of products they didn’t necessarily come for.

It’s called shopping.

Some folks like it, and some (including me) just want to go in, get what they came for, and get out.

In my humble opinion, the current design was friction-city because of all the competing information blasting visitors.

“They’ll lose to someone with a cleaner Value Proposition,” said Flint. “Why should someone buy from this site and come back again?” AKA, no relationship was established.

Another problem was no—zero—eyepath, due to competing constituencies. It looked to me like LP turf battles had brand managers and co-op manufacturers fighting like The Brady Bunch kids (plus Alice) all trying to get in the front seat of Carol’s 1970 station wagon.

Key questions began to emerge. We needed a framework.

What new page design would result in the best "mind trail"? That is, what are people doing now; what do we want them to do; and what’s in the customer's mind? Where do we want to send people to make the most money?

This page needed help, and that’s what we’re all about—what will get it done; for the partner, and for the customers. We’re here to eliminate that unsupervised thinking and clean up that confusing clutter.

Stay tuned to find out how we do it.

FADE IN: GENERIC OFFICE CUBICLE — DAY

A man, shopping online for a high-end product.

We see flash cuts between the man’s face—looking increasingly aggravated—and his computer screen, showing dozens of similar Web sites, each one just as friction-inducing as the next. . . . .

CUE DRAMATIC SCORE, CUE MOVIE TRAILER VOICE-OVER GUY

In a confusing world of online stores all offering hundreds of choices, the lowest prices, and discount delivery, two companies will join forces to optimize a landing page that millions have seen before. . . .

SWITCH TO SUMMER FUN SCORE

This summer, get ready for one of the best landing page tests ever, when MarketingExperiments and their ideal partner team up to increase conversions beyond their wildest dreams. . . .

CUE NEEDLE-BEING-PULLED-OFF-A-RECORD EFFECT

Or not.

CUT TO: OFFICES OF MARKETINGEXPERIMENTS – DAY

FULL SHOT: Director Flint McGlaughlin, backlit by 10 computer screens.

FLINT: What’s the partner’s willingness and ability to make changes?

CUT TO. . . .

Okay, so this work in progress is not likely to be a blockbuster hit, but it does have a viable premise — “X leads to Y” — the movie producing equivalent of a Value Proposition.

In this scenario, lack of data leads to less than ideal project results.

“The point of finding the right thing to test is having the right data,” said Flint in a recent Training and Solutions meeting, as we discussed the challenges of a recent project. “All of the information we need to design this [landing] page is in [the company’s] metrics program.”

Unfortunately for those who partnered with us, someone in their organization was unwilling or unable to get us the information we needed to help them achieve the best results.

Whether you’re a hired gun or an internal testing team, the roadblocks to optimization efforts — corporate politics or culture, bureaucracy or malice, laziness or indifference — may well be insurmountable without one critical element: professional project management.

Marketers should insist that their online testing and optimization projects follow the same project management best practices that have shown real results for corporate IT projects.

Let’s start with project sponsorship. If you’ve ever attended a Project Management Institute course, you know the importance of engaged sponsorship from a strong leader or leadership team within the organization. That sponsor must be empowered to cut right through the smoke, the flak, or anything else being blown or thrown, in order to achieve the established objective of the project.

Having said that, if the project’s scope isn’t adequately defined at the outset, if establishing the key requirements (for example, access to specific data) has fallen short, then producing the deliverables will be a nightmare. The project is set up to fail from the beginning.

Paraphrasing Flint, it’s all about a company’s willingness and ability to make recommended changes.

Even if you get access to the data you need, if the IT team can’t or won’t help you implement, if the project sponsor is a sponsor in name only, or if inertia cannot be overcome . . . well, your optimization project is really done before it’s over. Stick a fork in it and . . .

FADE TO BLACK