Key to the difference is Powerset’s ability to glean meaning from the sentences. While other search engines primarily look for instances of words on pages, Powerset understands those words. Or something like that. The Search Engine Land article illustrates the concept with a search for Henry VIII. The Powerset results include “factz” based on verbs, such as he “granted” land and “married” a bunch of times.

I was suspicious of the “z”.

But, I figured I’d try it out myself using the tried and true ego search method. If I there’s one search result for which we should be able to judge accuracy, it should be the one about ourself. (Keeping in mind that the current version of my Wikipedia entry is woefully out of date and has been flagged for depressing grammar issues.) So what does Powerset think that Wikipedia has said about me?

That I have declared bankruptcy and received email.

Sigh.

I figured hey, why not start organizing some events for developers about search! So, I did.

Ideally, I’d like to hold these once a month, and bring together experts to review sites from the audience. And have lots of food and drinks. In our inaugural month, we’re holding two events!

Tuesday, May 13th at 6pm
Solo Bar, 200 Roy Street, Seattle
This event is sponsored by Microsoft, and they’ll be providing lots of swag in addition to food and drinks. We’ll chat a bit about search, look at a few sites, then hang out and chat. You can sign up at Upcoming.

Thursday, May 29th at 6pm
Google Seattle office, 651 N 34th St. Seattle
This event is sponsored by Google, and we’ll look at some diagnostic issues sites may encounter while we snack and drink. You can sign up for this event at Upcoming as well.

Wednesday, June 4th
Bell Harbor Convention Center, Seattle
Of course, if you’re looking for more in-depth information about how to build crawable sites, you can check out Developer Day at SMX Advanced on June 4th. We’ve got speakers from the major search engines to talk about the infrastructure details of web applications from a search perspective, Duane Nickull from Adobe to talk about making Adobe technologies search friendly, and web developers to give real-life examples and case studies. We’ll be ending the day with an expert panel to review your site!

Brought To You By Jane and Robot
The free networking events are the first activities organized by a new project I’m working on with Nathan Buggia called Jane and Robot. The idea behind Jane and Robot is to provide definitive content to developers about building web applications for both users and searchers. We’re focusing on the developer audience, rather than search marketers, so we’ll talk more about implementing 301 redirects in PHP than we will about optimizing content for particular keywords. The site is in “soft launch” mode now, but watch as we evolve it.

So far, we’ve got slides up from the SEO for Developers workshop we did at Web 2.0 Expo a few weeks ago (along with diagnostic checklists), as well as an events page where you can watch for more events like the ones we’re putting together in May.

And check out our first article, on domain canonicalization.

That leaves Tuesday, Wednesday, and Thursday for any serious email marketing effort. Some say Tuesday early afternoon is the best time for optimal open rates. People are ready to tackle the drudgery that is the inbox, and your mail is the first thing they see. Others say Wednesday, as perhaps people have conquered the worst of it and feel they deserve a reward such as idle email shopping. Choose either day, but make sure you send early afternoon.

Finally, testing and research have given us definitive answers for something and we never have to worry about it again. We now know not only the ideal days but the ideal time. Hooray!

Except I’ve found a potentially fatal flaw in this plan.

And that is that everyone is now sending marketing mail on Tuesday and Wednesday, early afternoon.

I don’t get much email marketing because I unsubscribe to just about everything as soon as the first piece of mail hits my inbox. Someone who declares email bankruptcy must become ruthless with incoming mail.

And yet there I was last Wednesday at around 1pm, and on came the mail. REI wanted me to know about their May events calendar. Alaska Airlines wanted to make sure I knew I could buy people flowers and earn miles at the same time. Microsoft Office Live Small Business thought I might want to know how to get my business online for free! Choice Hotels has my room ready! The mail just kept coming.

And I realized, all that research was going to have to start over with the addition of a new variable. Not only do marketers have to avoid sending mail when people are off for the weekend, they have to avoid sending mail at the same moment everyone else is sending mail. And so Thursday at 10am will become the new Tuesday at 1pm. At least until everyone adjusts their email schedule. And then it all will start over again.

Of course, rather than look at averages for “the” customer, you could look at the particulars of your customer. I was thinking about this last Wednesday at 1pm when my mail started filling up, but apparently I’m not the only one.

Last night on the plane, I was reading Fast Company and happened upon this article about how Barneys is personalizing mail based on individual behavior on the web site. Targeting mail seems like a much better approach than the old fashioned blast, although I’m not sure about their assertion this rosy new relationship with the customer means that people embrace getting up to five emails a week. They do say they’ve had a ten-fold rise in response rates, which totally makes sense. If you send a promo for hip new purses to your entire email list, you’re percentage of conversion is going to be lower than if you send the purse promo to teenage girls and the power tie promo to older men.

Although Barneys is getting better at segmentation, they seem to be hesitant to go the next step: stop sending mail to people who don’t respond. I have never shopped online at Barneys and haven’t been in a store in at least eight months. But that doesn’t stop them from sending me mail day in and day out. Mail, by the way, that I never open. (I finally opened one last week solely to click the unsubscribe button.) The incessant mail (10 messages during a 12 day period last month) actually made me less likely to shop at Barneys because I was so irritated that they continued to clog up my inbox.

Ryan Warren of Exact Target brought this up today at the eMetrics Industry Insights Day. He said that sometimes the best thing you can do is stop sending mail to people who don’t open it. Spend you energy on those who like getting your mail and take action on it.

His data supported Barneys’ direction. He said that only 11% of companies send targeted mail and only 7% leverage click stream data, but doing so can raise conversion rates from 1.1% to 3.9% (and can raise click through rates from 9.5% to 14%).

He talked about sending mail not on Tuesday afternoon at 1pm but based on when the customer was interacting with the site. For instance, if you have a travel site and someone puts a trip on hold, send them an email to remind them the hold is about to expire. Or if they were checking out a vacation package, let them know when the price drops. Or better yet, if you know they’re in Seattle and they were browsing trips to Mexico, email them when you see that the Seattle weather forecast calls for rain. (Although now that I think about it, you might need to tweak that last one, or you may end up with the Barneys mail sent every day dilemma.)

Domain Roundtable
I’ll be speaking on the SEO experts panel on Saturday about the key things to look at when thinking of developing a portfolio of domains into content sites. Building web sites with content aimed at users can be quite a bit different than managing domains for their potential inherent name value, and my advice will be focused on building long-term value. Even from a purely domain perspective, a site that’s built for long-term value should be easier and more lucrative to sell. (Of course, there are a myriad of other benefits from approaching site building this way as well.)

Web 2.0 Expo
I’ll be speaking at two sessions on Tuesday.

In the morning, I’ll be doing a session with Nathan Buggia in the development track about search-friendly design for web developers. We’ll be talking all about how to build solid infrastructure that takes into account both usability and search engine crawlability. The cool thing is that you can code the site in such a way that you accomplish both goals at once.

In the afternooon, I’ll be speaking with Dave McClure and Hiten Shah on startup metrics. At this session, I’ll be talking about the marketing side of search (rather than the development side that I’ll be talking about in the earlier session), particularly about the search metrics that matter most and how you can make them actionable.

Ignite San Francisco
On Tuesday night, I’ll try the whirlwind that is Ignite. 20 slides in 5 minutes! If you don’t have time for the three hour session Tuesday morning, you can check out the 5 minute version: 5 things developers should know about search. First thing! That you need more than 5 minutes.

Web developers generally make sure their applications are accessible by the major browsers. After all, [Internet Explorer users represent 54% of those online and Firefox users are about 37% of those online], so you’d be cutting out a large chunk of potential customers if your app didn’t work for visitors using those browsers. EDITED: This number is obviously reflective of one data point, but as someone in the comment points out, it can vary. For instance, Wikipedia has IE at about 75% share. My visitors use IE just over 50% of the time and Firefox nearly 40%.

But what about that 95% who are searching for you? Doesn’t it make sense to make sure your application is accessible via search as well?

Search engine optimization (SEO) is sometimes seen in a bad light, filled with spamming and deception and trickery. But what’s sometimes called “SEO” is more accurately called “spamming”. True SEO is much more akin to making sure your pages render on Firefox and when done well, increases not only findability, but accessibility and usability as well. Pages designed using SEO principles render well on mobile devices, screen readers, for those with slow connections who turn off images, for novice users who are still on older browsers, and for savvy users who block Flash and javascript.

If you’re developing web applications and you want to make sure your code is well-optimized for search, where do you start? You can check out these resources:

• Nathan Buggia’s MIX08 Presentation: Advanced SEO For Developers
• SEO and asp.net
• Colin Cochrane’s asp.net/SEO posts
• A List Apart: Findability

A Conference for Web Application Developers: All About SEO
For a comprehensive deep dive into infrastructure issues and solutions for both the Microsoft and LAMP stacks, diagnostic tools and checklists, and practical tips for building web apps that searchers can easily find, check out the Search Marketing Expo Developer Day conference, happening June 4th in Seattle. We’ll end the day with in-depth, technical site reviews that bring together everything presented throughout the day.

Interested in speaking? Submit your pitch now!

If you are a web developer, I’d love to hear about the issues you are most interested in learning more about. Let me know in the comments!

I also thought I’d ask the Twitter audience what tools (specifically Firefox plugins) people like the most to see what I might be missing. Overwhelmingly, the favorite was Firebug. I have to agree with the crowd; Firebug is probably my favorite too. But people recommended lots of other tools, some of which I use all the time and some I’d never heard of. Here’s the list for your debugging pleasure.

• Firebug: everyone’s favorite plugin. Sure, it’s a much better way to view the source code of a page, debugs javascript, and lets you see CSS. But it also lets you tweak code and see how it would look rendered on the page, and it shows you exactly where code is located on a rendered page, and it helps you from going crazy trying to figure out why things aren’t lining up, and it… Well, you get the idea. My favorite use of it lately is to search the code for a URL, then inspect the element to find that link on the page. Sometimes those links are sneaky and Firebug makes them super easy to find.
• Web Developer Toolbar: another plugin I use all the time. It adds a cool toolbar to the browser that lets you easily disable things like javascript, images, and CSS, view meta data, see a page in a different resolution, validate stuff… and I have barely started. One thing to note is that when you disable CSS and then images, CSS becomes enabled again (this is just a bug in the toolbar). So, you have to disable CSS after you disable images.
• User Agent Switcher: This is useful to see if a page is cloaking, but obviously, only if it’s cloaking at the user-agent level (vs. for instance, by IP address). Simply add the user agents you want to check out (for instance Googlebot, MSNbot, or Slurp) and then select that user agent and reload the page. The most well-known example of how this works is the nike.com site. Here’s how the home page looks when the user agent is set to default (in this case, Firefox):

  

  And here’s how that same page looks when the user agent is set to Googlebot:

  

  (As an aside: note that for whatever reason, the Flash page isn’t loading for me. Yet another reason why I just don’t buy the argument that all Flash sites are good for users and that webmasters have to jump through these cloaking hoops to make sure search engines index the pages. If I’m a customer and I want to buy some shoes, I don’t want to have to debug why your site isn’t loading for me. (Not to mention if I’m at the mall and want to check out your shoes on my mobile phone…) People want non-Flash links too!

• Live HTTP Headers: view the HTTP response of a page. For instance, is the page returning a 200 response when it should be returning a 404? Is the redirect a 301 or 302? And speaking of redirects, how many hops does the server take you on before you land on the destination? And that’s just the kinds of things you can get from the status code. There’s all kinds of value to be found in header information. Is the page setting a cookie? Is it sending data compressed? What really is being downloaded?
• Header Spy: Similar to Live HTTP Headers. I haven’t used this plugin, since Live HTTP Headers works pretty well for me.
• Yslow: indeed, why slow? This is a great example of a tool that not only provides data, but shows you how to make that data actionable.
• Flashblock: replaces Flash with a play icon. I think this is great because the page loads without Flash, but you can view it if you want to.
• Aardvark: view page elements, remove elements from pages, do lots of analysis.
• ColorZilla: I hadn’t heard of this plugin, but it sounds pretty cool. You can easily grab any color that you see on a page, as well as lots of other little things.
• HTML Validator: exactly as it sounds.
• SEO For Firefox: This is Aaron Wall’s plugin and he has a video on his site describing it. This plugin adds a bunch of links under the Google search results. You can click on a link to find out more information about that result. For instance, below is the first result for “vanessa” (this site). I’ve clicked the first few links to get more information. The rest show what the page looks like by default for each result (a question mark appears until you click it to retrieve the information). (As an aside, I don’t know that all of these factors influence rankings, but it’s still useful data to have and nice to have it all at a glance. It’s also handy to be able to click the “whois” link next to any result.)

  

  This plugin also highlights nofollow links on a page.

• SEO Quake: also has a toolbar and search results parameters that are customizable.
• Search Status: this plugin provides a lot of the same information as SEO for Firefox, as well as things like robots.txt and keyword density.
• SEO Link Analysis: provides lots of things beyond links including anchor text (helpful) and PageRank (perhaps not as helpful). I’m having trouble getting it installed though.
• View Dependencies: shows you a list of all the dependent files and lets you open them to take a closer look.
• Meta Info Sidebar: shows a lot of seo-related and meta data in a sidebar.
• SEOpen: I haven’t used this, but it looks like it provides many of the same details as some of the other plugins. This list was in part recommended by Twitterers, which means that they all had preferences about which SEO tool to use. I think it’s all about finding the one that you like best.
• View Rendered Source: A much easier-to-read version of view source. Since I use Firebug for viewing the source of a page, I haven’t found a need for this, although it looks like a pretty readable way to look at page source code.
• ShowIP: see the IP of a page, and query info such as whois.
• Advanced Dork: easy access to Google’s advanced operators.

And if you want to know even more about Firefox plugins, cshel blogged her list back in January.

I’m forming a company around these ideas (although I’m not ready to talk about the details just yet) and am looking for just the right people to join me. If you like getting in on the ground floor of cool new stuff, have a developer background or online marketing background (beyond SEO), and want to talk more, let me know!

And if you’re working for a company that would like to get a new perspective on online marketing strategy, or simply talk through your issues and needs, ping me on that too.

You can reach me at vanessafox at gmail dot com.

Fit neither of those camps but just want to know what the heck we’re up to? Watch this space.

Open in a New Tab?
The most irritating thing, which is surely a bug, is this. You click on the spreadsheet from the docs list. It opens in another tab. Then from that tab, you go to another page. Perhaps you start to compose a blog post. Then you switch back over to the docs list and click the spreadsheet again to open it. It opens in the tab where it was previous open, rather than a new tab. Perhaps this is the tab that contains your newly composed blog post. Or did contain before that content was replaced with the spreadsheet, wiping out all of your writing. Ahem.

Anonymous Results
Another thing that I suppose makes senses but isn’t clear up front is that you can send the survey to a list of email addresses, but the survey results are anonymous. So my question asking if the respondent is willing to do a follow up survey? Not all that useful to me. I’ve since added a new field asking for that email address.

Limited Question Types
Two things I really needed in my survey but that Google Docs didn’t have were:

• An ordered list (rank a list in order of importance)
• An “other” choice with a text field (so respondents could write in additional answers)

No Custom Email Messages
You can send the form out via email, but any text you write ends up as part of the form. You can’t send a separate email message as you can with other types of Google docs.

Overall all though, the survey was easy to put together and it’s easy to see the responses. I can’t seem to manipulate things much in Google Docs, so I’ll likely have to export to Excel. It would be nice, for instance, if my single-choice answers had graphs or pie charts displayed by default that showed distribution. Instead, I just have a non-compiled laundry list of answers. You can now add Gadgets to spreadsheets, and likely one of these does what I want, but I’d like to have some of that functionality built in. Not that I’m saying I’m lazy.

If you’d like to take my survey about how you approach and measure your online marketing activities, you can check it out here. All results are anonymous. Unless you fill out the newly added email question. Which, like the rest of the survey questions, is entirely optional.

I find that the best way to start any online marketing effort, link building included, is assessing where things currently are.

What links do you really have?
You can use Yahoo! Site Explorer and Google Webmaster Tools to get a general idea of the number of links to your site and where they’re coming from. I like that Site Explorer (supposedly) ranks the links in order of importance, and I like that Google lets you download the links so you can open them in Excel. Yahoo! shows more of a full count, but only lists 1,000. Google may not have the complete list of links, but lets you see (and download) the full list they do show.

Note that Excel 2007 lets you load up to a million rows, but earlier versions only let you see up to 65,000 (ish), so if you have more than 65,000 external links and want to view them in Excel, you might want to spring for the upgrade.

In Site Explorer, make sure you change the Inlinks options to “Except from this domain” and “to Entire site” to get an accurate picture of your external links.

From the list, filter out the links that search engines could potentially be devaluing or discounting entirely. This includes links that are paid or the result of link networks or exchanges, links that generate no traffic to your site, other sites that you own, and spammy directories. Then consolidate multiple links from a single site. What’s left? For some sites, things can initially look pretty good, but once you filter and consolidate, you find that you’re left with much fewer links than you had thought.

(I find this exercise is sometimes useful when people ask me why a competitor site ranks above them when the competitor site has fewer links. There are lots of reasons this could be, of course, but one of them might be that once you subtract the number of potentially discounted links from the total, you end up with fewer than the competition.)

Your Most Compelling Content And Your Target Audience
From the remaining list, what types of sites are linking to you and what types of content are they linking to most often? Easy ways to get links are to:

• create more content like that which is already being linked to (people are clearly interested in that kind of content).
• target other sites in the same categories that link to you already (those types of sites have audiences who are interested in your content).

Your Biggest Fans
Make a list of who is linking to you that would link again if you had something new to talk about (like bloggers and reporters). Without doing link analysis, you may have no idea who’s out there talking about you! Make sure to ping them about new features and content.

What links bring the most traffic?
You can get this information from your web analytics. Determining what links bring more visitors can help you with audience analysis. What sites have audiences that are most interested in your content? Don’t just look at page views, look at bounce rates, time on site, and number of pages viewed. What audiences are most actively engaged?

What social media sites bring the best traffic? Again, those sites are more likely to cater to your target audience.

Looking at my referrals, for instance, the StumbleUpon audience seems to be a good one for this site. They have a super low bounce rate of 37%, stay on the site for over a minute, and look at nearly two pages while they’re here. Compare that to the Reddit audience, who have a bounce rate of 92% and spend only 13 seconds on the site.

Who’s linking to your competition and why?
Use Yahoo! Site Explorer to check out your competitors’ links. What kinds of content of theirs is linked to most often? Maybe your site doesn’t fill a need that theirs does. What kinds of sites are linking to them and not to you? Those are audiences interested in your topic who may not yet know you exist.

Anchor text: What are your external links saying about you?
What does your anchor text look like? (You can get an anchor text report from Google Webmaster Tools, as well as from third-party tools.) Search engines use a combination of on page and off page factors to determine what your site is about.

In the next segment of this series, I’ll talk about ways to influence external anchor text to your site, but during the assessment phase, just make note of what the anchor text looks like. If you don’t have any anchor text for keywords that you care about, that may partially explain why you’re not ranking the way you’d like to, and even why those external links aren’t bringing the traffic you’d like. Ideally, the anchor text compels people to click the link and visit your site.

As the result of this assessment, you should have the following lists for the next phase in the link building process:

• Types of sites that tend to link to you
• Types of content on your site that is linked to most often
• Reporters, bloggers, and others who seem interested in your content, your competition, or your industry/topic
• Types of sites that tend to link to your competition
• Types of content that the competition provides that is linked to most often
• Anchor text and where it’s coming from

Yep, link building is a long and arduous process. But if you’re building links for long-term value, it’s well worth it. The next post in this series will be about preparing your site. Stay tuned!

Apparently, it’s the Leeson Inn.

(In case you see different search results than I do, 8 of the top 10 results are for that hotel.)

I can’t imagine this is the result of search engine manipulation, since all kinds of sites are showing up, including Yahoo! and Trip Advisor, which as far as I know, are separately owned. (Although in this age of user-contributed content, I suppose we could be seeing a new era of review manipulation for SEO purposes, as Larry points out in the comments.)

Is the Leeson Inn really the definitive answer to my question?