The newspapers are so good about reminding me when it’s time to change my clocks; I thought it was only fair to return the favor and do a copyright head’s up.

Unfortunately, I needed to write this article on my website so have to ask interested readers to click through. Enjoy this article about the conversion vs. legal issues of copyrights (and should it be copyright 2007 or copyright 2005-2007?)

My New Year’s resolution is to get off of Blogger and move my whole blog to Wordpress on my website. What’s yours?

Many thanks. And of course, Happy 2007.

Robbin Steif
LunaMetrics

Are there any best practices in landing pages?

The only thing that ever matters is what the customer thinks. It is no coincidence that Matt Roche, the CEO of Offermatica, has the url www.landingpageoptimization.com for his blog.

Having said that, we don’t always get the luxury of testing. Sometimes the page in question doesn’t have enough traffic to run a significant test in a reasonable amount of time. The customer is unwilling to spend the money on alternate versions. Your boss doesn’t believe in testing. Etc. And even if we test, we still need to know what to test.

So with that important caveat (i.e. in this arena, the customer is always right), here are a very few ideas.

1. Are the calls to action strong? “Strong” can mean “4 seats left at this price.” But often it just means, is there a perceived value in (downloading the whitepaper|signing up for the email marketing|attending the webinar)? Does the button say “Submit” or does it say, “Get my free whitepaper now?” (When I went back to get the URL for that last link, I saw that the Matt had written in the comments, “I have a simple rule [about what a button might say.] The button should finish the sentence…”)

2. If the page is longer than one screen: is the call to action at the bottom of the page as well as above the fold? Some people will want to see it at the top, but others will read the whole page. If the call to action isn’t at the bottom, they may scroll up, but they may also wonder “Now what?” Speaking of above the fold - yesterday I looked at the new AVG anti-virus software that I have to buy and it comes with a free gift, I was told. But there was no way to click on the “Free Gift” to find out what it was. Very frustrating. Only after poking around to figure out the price did I find the gift, hidden below the fold (it was another kind of software that AVG sells.)

3. Do you bury your calls to action in a colored box? If it’s a form (so there are boxes to fill in), that may not be a problem. But mere words may carry with them ad blindness.

4. Is it clear where the customer should click? I wish I had a nickel for every time someone told me, in person or over the phone, “You just have to roll your mouse over that one spot and it tells you what to do.” Similarly, I did an analysis for an enterprise software company recently, and noticed that it was hard to find the one hyperlink among all the webinar information.

5. With an instant’s glance, is the customer reassured that she has arrived at the right place? Even though I am in the camp that leans toward the “people don’t really care about pictures,” I do believe that pictures give a subliminal message.

6. Does the page represent someone who can solve the problem? For example, I worked indirectly with a fertility doctor, and the draft of a landing page had a picture of his clinic. It was a low building with palm trees in the yard. It looked like the kind of place where you get a face lift — not the kind of place where you have a successful in vitro fertilization. (I hear that the converson rate of the final version is over 30%, but I only get a piece of the credit, since LunaMetrics just consulted on this one. And I don’t mean to imply that this was the only change, but it is a good example.)

7. Does the page speak the language of the target customer? On the one hand, if your customer is buying tools for embedded linux, they may be disdainful and/or insulted if you take the time to explain how uClibc differs from glibC (or even, heaven forbid, write out what the acronym means.) On the other hand, the customer who is purchasing light fixtures may not have a technical vocabuly. Use words that meets his needs, and where you need to use more technical terms (e.g. xenon vs. halogen), define them.

8. Do you use company jargon that is meaningless to the customer? For example, I have friends in the design business who describe one of their services as “vision catching”. Now that I understand, I realize that the company has done a fabulous job: they have put together words and made them their own. But before I understood, it was just jibberish to me. The landing page visitor may be seeing your site and your company for the first time; it’s not the place to make him learn your terminology.

9. Are you using the appropriate amount of punctuation and capitalization? I have seen this problem a number of times — if you have to yell at someone that this is THE DEAL OF YOUR LIFE!!!! WHAT ARE YOU WAITING FOR??? don’t you think they will wonder just how spammy it is? On the other hand, there is clearly a target market for that kind of message (because most spammers really do test, and they have large “audiences”. And they probably would have learned by now to delete all those exclamation points if that’s what they learned.)

10. Do you really want navigation on the page? This brings up the question, what is a landing page anyway? Maybe it is a page that you design very specifically for a Google AdWord or a banner ad visitor to land on (in which case, you can decide if you want navigation.) But it’s harder to choose what comes up the organic search, so you’ll probably have a page from your real site, with your navigation. So now let’s rephrase the issue: If you designed your landing page for a controlled situation like a PPC campaign, does the navigation on the page give the customer too many opportunities to leave the page without converting? This may well be the case if the ad was about signing up for a free whitepaper, and the job of the page was to get the information. On the other hand, does no navigation give the customer too few opportunities to find what he really wants, so is there an easy way to get to the rest of the site?

11. Does it work with Firefox?

12. Is the type large enough to read without using a magnifying glass? This one sometimes goes hand-in-hand with the Firefox issue — the type size works nicely in IE but is much tinier in Firefox.

13. If you have a form on your landing page: do you really need all those fields? The more you ask for, the less likely the customer is to fill them out, even if only three are required. Check your web analytics overlay to see how this observation dovetails with your reality. Don’t have an overlay? Go to Crazy Egg and install one for free.

14. Do you have web analytics on your landing page? Lots of times, landing pages are created without using the site template, and the analytics, which may be in the template, never get transferred to your landing page.

Robbin Steif
LunaMetrics

I would give the world to see Rolling Stone’s web analytics and to know for certain if they are converting as poorly as I think they must be.

A friend got on their website to buy a gift subscription and was completely unable to do so. Here is how the conversation went:

Robbin: I can’t wait to look at their site to see what they are doing wrong.
Would-be purchaser: Maybe it has nothing to do with their site, maybe it is just that I’m unable to do it.
Robbin, somewhat incredulously: You mean, you had your credit card sitting next to your computer, you were unable to give them your money, and you think maybe it’s your fault?

Well, I love a challenge, so I got on the site and saw the problem immediately. Here is what their home page looks like (and that is how I started, by typing their address right in and landing at Home):

“Subscribe to Magazine” is right up there on top and I chose it. (I marked it with a white arrowhead in the screenshot in case the screenshot came out too tiny.) But when I got to their very nice, compact form, it did not give me the opportunity to enter a giftee’s address, or including a note to the recipient.

Hmm, there must be a way of doing this, I thought. So I poked around and further down the page found all the ways you can subscribe (blue highlighting added by me): I also found that they have not one, not two, but three different checkout carts: one for new subscriptions for yourself, one for renewing subscriptions, and one for gift subscriptions. Oh, I forgot to mention the one that you use to purchase back issues.

Maybe they’ve tested it and found that it converts better this way, but I’ll bet that it’s just easier for them to administer this way. And then you have to ask, is the goal of my site ease of administration (sometimes it is) or is it making money on my print publication when the whole world is moving to digital media?

Robbin Steif
LunaMetrics

For anyone in the blogosphere who has been living under a rock (that must be hard to do at the same time), there is a game of blog tag going on. The idea is, you tell five personal things about yourself and then tag five other bloggers to do the same. June Li tagged me but I am ready to turn the game upside down now. (No offense June, and your article on Mountain Dew not owning their own chatter was excellent.)

Let’s face it, readers really don’t care whether I hate my mother-in-law or that I got kicked out of the sixth grade for telling the teacher that she was wrong. On the other hand, I am dying to know who you all are. (Well, maybe not dying. But certainly, incredibly curious.)

Every once in a while, I find out who reads my blog. Thrice I’ve gotten emails about things that were broken on my blog. Sometimes I find out by seeing myself in another blog. Sometimes I find out through my role in the WAA (I’m co-chair of the marketing committee), and I’ll be going about my WAA business, working with a volunteer who then volunteers, “Hey, I read your blog.” On Friday, someone sent me a RegEx question.

So, in the same vein that Time Magazine voted you Person of the Year — do tell about yourself in the comments. You don’t have to say how you embarrassed yourself in junior high (so what else is new?) or that your boss doesn’t approve of blog-reading time or anything else personal. Just… tell us who you are. But if you don’t know what to say, or just don’t want to say, maybe you will answer some of these questions:

• What do you like about this blog and/or why do you read it?
• What do you hate about this blog? (I’m a big girl, I can take it.)
• Are you more interested in web analytics posts or conversion rate posts?
• Do you have a Regular Expression that you are trying to figure out (and if so, what is it?)
• If you are one of a handful of people who come to the blogsite all the time but never subscribe — why not?

Well, that’s five suggestions to my five readers, maybe I can feel like I still played the game.

Robbin Steif
LunaMetrics

I managed to skip most of the analytic seminars at SES Chicago (after all, Brett Crosby and Matt Belkin didn’t even show up as advertised to the analytics vendors seminar), and so didn’t hear a lot about analytics (except here.) Every once in a while, someone would say, “Check your analytics before you decide.”

The really interesting “Check your analytics” comment was with regard to duplicate content. Adam Lasnik from Google Webmaster Tools (formerly known as Google Sitemaps) insisted that Google doesn’t punish site owners for duplicate content since most of it is innocently created. But I think he really meant, “We don’t hand out a minus 30 penalty for duplicate pages.” In fact, Google indirectly punishes you by putting duplicate pages in the supplemental results where my experience has always been, they rank poorly.

His comment about duplicate pages being created innocently is a good one. Just think, we have www.mysite.com and mysite.com. And then there are pages that are “print only” versions but are basically dupes of the real thing. We end up with secure and not secure pages for the same page by accident (I’ve seen that twice in the last week.) For most of these problems, the answer was, check your analytics, see which page is the most popular, and then be sure that the search engines don’t try to index the less popular one (using either your robots.txt file or a noindex tag.)

Robbin Steif
LunaMetrics

Just because this post is about something cool I figured out in Google AdWords doesn’t mean that it’s not about analytics and conversion. (Well, ok, it’s not mostly about analytics and conversion….)

I’m probably not the first person to figure this out, but I definitely stumped “my” Google Ad rep. Here was the issue and the answer, and I will pretend that my company is the customer so that I can use examples.

Issue: If I use a big single head word like analytics and I let Google AdWords use a broad match configuration, the term pulls really well when the searcher uses at least one other word with it. So, for example, if the visitor types in web analytics or analytics consulting or Pittsburgh analytics, the click through and conversion is excellent. You might be thinking, what about when they type in stock analytics or financial analytics? Those don’t matter because I already told Google never to match when the searcher uses stock or financial. You know, negative match.

The problem is, the term doesn’t do very well when someone just types in analytics - it can mean too many things. I could try to find all the phrase matches possible, but I am too lazy for that.

Answer: “So,” I said to the Google AdWords guy, “Can’t I do negative exact match? Like this: -[analytics]. That way, analytics is still a broad match term, just not when it appears alone.”

He looked at me, and he looked at my paper, and he looked at me again. “Well, I guess it should work, theoretically.”

With that I thought, this is ridiculous, I am just going to try it. So I stood there while we talked and added it and AdWords took it. Later I went back and verified that it worked, and that the click throughs and conversions were increasing. I only have a couple days of data right now, but am at a click through rate of 6% for that campaign with no decrease in conversion rate. So I am getting ready to lower the bids.

I learned that I needed the exact negative match by implementing this Google Analytics hack from GA-Experts.

Robbin Steif
LunaMetrics

Today, Matt Cutts showed the browser technographics for the last month of his blog. He was surprised that his IE and Firefox users were split, given that techies are so inclined to use Firefox (I wasn’t at PubCon but apparently they did a quick survey and found that it was two to one in favor of Firefox.)

He has a techie blog, right, a techie audience. So why is his audience not more highly skewed toward Firefox? My blog is just about the same as his, half and half, so I used my stats to evaluate the thoughts below. (Not a great idea, but better than no data.) These were the things that I considered:

Hypothesis: Firefox users “get it” faster than IE users so don’t have to visit each post as often. After all, he was showing visits, not unique visitors. Hypothetical answer: Not only is this technological snobbery on my part, but it’s just not true. I couldn’t find any trend to prove that one.

Hypothesis: IE users are more likely to visit (e.g. to search and click through from the SERP) when the blog post is going to be less technical. Hypothetical answer: Possible. He did show how skewed his WA were when he did the post on hacking. However, to really feel good about this one, I would want more data.

Hypothesis: SEOs may like to use IE but don’t want to admit it in a survey where everyone can see. Hypothetical answer: I like this one. I was very surprised at how many SEO presenters used IE while speaking at SES Chicago.

Hypothesis: Tech-heavy Cutts readers are feed subscribers and they don’t need to visit his blog to read the post: it comes through in their feeds. Hypothetical results: This is my absolute favorite, especially because he showed how Firefox-heavy his analytics were when he did an extremely techie post. The post had over 100 comments (I think it was 132), and to comment, you have to actually visit the blogsite; a feed isn’t good enough. When they visit, the WA finally sees them. Of course, there are probably people who subscribe and have to click through to actually read the post (like My Yahoo! subscribers), but many do not.

Robbin Steif
LunaMetrics

It was interesting to hear, this morning at one of the last SES Chicago 2006 panels, how some retailers are unhappy with all web analytic packages and are dying to design their own in-house packages. Some have already started.

This was a panel with five retailers (Pier One, Baby Age, SpinLife, Tool Barn and Vintage Tub & Bath.) One agency/consultant, Range, was on the panel. So, not very statistically significant but interesting nonetheless. The moderator asked all the vendors to address what they liked/didn’t like about web analytics. Note: I didn’t tape record this, and did the best I could to capture what people actually said, but as usual, the mistakes are mine, mine, all mine. All the comments in square brackets are mine (and aren’t a list of regular expressions…)

Pier One. We use Webtrends 7. Our site generates multiple duplicate URLs, especially after someone uses our on-site search, and the analytics packages can’t handle it. We are trying to create our own software. [I found that Pier One has GA runnning on their site, but obviously can't see what server side analytics they have.]

Baby Age: We have pretty much used every package out there except Coremetrics and have been pretty unhappy. For a while we were using Omniture and a couple of ROI trackers and they were off by 30 percent. We are using GA now and it seems to be the most accurate, but we are writing our own analytics now. We are working with local universities who have students getting masters degrees in mathematics and statistics and are getting histograms that we would never get from our analytics package. We are working with Penn State — their professors are looking for projects like this. [Baby Age currently has both SC and GA installed on their site.]

SpinLife: Our experience was similar to Baby Age. We launched in 1999. We had our own analytics up front. We couldn’t wait until we could afford a big package, and decided to go with CoreMetrics. We found that the large packages are skewed toward customers that sell simple SKUs. It was inconsistent with the kind of products that we sell. Coremetrics tried very hard and they could not make it work for us (lots of our transactions finish over the phone.) We went back to our home-grown package. We are moving toward a package called RedZone. We need something that handles phone orders well. [Sorry, I couldn't find the package she was referring to, any comments are welcome.]

Baby Age: Our biggest challenge was tracking deferred revenue.

Tool Barn: We use Index tools. We find that we don’t have much time to read our reports [audience bursts out in laughter], and we like Index Tools because the info is force fed to us — it is emailed to us. Index Tools also emails exception reporting, huge spikes etc.

Range: from the agency side, we have looked at just about everything and seen all these problems. The thing that drives us crazy is when people have four packages. None of those things are ever going to line up.

Tool Barn: Along those lines, we are participating in a case study with a dozen packages that will all be on our site. [I assume he means, the study that SEOmoz encouraged. ]

Vintage Tub & Bath: We use Omniture. You can have the best tool in the world but if you don’t have someone looking at it, who cares? [I swear that I heard him say this, but when I went to his website to get the url correct, I found that they only use Google Analytics. Perhaps I misheard.]

Audience question: How do you get buy-in from management to create your own tool? Answer: Buy gifts for the IT staff. [Laughter]
Pier One: I really mean this, in a way. IT loves to make things work correctly, and I watch them create stuff for me on their spare time in a few days that I expect would take months. [This reminds me a little of the guest post that Jodi McDermott wrote on IT vs Marketing.]

Robbin Steif
LunaMetrics

Although SES is not primarily a WA conference, there’s still news and gossip.

Last night, before I went out drinking with my friends from FeedBurner, I got to meet Jeff Turner, formerly of Blogbeat.net, whose company was acquired by FB in July. I sat down with him (well, I stood) and reviewed all the ways that I love Blogbeat and why I have to turn to Google Analytics when I need hard blog data. However, I’m expecting great things of the new integrated FB and Blogbeat.

By now everyone probably knows that WebTrends purchased ClickShift and announced it yesterday. I met the ClickShift CEO, John Rodkin, at the WebTrends party tonight. However, he was wearing a WebTrends shirt, so I didn’t really understand that he was “the man” until I started to wonder why he didn’t know any of the other WA vendor names. “So,” I asked him, “Did WebTrends buy Clickshift [which is a highly automatic, intelligent paid search tool] because WebTrends wants to get into search more?” I didn’t have a tape recorder but I think John’s answer was, “WebTrends wants to be the CMO’s best friend.” Then I talked to Jason Palmer, VP Marketing at WebTrends, who is incredibly impressed with John and his team - he thinks John is a genius.

Moving on, I was dismayed that Bill Bruno of Stratigent gave away all his passes to the free Google party and didn’t leave one for me. Bill and the Stratigent CEO, Josh Manion and I all got to meet the analyst from Oregon who calls Josh and me all the time, Chad Bartley. (Parentheticaly, many thanks to all the analysts who signed up to talk to investment bankers.) Speaking of which, I got to meet Gary Angel of Semphonic who was exhibiting here too — Gary signed up to talk to investment bankers after I posted on the Forum but wrote, “Robbin, your request was the strangest one I have ever seen on the Web Analytics Forum.” On the trade show floor, while I was waiting to talk to Gary, one of his employees tried to give me a printed book of all his blog writings. “Do you work with any web analytics?” she asked me.

Chris Spiek of Awecomm (WAA Marketing lead for the WAA website) and Nancy Taffera-Santos (co-chair for WAA Events) crashed the Google party with me, after the WebTrends party, but no one noticed. The music was too loud and there was no food (I am not sure which is worse.) Jason, your party was much nicer, wish I had had the chocolate-covered strawberries.

I rode up in the elevator with John Marshall from ClickTracks, musing about writing this post. “Well,” he said. “Everyone loves gossip.”

Robbin Steif
LunaMetrics

post script: At 12:41 this morning, one of the many people I mentioned wrote me to tell me that my blog gets crawled fast. The post had already pinged the services, been crawled by Google, and he had received an alert that someone was writing about him. (Always a good idea. I use Google Alerts but am not too happy with them, anyone have an alert service that they love?) rfs

So here I am, actually writing my post in the FeedBurner office. I figured that while I was here, I would learn about IE7 and Feedburner.

But first, let’s not forget what’s really important: The FeedBurner offices. Every bit as cool as the company. All the walls are different colors, and the tables are set up like industrial picnic fences. You can feel the energy in the air.

Anyway, John Z, my always-responsive correspondent, explained the issues to me.

The new IE 7 has the ability to subscribe to feeds (just click on the feed button.) But since it’s not a “standard FeedBurner thing,” I couldn’t understand why I was getting FeedBurner statistics showing me that I had IE7 subscribers.

The issue, John said, is about autodiscovery. Many blogs and podcasts have set feed preferences in their source code. That way, when someone goes to subscribe to a feed with IE7, if they have their autodiscovery set up as FeedBurner, they can track the subscription through FB. Blogger users like me are out of luck - Blogger doesn’t support autodiscovery.

So why do I still see IE7 subscribers? When someone clicks on a generic feed link on my blogsite, and they are using IE7 as their browser, IE7 knows that I have a feed (and knows that I have FeedBurner, since that’s the spot that the orange feed link usually goes to), formats it like a feed and serves up the opportunity to subscribe. Like this - see the Feedburner address in the address bar (with their flaming icon)?

Summary: If someone uses IE7 to subscribe to a feed with FeedBurner autodiscovery enabled, FeedBurner always measures it. If someone uses IE7 to subscribe to a feed like mine (no autodiscovery) they have to use the icons on my site, not merely the IE7 icons, for FeedBurner to notice it.

So for all you IE7 subscribers whom I don’t know about: hello out there! (And it is oh! so cold in Chicago.)

Robbin Steif
LunaMetrics

ps John Z gets all the credit for this post. I am just the journalist here and get credit for all the mistakes.