When you look at your data, you might just find that CRT is far from dead and 800×600 still defines a majority of your audience.

Designers: Time to Go Back to School

In the US, CRTs still rule the schools. There are several solid reasons why this might remain the case for some time to come:

• Durability – It’s pretty hard to scratch the glass on CRT.
• Security – after all who steals a CRT?
• Financial – While LCD prices have come down, many older computers need video card upgrades to tap into them. In addition, there’s another good argument: “if it ain’t broke, why fix it?”

On the educational sites I’ve worked with, it is not uncommon to see 80% of the visitors coming in at 800×600. This might bum out the designers, but they have to be sensitive to that data. If the visitors are captive to the lower resolution, they’re not going to take kindly to a fancy design that causes endless scrolling.

Marketers: This is Your Father’s Internet

Going back to my last article, I mentioned early adopters and how they are likely to upgrade to the latest and greatest. Looking at demographic studies, the top end of that group generally falls into the under-50 crowd with some reports inching up toward age 55.

Even though marketers have good reason to salivate over the potential of the most well-off group of retirees in the history of the country, the fact of the matter is that new computers are pretty far down the list for the majority of that group. They might have a 50″ television set in the living room, but there’s a dusty old CRT in the den.

And it’s probably going to stay that way for quite awhile.

So what do you do with this data?

You dig deeper and you test.

Based on my experience, I’d focus on developing a design that places less emphasis on fast scanning and links between areas of the site. You are likely to find that this segment will spend more time reading your copy and less time looking for the “scent of information”. This means they’ll also get quite frustrated when you make then dig to get what they want.

Of course, I might be completely wrong about your audience. That’s why I recommend you dig into the data and test.

I realize it’s hard not to go glassy-eyed when designers and web developers start spouting phrases like “aspect ratio” and “backwards compatibility,” but listen up, marketing folks! Browser and screen resolutions matter because there’s money to be made.

“What’s that Browser Capabilities menu?”

Every web analytics package has that seldom viewed report called “browser capabilities” (or something similar). In that section, you will find such thrilling choices as:

• Browsers
• Operating Systems
• Screen Resolutions
• … More things you don’t want to know.

But drop on in there and ask yourself these questions:

1. Who converts better? Mac or PC users? Vista, XP, or OSX?
2. What is the average order size of visitors who use 800×600 vs. 1280×1024?
3. Do Firefox users buy more than IE6 users? What about IE7?

That’s Nice. Why Do I Care?

When I looked at recent stats for one site, I was surprised to find out that Mac users were the most valuable customers. In fact, the lift on PC users was over 80%. When I dug a little deeper, I found that XP users, while a very large percentage of the visitor pool, lagged well behind Vista.

What kind of conclusions might we draw from this information?

Affluence is one obvious answer. Vista might mean the visitor has a new computer. It may also mean they’re just early adopters, which again may signal affluence. Apple users tend to have higher incomes as well (at least according to some surveys).

As you may expect, I also found that the higher the screen resolution, the better the buyer. Even more interesting was the insight that widescreen aspect ratios (16:9) had the best rates of all. At the moment, most widescreen monitors are attached to shiny new laptops as opposed to desktops (though this is changing).

Can we infer that laptop users buy more? And if so, why?

The browser question drove similar results… IE7 beat out IE6 by a wide margin (even though visitors were even). Firefox beat them both.

So what do you do with this data?

You dig deeper and you test. These aren’t small scale tests either.

In the case of Vista vs. XP, we’re talking about more than 60% of all visitors to the site. You may find that IE6 has a lower conversion because something is actually broken on the site. If you try out different resolutions, you might find that your site is easier to navigate at higher resolutions.

Geek out aside, I’m talking about large audiences with a clear differences in response. These are segments worth investigating. You may be leaving enormous amounts of money on the table, or perhaps it is just walking out the door.

Rally with your designers and developers, test and QA. Champion their ideas, and you just might find that next campaign hidden amid the most boring data in your analytics package.

Oh, and lest you think I’m biased on the side of big, shiny Macs, stayed turned for the next installment on browsers and screen resolutions – Why Bigger Isn’t Always Better, Long Live the CRT.

Every Page Is Your Homepage:

The premise is simple: every page of your Website must now be considered a homepage. The hours spent tinkering on your current homepage needs to shift into ensuring that every page is a brilliant representation for the keywords and external links that drove someone to your site. Fewer and fewer consumers are coming in via the homepage. They could also care less about what your company does or the other products you sell. They were searching for something specific, and if it’s not there after they’ve clicked on a search result link, they’re you’re roadkill on the information super highway.

[Via: Six Pixels (Mitch Joel)]

I started the month with a visit to Omniture’s Summit in Salt Lake City. The conference was great, even without swell keynotes from Lance Armstrong and Seth Godin (which were both awesome in their own way). Omniture has an ambitious vision for their product line. I’m especially excited about the deep automation they are building with their Genesis platform. If it works even half as well as they claim, it will change the game for many of Omniture’s customers.

Recent Links of Interest:

Avinash Kaushik has a great article on the importance of putting context in web analytics reports.

Debbie Pascoe gives us a look at the new page load time calculations being used to grade AdWords advertisers.

Charles Thrasher has a bitty note about Gord Hotchkiss’ recent presentation at Microsoft. Gord’s talk revolves around Malcolm Gladwell’s thin-slicing hooptedoodle from Blink. Here is a snippet from Charles’ post to whet your appetite:

On a SERP, thin-slicing occurs at the very first listing, either sponsored or algorithmic. The very first listing determines the credibility of the rest of the page. If the reader finds the first listing relevant to their intent, the rest of the listings get a lift in credibility, even when the only thing changing between test SERPs is the first listing! Conversely, poor relevance in the first listing can taint all other listings on the page.

Interested in the Taguchi method? Billy gives a nice overview of the Taguchi Method in this post:

The Taguchi method (also known as robust design) belongs to an engineering discipline called Quality Engineering. The main objective of the Quality Engineering design is to minimize variability in the performance of a product under different environmental conditions.

[By the way, like Dave Taylor, I always think about Tamagotchi when I hear or read Taguchi.]

Ian Thomas gives a solid overview of Gatineau Microsoft adCenter Analytics. Big congrats on the launch of 2.0, Ian!

I hate the term customer engagement, but I respect Eric’s attempt to bring mathematical rigor to his framework. I hope he can devise a decent and reproducible model, but I pray that he can get a better name for his concept.

VeenAndy has some super links in this post about learning AB testing from the pros. I really appreciate his recap of the 7 pitfalls of controlled experiments on the web.

And to round out the lot, here is a post from Matt Belkin at Omniture about spending smart money on internet advertising during tough times. I attended this presentation at the Summit (it was packed). Matt is a great speaker (funny and knowledgeable) and the information in his long post is timely. I wish he would post the crappy PPT his team made. Very funny.

I really like this idea, but I think it is imperative for businesses to quantify the shift rather than just jump in with blinders on.

Here’s an example in the form of Chris’ smackdown of Spencer Wang’s analysis of a free WSJ.com:

There is one thing clearly missing in this analysis: the indirect benefits of the Wall Street Journal reappearing in the online business conversation that it has largely ceded to others due to its subscription wall.

For instance:

* What about the new newspaper subscriptions that a 6x increase in web traffic will generate? (Print subscribers are typically worth five times what online viewers are worth, due to the higher effective CPMs of print media.)
* What about the increased buzz and respect that the ability for bloggers everywhere to link to wsj.com stories will engender, bringing the paper back to the front of mind of media buyers and thus bringing in more ads?
* What about the fact that, in a fierce competitive battles with its cross-town rival, the the New York Times, once nytimes.com went free, wsj.com had no choice but to do the same to maintain mindshare with an audience who are increasingly shifting online?

I don’t know how to quantify any of those factors, but I know they’re all non-zero, and in the case of second, at least, could be large.

I’m sorry, but if you can’t quantify the impact you can’t include it in the analysis. It might be a feel good in a discussion with the board, but at the end of the next fiscal year Lucy’s going to have some ’splaining to do. Of course, with a little noodle work, you can show that even under a conservative scenario the WSJ is going to do better under an all-free model. It just isn’t going to come in the form of paid subs.

(Analysis after the jump…)

Paid Subs

With an ad-driven content site, we might expect about 0.1% of visitors to purchase paid print subscriptions.

Really? How did I arrive at this number?

Quantcast puts the NYT daily traffic at about 1.5MM unique visitors per day. The NYT maintains a subscription portal at nytimesathome.com which has an average of 2.8K unique visitors per day. Let’s be very generous and say that just over half convert. This works out to 1.5K subscriptions per day, which happens to work out nicely 0.1% of the 1.5MM.

(Believe it or not, I didn’t fudge this as much as it looks. I actually wrote down 0.1% and then thought I ought to see what the “real” number might be.)

Alright, now we need to figure out how many visitors the WSJ will get with their new and exciting, all free format. I will use Wang’s peer analysis of roughly 6X pageviews increase as the “break-even” analysis is ridiculous.

Quantcast shows the NYT running at 3.25 pageviews per visit. The WSJ currently runs 2.71 pageviews per visit. If we split the difference, the new WSJ might get 3 pageviews per visit. This should put us at 240MM visitors per year. The NYT averages 2.92 visits per unique so that brings us to 82MM unique visitors.

If 0.1% of those visitors convert to paid print subscribers, the WSJ will see 82,000 subscriptions These are not incremental subscriptions, just raw subs. At $100 a year, that’s $8.2MM. A nice and tidy sum, but not something I’d include in my analysis of free versus paid content considering they are drawing $78MM for WSJ.com subs today.

Ad Revenue

I have to agree with Chris that $6CPM is way too low for the WSJ. They ought to be able to draw double that figure at least. At $12CPM, they are just about at break even. At $15CPM, they are $20MM to the good.

Conclusion

At worst, going free will be break even. Yet, a free WSJ.com should bring in 4.9X more visitors than it does today (taking the difference between traffic increases and pages per visit). This increase will provide substantial opportunities for email address acquisition, RSS subscribers (more advertising), along with other ancillary sales. I’m sure that the WSJ folks have figures available for the value of an email address and their revenue from ancillary sales (and list rental).

So unless there is a major impact on offline paid subscriptions to the WSJ, it seems like the all-free model is the way to go. Paradoxically, Wang doesn’t think this will be a problem:

A potential risk to the free Web site model is the breakdown of the Wall Street Journal’s print subscriptions, but Wang considers this unlikely given steady growth of the New York Times’s print subscriptions after offering free online content and because the core Wall Street Journal reader has above-average disposable income.

“We think a large portion of [Journal] readers also appreciate the physical hard copy of the WSJ, particularly for those who commute to work in the morning,” he said.

[Via: TradingMarkets.com]

Related Links

Free WSJ would need 12X to offset loss [PaidContent.com]
If WSJ.com Was Free: The Numbers at Stake [PaidContent.com]

18 most popular Web Analytics blog posts of 2007:

I am jam-packed on Muffins, Marzipan and Scandinavian Glogg. Nonetheless, I managed to stumble into the living room – in my new Santa underwear – Barely able to turn on my Laptop, I decided to do my general 2007 Web Analytics status using Avinash’s Top Ten Web Analytics Blogs: July 2007 as a foundation (Since most of them happen to be friends of mine anyway) and ask them to provide me with their 3 most popular posts of the year. Find the result below; which is simply a minimum MUST READ list – if you are working in the Web Analytics industry.

[Via: Visual Revenue]

Robbin Steif has a brief (and real-world) example of why E-Commerce folks should go the extra mile to take advantage of the biggest asset of their website – real estate!

Selling with more than features and benefits:

This is the same issue I saw last week in user testing. We had a site with product after product, and they were so similar. The owner certainly understood how and when to make a selection, because he knows the product selection intimately. Some of the users, however, were overwhelmed. He needed to tell them when to choose each one — just like Melitta needed to tell me how to choose a flavor of coffee.

[via: LunaMetrics]

Tell a story. Romance your products. Explain, in-detail, why someone would want to buy your product and how they might use it. Or even better, make it possible for customers to tell their story too.