This is your downloadable version of a presentation first conducted April 12, 2007, by:

Anne Holland, President
Stefan Tornquist, Research Director

#1. Click this link to download the PowerPoint presentation PDF, including eight new data charts and an eyetracking heatmap (yes, you may share with colleagues):
http://www.marketingsherpa.com/tele/ECOMM-IN-041207.pdf

#2. Click this link to download the MP3 audio file:
http://www.MarketingSherpa.com/tele/ECOMM-IN-041207.mp3
(Audio download note: This is in MP3, which almost anyone with a PC or Mac can play. If you'd like to listen now, just do a regular mouse click (left side) and it will start playing fairly quickly. If you'd like to download and save to listen later, just do a right-click with your mouse.)

#3. Here is the transcript of the teleconference:

Anne Holland: Good afternoon and welcome to MarketingSherpa's annual Ecommerce Search Marketing teleseminar. This is Anne Holland. I'm the president here at MarketingSherpa, and joining me today is Stefan Tornquist, our research director. Hello, Stefan.

Stefan Tornquist: Hi, Anne.

Anne Holland: Now, I'm all excited because this is the second year in a row that we have presented new data from our research on ecommerce marketing. Can you tell me where the data is coming from?

Stefan Tornquist: Sure. If folks want to go to the second slide, like all of our reports, this one is based largely on our own primary research. In fact, more than any report in Sherpa's history, this one comes in at about 95% Sherpa-generated content. The primary part of that is our own benchmark survey this year of over 1,900 ecommerce marketers.

We were also in a position to conduct several surveys of consumers. I think, in total, over 2,000 consumers were surveyed on a variety of topics. Then, we also do lab and partnered research, including eyetracking research, that we'll be taking a look at today, and then we include some of the "Best Of" research that's out there to fill in the gaps.

Anne Holland: Great. Now, you should have received a hotlink to a PDF along with the dial-in number for today. Click on that hotlink, download your PDF, and that contains the charts and examples that Stefan and I will be talking about. Now, we're going to move to slide #3, and this is the first data point from our survey of more than 1,900 ecommerce marketers. This was conducted in January 2007, and one of the most important things we asked is, "Where are you getting traffic from for your site? Where are your shoppers coming from?" Of course, here, we're comparing 2007 to 2006. What are the biggest changes that you're seeing?

Stefan Tornquist: Well, the No. 1 surprise here was that the number has changed so much. If you look at the bottom of the chart where online advertising and affiliates, where the needle's barely moved, that's really what we expected. Now, when you look at the changes in paid search versus direct-to-site -- in other words, if somebody clicks directly to a retailer -- those are very big differences in just one year, and there are a couple of things that we know are going on and a third that we suspect.

First off, this represents the whole universe of companies responding to the survey -- and I should point out that when you carve out the larger companies, this change that we see in this chart is a lot smaller and direct-to-site edges out paid search. So, for the larger companies, not so much of a seismic shift.

Second, we do know that paid search grew last year in terms of companies' keyword volume and spending. So, some increase would be expected. But I think the other thing that may be going on here is that we've seen a big adoption of Web analytics over the last 18 months, especially in the SMD -- small to medium-sized company -- range and, so, I suspect that a part of this upswing is, in fact, better measurement and a better ability to allocate where visitors are coming from and not entirely just a groundswell in paid search.

Anne Holland: One key, of course, that everyone should look at is that of the three that you're measuring that are not just people typing in your URL directly, two of the three are search. Search is huge.

Stefan Tornquist: Absolutely, and one note on SEO. This is a case where you'll see 14% reporting -- that that's where 14% of the leads are coming from -- are visitors. One thing to note about this is that really there are two groups, and one is seeing a very low SEO contribution, under 10%, under 5%, and then there's another group that sees quite a high amount, 30% and sometimes higher. Now, that all shakes out to 14%, but this is a case where the average doesn't necessarily reflect where most companies -- it's not a bell curve.

Anne Holland: The fact of the matter is, though, if you're in one group, the under 10%, or even under 5%, group of SEO traffic, you are spending, too much money on getting traffic to your site. You need to be spending more investment in search engine optimization because the traffic that does come is generally very high-quality traffic, and there's a fairly low cost because you're not paying per click. You're just getting yourself optimized.

At the same time, if you're totally dependent on search engine optimization for your traffic -- if 30% or more of your traffic is coming from search engine optimization, then, of course, you're still vulnerable because if somebody like Google comes up with a big change in their algorithm, you're spheiling off staff next week and then hiring them back again. It's just so delicate. It's terrifying, in a way, to have your entire company be balanced on somebody else's decisions.

On one hand, you have lots of control. On the other hand, you have highly economic marketing. I would actually like to give a pitch out -- everyone should be doing more emailing to their house list -- but this speech is about search, so let's move on to our next search slide, page 4. Now, this is from a report card. This is from an observational study, isn't it?

Stefan Tornquist: That's right. This is from this year's Ecommerce 250 Audit. We increased the number of sites we looked at, dividing them between small to medium sized and then sorted by brand retailers, of which we studied 150 and this chart is of those larger retailers. What we did was, for each site in the audit, one of our analysts identified key words that reflected primary product categories and specific products that were featured on the site at the time of the audit. In other words --

Anne Holland: So, a category would be like automotive, whereas product would be a Honda CR-V, or something?

Stefan Tornquist: Yeah. Books versus John Grisham, say.

Anne Holland: OK.

Stefan Tornquist: Our analyst went through and examined how these different companies in all the different product categories performed against these key word groups because these were the ones that were most likely to be emphasized and paid in natural search.

This is a chart across all product verticals. When we split them out, we actually see a very big difference in their performance. For example, almost 80% of the housewares and home furnishing retailers appeared on the very first page of results in those keyword groups. If you compare that to drugs and beauty, a much more competitive paid search space, that's only down around 36% of those companies appearing.

Anne Holland: There are some really wild swings by industry vertical.

Stefan Tornquist: The other thing we wanted to look at what was their search strategy, whether the emphasis was on category terms or product specific terms, and there's also quite a swing by vertical in that regard, as well.

Anne Holland: What I think is interesting here, of course is, bearing in mind that you do have to appear on the first page -- no one -- hardly anyone ever clicks through to page two, and what is interesting here is that, again, SEO is much harder. That optimized listing, which is worth it, so many fewer people appear in it.

Now, of course, you're competing a lot, but you're also competing for the paid search. It isn't like there are so many more paid search spots to get on that page. I think people aren't putting as much effort into organic as they could be. I think this is just something that is worth more investment. In particular, you're going to want to look in the guide under vertical and see the verticals in which, if you made a little more investment, you really could own the organic spots in that vertical because nobody is really, perhaps, trying as hard as they could.

Let's move onto the next slide because I want to show you an example of one clever campaign. It's a really clever search marketing campaign. It was very inexpensive. This is actually -- as you see here, here's a picture of Mischa Barton, the TV star. This is a campaign from the ecommerce site Ice.com. They sell jewelry. They tested launching a blog about celebrities wearing jewelry. Now, we all know, if you've ever studied search engine reports, seeing what types of terms people search for, a lot more people are searching for Mischa Barton than are searching for, well, pretty much any other term you'd want to be under. This is fabulous.

If you're searching for Mischa Barton, up comes this blog that actually is sponsored by Ice.com and it says, "Oh, here's a picture of Mischa, and, by the way, she was wearing this wonderful piece of jewelry, and by the way, we happen to carry it, or we carry something that's a knockoff of it or that looks like it, and click here if you'd like to buy this." It started out slowly. They got a little bit of traffic at first, and it ramped up pretty quickly. … What was interesting was that around the holiday season, which is the peak for this particular site, their very first December with this blog, they ended up with more than $50,000 in sales that just came from people who sort of randomly came to the blog, were surfing celebrities, found this on a search engine somewhere, went, "Oh, gee, isn't that neat, and yeah, that's great. I'll buy that as a gift.”

That's $50,000 in found revenue from search engine optimization, and it really was just a couple of hours a week to toss this blog up there and look for celebrity pictures, which, I believe, they had an editorial assistant that was doing anyway. So, you can do some clever stuff for pretty low cost if you're considering it and take advantage of the terms that people really are searching under, even if they're not searching directly for the name of your particular product. I love that celebrity tie-in.

Now, why don't we go onto our next slide. Of course, the key is, once you've done all this wonderful search marketing, what happens when they hit the page they reach? What happens when either they hit your home page because they see your URL and perhaps they type it in directly? What happens when they hit your landing page? What happens when the shoppers actually come to you? And, we believe that there's another role of search in this end. Stefan, can you talk us through this chart?

Stefan Tornquist: What we basically have here are two different groups of people -- the search box people and the links people. This divides up all the visitors to retail sites thanks to a survey we did with Guidester and what this basically supports is this idea that people do fall into these two groups and they kind of stick to their preference. People who like menus stick to menus and search box people, as well.

When you take out the small percentage who say that they decide to do what they want when they get to a site, that they could go either way, you basically see an even split. 43% use the search box and 44% use links. Now, these numbers are going to shift over time. You know, the search box does seem to be gaining on traditional navigation, but for the foreseeable future, both are so important that you have to spend a lot of time testing, tweaking and optimizing those, and that's what we'll spend some time on now.

Anne Holland: It's interesting. We've done eyetracking studies of home pages in the past, and what really interested me was how many people came to that home page, whether it was Wal-Mart or it was Best Buy or wherever it was, and didn't look at anything in the middle of the page. They just sort of looked at that edge. They looked at, where is the search box or where is that little text navigation and moved on.

The eyetracking studies indicated that people really aren't shopping for entertainment or browsing. They're really searching. They're treating your website as an extension of a search engine, and this data kind of shows that, indeed, yeah, they're coming in and they're not looking at the pretty pictures and seeing what's around. They're really coming in and they're like, "All right, where's the search box? Let me use it. OK. Where's the navigation? Let me use it." In a way, shopping is still search activity.

Now, this next slide you'll see -- actually, these are real-life samples from three completely different ecommerce websites, all of which were tests that turned out to be massive winners. The one on the top left is from a website called Carrot Ink. They sell inkjet cartridges. The one on the far right, Book a Flight, is from Jet Blue, and the one on the bottom, which says women and men and has that little alphabet there, that's from Perfumes.com.

In all three cases, the Web designers created something to help people when they got to the home page, or to a landing page, or in some cases when they saw an email -- to help them search. So, sort of navigation, making the search box even more useful. It's sort of navigation beyond your classic just search with the little box there, making it more engaging, helping them get where they want to go more. What they found in all three cases was it improved the conversion rate of the site overall, sometimes substantially.

For example, for Carrot Ink, that little "Find your cartridge" search box gave them a 96% lift in active shopping; in other words, a 96% lift in the number of people who came to the home page and interacted with the site, as opposed to just going, "Uh, I don't see what I need," and leaving.

For Jet Blue, they saw such a substantial lift from adding this easy -- they call it the mini-booker -- to their home page, that they also began adding it to almost all of their email templates, as well, and for Perfumes.com, they actually saw a 6% sitewide gain on conversion just by adding this little box. What you do is click on the little letter of the alphabet to spell the perfume name instead of having to type the perfume name.

This, to me, is fascinating because it sort of says, you know what, maybe consumers aren't typists. We're used to typing all day long maybe for our jobs. Maybe most people don't like to type. Maybe they'd rather just click on the word and in a way, it's like they're clicking on a keyboard. It works like crazy. They actually found that seven times more people would click on that little alphabet to spell the work with the click than would type into a search box on their home page. This is Perfumes.com. This is not a cheap site. These are pretty pricey products.

I found that fascinating. It's saying, let's rethink even the way we offer search, first of all, in our home page, in our landing pages and even in our e-mails. What are easier ways beyond just a box and the word search in it to get people more involved and more active? These are all super clever ideas and I think anyone can steal ideas from this.

Now, why don't we move on to category pages. This, of course, is one of the fun eyetracking heatmaps. Stefan, can you talk us through this? This is from Best Buy.

Stefan Tornquist: Here, we focused on how people use internal navigation, and we added a new wrinkle to the test by conducting mini-focus groups after people interacted with the pages. Not only did we see how they interacted, but we asked them why, what they liked and so forth.

This category page is a good example of an increasing trend in using text lists essentially as lenses with which to see different product categories. If you've got something like computers, you can use different brand names, different price ranges, capabilities, etc. On this page, you see a number of different ways of approaching this very large product category. People seem to respond well to these text lists, especially when they're put together well.

It's interesting that the days of advanced search, when everybody was working on having these very complicated advanced search functions on their site, basically what these text lists do is bring some of that functionality into a very easy to use format, right up to the category page.

What we did here was, after people interacted, we asked them, "Which of these different choices did you like and how much?" Most people said price was very convenient, to look at price ranges, etc., and if you see -- I don't know if you can see on this slide, but the choice to the far left is to search by lifestyle -- these are things like home office, mobile entertainment, that kind of thing, and, at first, if you look at the ranking, the fewest number of people said this is a compelling navigation choice for me, but when you dig into their individual responses, those few people -- this was about 20% of the people in the test -- those 20% really loved that option. It underscores the point that you have to create compelling navigation, different navigation that's compelling for a lot of different groups of people.

Anne Holland: Sort of serving different search personas, in a way.

Stefan Tornquist: That's exactly right.

Anne Holland: The thing I think is really worth looking into is when was the last time you tested your category page layout and what it looked like, especially for those of you who are spending a heavy amount of money on paid search marketing. A lot of the time, what you're doing is you're paying for category. You're saying give me everything in -- well, it might not be books, but it might be raincoats or women's sweaters -- and if you are really spending in these categories and then sending people back to your handy category page, when was the last time you actually tested the design of that category? Often enough, unfortunately, the category page, the design and all of that was sort of picked during the wire framing session with the Web designer a few years ago, and nobody's really heavily tested the layout of that since.

You should be looking at the number of things you have lined up across, the size of images. Do you have price? Do you not have price? How much text is there? Can they click on text links or just buttons? Are there buttons and links? You ought to do a lot of testing around this category page if you're doing heavy search marketing to it. I think there's a world of response lift that's available there. It's very exciting.

Why don't we move onto slide No. 10. Now, Stefan, can you talk us through this chart?

Stefan Tornquist: These are the various website tests that we polled folks about, and what we're looking at are the positive reviews. Those who said the tests were very effective, those are on the right-hand side. Those who said they weren't very effective are on the left-hand side.

We get a very good graphical representation of the ratio of positive to negative, and it's interesting that tweaks of internal search is clearly the winner. It has a very low negative percentage, very high positive ratio, even doing better than those folks reporting on shopping cart design and function tests, even that sort of premier test that almost 60 percent say this is worth doing. But, 11%, a significant percentage, had missteps in their testing, only internal search, but basically anything you could do to improve internal search had a real impact and quickly.

Anne Holland: That's fantastic. In other words, again, internal search, often powered by actually a vendor, sometimes ... often, you'll have a third-party vendor who powers the internal search on your site and maybe your layout is just dependent on what their default specs are. Go back, look at their default, see if you can test it. See what you can test. Here, of course, are some ideas. In the very next slide, you'll see some of our top ideas based on anecdotal data from case studies of real-life marketers who tested stuff. Here are some of the ways you can test your internal site search landing page to make sure that you get the highest possible conversion rate:

Horizontal versus vertical display, the number of answers displayed. In particular, are you just displaying 10 still? Are you displaying 20? Are you displaying all? I think a lot of people picked the default number they would display way back in the days when everyone was on dial-up and we only had to show a couple of things at a time. Broad matching. That's of course, whether you have -- if someone types in 'umbrella,' will it also include things maybe called Brawley or weather gear or whatever, or is it just going to show anything with the word umbrella? The relevancy of the top answers. Are you going to show the best seller in that category? Are you going to show the lowest-priced item in that category? Are you going to show the thing that was the most relevant to the particular search term?

There's all sorts of different ways of deciding what item goes first. It's pretty complex. The size of the image. If you have images in search, and hopefully you do, how big are they? How small are they? Are they clickable? And, of course, what tops of copy and hotlinks, including pricing, do you show right there on that search page? These are all tests that you can conduct.

Now, the first one actually is the one a lot of people ask me when I say horizontal versus vertical display. They say, "What do you mean by that?" Here, we've shown you, as our absolute favorite test of all time for internal search landing pages, we've actually put together a couple of slides to show you what the difference is.

The very first slide, on page 12, is what we would call a horizontal display. We call it that because you see a list of images or search results in a horizontal row and we've got them from two different very famous websites, and here they are, displaying three things in a row. This seems to be the default in particular for apparel sites. I'm not sure why, but there you go.

And then, of course, the vertical display and that would be everything sort of in a list and, of course, you see here two websites, and they're very, very, very famous websites, Amazon and eBay, that have chosen to go with the vertical display for most search results. Now, it's interesting. I think a lot of people went with horizontal display initially because we were all so terrified of the fold.

We all thought, "Gee, I'd better get as many results as I can up above the fold, so I sure as heck won't make a list because that would require scrolling." But, the fact that these two very big, very famous sites that we know have metrics out the whazoo, are actually deciding that they're going to go with a list means something, means at least horizontal versus vertical is worth testing for your site. I'm deeply interested in this, and we're going to certainly be watching data on that for the rest of the year.

Well, thank you so much for joining us today on our ecommerce and search marketing presentation. All of the data that we mentioned today came from the new MarketingSherpa Ecommerce Benchmark Guide. You can get your own copy online. There's a little hotlink for you if you'd like to get a copy or see what it's all about. If you have any questions about what it's about, no problem. Just e-mail us at feedback(at)marketingsherpa(dot)com, and our research team will get right on it and let you know whether we have data on anything. Chances are, if we don't have the data, we know where to find it. So, if you have any questions about data, please let us know.

Thank you so much and thank you for joining me here today, Stefan.

Stefan Tornquist: Thank you.

Anne Holland: Good afternoon.