Yesterday I got confirmation from one of delicious’s engineers that the recently released new site was indeed built on Symfony. That’s great news for the framework, which was already used to create Yahoo! Bookmarks, proving it’s enterprise-ready and able to scale up to millions of active users. Watch the delicious blog for a promised update on “what we learned and how we made certain decisions”.

The search giant’s argument is that too many people are buying links on websites in order to artificially boost their PageRank, a factor in determining search result ranking.

If you don’t know why buying links affects Google, read Chris Beasley’s Search Engine Optimization Guide. It’s truly the best SEO resource on the web. All of the facts, none of the myths, based on tried and tested experience.

A poor quality site could potentially rank higher in search results than higher quality sites by buying many links from high-PageRank sites. When Google’s search results don’t bring you the best sites first, they’re less useful to searchers, and those searchers are more likely to consider alternative search engines like Live, Yahoo! or Ask. I recommend you try them all yourself anyway, as they’ve all been working extremely hard over the past year to compete. Especially Live and Yahoo! have made huge leaps forward. Their indexes are fresher, their algorithms smarter, and their presentation often gets you what you want faster than Google’s bland result pages.

To get back on topic, Google needs to know the difference between a site linking out because it found the destination site useful to its visitors — an editorial vote for the site as PageRank is based on — versus a purchased link there because an advertiser paid for it. They’ve tried to do so algorithmically, and they’re pretty good at it. A link among a cluster at the top or bottom of a page, or near text like “sponsors” is worth less than one occurring amidst a paragraph of text in an article. That’s great. That’s what they’re paying all those Ph.D.’s to do — make Google smart enough to rank the web no matter what any site tries to do to manipulate the rankings.

But somehow they feel they’re not identifying enough this way, and took it further, much further. Official word now is that if Google knows you’re selling text advertisements and that the buyers are purchasing them in order to benefit from your PageRank, that Google will penalize the seller’s site. Not penalize the links, but penalize the entire website selling them. If it’s a high-PageRank site, they’ll knock down your PR. If it’s not, they’ll directly hurt your search result rankings. Either way the effect is to make your site appear less, if at all, in search results, reducing your traffic from Google. Perhaps if you stop selling ads and swear off the practice, you’ll get the penalties removed.

This is where a whole bunch of people cheer “it’s about time!”. This is good for Google users, right? It’s their search engine, they can dictate the rules however they like, and something that hurts those manipulating results by buying links is a good thing. But I wholeheartedly disagree.

The premise that this improves the search results isn’t entirely true. Sites selling text links aren’t somehow less useful because they support themselves on that type of advertising. They’re not somehow a less perfect match to what a Google searcher typed into the search box. They’re not somehow less worthy of the votes they got from other sites that gave them their result positions in the first place.

If at issue were only off-topic, clearly useless links that have no value to users of the linking site, I wouldn’t have a problem with this. It’d still be a further reaching “rule” than I’d like, but I could accept that Google wants to devalue sites that are devaluing themselves by mucking up their content with off-topic links. Those are less useful sites and deserve lower rankings. But this is about all link advertising. If a site only accepts advertisers whose sites complement their own, sites they wouldn’t be ashamed to vouch for anyway, then there’s no reason to penalize that seller.

The amazingly informative and useful web we have today, which contains so much more in quantity and quality than the web of a decade ago, exists largely because there is money to be made here. Commerce makes the internet go ’round just like the rest of the world. Content producers and service providers create these websites because there’s a potential to make money. And text advertising, even when advertisers specifically want text links without “nofollow” attributes, support tens of thousands of these sites.

Google’s success is built on the back of the high quality of the websites it indexes. If doing a Google search didn’t bring you to a website with what you desired, there would be no reason to search. These ad-supported websites are why Google has a market cap of $190 billion. By penalizing sites for selling text advertising, Google is damaging their own search results. They’re removing good results from the list instead of the bad ones — the buyers that ranked artificially high. This is wholly unlike anything Google’s ever done.

Let’s not forget that Google is the largest seller of text link advertising in the world themselves. Buying text ads from Google is perfectly alright, but anyone else selling them without the “ignore me please Google flag” attached to each link is someone not worth indexing. It’s almost monopolistic to control so much of the advertising landscape, so much of the search traffic, and to use bullying like this with their search dominance to discourage advertising outside of their own service.

I think what Google has said they’re doing is wrong. It’s wrong to penalize good sites that make money selling on-topic text ads that happen to not be Google AdSense ads. It’s wrong to screw up their own search results to make that happen. It’s wrong to apply the penalties manually, by people making judgement calls on a tiny subset of the web. The only thing that would make me less disgusted with Google on this issue is if they were to backpedal and explain they are only concerned with the sites that sell completely off-topic links that have no value to users and only serve to manipulate search results.

It’s completely free and smartly monetized through mutually beneficial offers from the site’s partners — telling you, based on your actual spending, how you can save money. For example, it’ll tell you what banks offer higher interest rates for your spending. It’ll tell you what credit cards have better rewards or lower interest than yours. It erroneously told me how to save $80 per month on my internet service by switching to Verizon. Unfortunately those regular $100 payments to Yahoo! weren’t ISP payments like it assumed but Yahoo! Search Marketing deposits.

Full, searchable transaction history and spending trends round out a nice feature set. With categorization and labeling of those transactions, it’s easy to see where you spend your money and how that spending changes over time.

I said I really like mint. I’d love it if they added loans. The cash vs. debt balance information is something more people need to be aware of, but without including loans, you don’t get the full picture. I especially need my outstanding student loans in there to really understand my financial situation. I put in a request already just to add my support for the feature… although it’s probably often requested already since it was a category on the contact form.

It includes full breakdowns by client of what tags, selectors, and individual CSS properties are supported. Attention is paid to the new webmail services of the year — specifically Windows Live Mail which will be replacing Hotmail and the new Yahoo! Mail. The bad news on the desktop side is that Outlook 2007′s new e-mail rendering engine is pretty technologically backwards, and makes it nearly impossible to use CSS for even basic layout. The compatability charts are available as a PDF download as well.

In Philadelphia, my blackened windows (from the accumulation of truck exhaust) overlook traffic on pothole-filled roadways of drivers barely missing eachother to get from one all-concrete part of the city to another. Here, I can look out the window and see a stream that runs to Lake Washington around the corner, and drive around Lake Sammamish on the way to the office each day. I can deal with the morning drizzle every other day for that. Even if I stay in Pennsylvania after I graduate, it won’t be in Philadelphia.

Getting Acquainted

My first day and a half was spent at New Employee Orientation. Those are held every single week; there were about 80 people in the group with me. Microsoft has been on a hiring spree for the past two years, picking up about 16,000 employees last year as it grew to nearly 80,000 in total. Orientation was a bore — a lot of CYA stuff for Microsoft — things like document retention policies, when to consult the legal department, what the org structure looks like, etc. I won’t go into details on NEO since it’s a bore. There were two hours covering health benefits and stock options I don’t get — the only difference between intern and FTE — during which I hopped on an MS shuttle to the company store. Everything there is heavily discounted for employees. Vista Ultimate would cost me $35 instead of $275, and Xbox 360 games just $10, for example. I’ll take advantage of that soon.

On the second day I made my way from the main corporate campus in Redmond to where I’ll be working in Issaquah, what they call the Sammamish campus since it’s next to Lake Sammamish. It’s an extra 10 minute drive compared to going to Redmond but traffic isn’t too bad and the scenery is nice. Compared to the commute I’d have from Philadelphia to Malvern for Siemens or Unisys or a lot of the big Drexel hirers out there, it’s nothing.

I got a quick tour of the building — just a few floors tall with lots of open spaces and open doors — Microsoft tries to promote open collaboration and a college campus feel. I met my immediate boss, his boss, the developers I’d be working with, and the very energetic female “admin” who apparently makes the real decisions for the department as well as providing free candy and a massage chair to anyone that wants the room. Speaking of free food — what we’ve all heard is true — there are big cases of free soda on every floor of every building. Developers really do live on Mountain Dew.

The New Day to Day

My official position is Application Developer for the Account Planning team, part of the Enterprise People & Groups division, which in the big picture is part of Microsoft IT. I’m not working on Windows or Office or XBOX so you won’t see the software I build on your desktop, but several thousand Microsoft employees will. It’s a small team, with my boss across the hall, one of the few other developers in the same office as me, and the program manager right around the corner.

One of the first things I asked my boss about was what hours everyone was putting in. It’s a salaried job, so there is no clocking in or time sheets to fill out, but most people work roughly 9-5. Supposedly, if I can get my job done in 20 hours, I can work 20 hours a week. If I can get it done working from 3PM to 11PM every day, I can work those hours. If I want to work from home, I can head over to Redmond and get a smart card and card reader for a laptop to connect to the corporate network from anywhere. I have a “blue badge” which provides 24 hour access to all buildings in Washington, with other colors given to vendors and partners and other types of employees with restricted access, so my office is always open to me.

There are very few desktops around that aren’t running servers of some sort — development sandboxes, test servers, pre-production environments, etc. Most of the people here are developers, and developers get laptops. They’re pretty beefy Toshiba Tecra tablet PCs with more than enough memory to compile large apps on. Even with all the collaboration tools Microsoft makes, there are constantly people walking down halls with running laptops in hand and bringing them into meetings. Along with the laptop, there’s a docking station, 17″ LCD and Microsoft Laser Desktop 6000 wireless keyboard and mouse. Not as fancy as the Wireless Entertainment Desktop 7000 I have at home, but probably a better keyboard to use for coding anyway. With the laptop docked I spread my desktop across the two screens. This is already a much better environment for this kind of work than DuPont or Math Forum provided.

There’s no micromanagement here. I’ve had one-on-one meetings with a couple levels of managers up from me already and all of them say the same thing. They don’t micromanage, they just get projects going, set milestones, and expect employees to meet them. Nobody tells me what to do each day and rarely does my boss come in to see what I’m doing.

There’s more information than I could ever need available. Everyone’s doors are open, literally, so I have no problem talking to the right person for some guidance or advice. Beyond that I’ve got all of MSDN, tons of online training, classes I can register for, and a huge physical library of paper books I can borrow.

Microsoft Immersion

Why did I take this job? why did I choose to go through all the hassle of moving to Washington for half a year instead of taking a nice management job at DuPont just minutes from my apartment in Philadelphia? Aside from the coolness factor of working for a recognizable name, it was the only route I could find to getting experience in the Microsoft environment.

Drexel University helps students find jobs to fulfill coop requirements by finding employers willing to hire students for 6 month periods and listing the available jobs in a database. When I searched through there each year, three times now, there were at least 20 positions for .NET developers of some flavor. All these positions required previous experience with Windows servers, Visual Studio, some .NET language and usually Microsoft SQL Server. While I’m confident I have enough experience to pick up any of these technologies quickly and become productive in just a day or two, I’ll never get an interview for those jobs without any of the experience showing on my resume.

Microsoft came to the rescue. They sent recruiters to Drexel looking for experienced programmers but they weren’t looking specifically for experience with their technology. They were the first company I ran into developing in the Windows environment that saw my resume and gave me an interview. Those of course went well, since I’m here now, and I’m getting what I wanted. I’m being completely immersed in the Microsoft environment.

• I’m running Windows Vista Enterprise.
• I’m developing C# .NET code in Visual Studio 2005.
• I installed that with Vista’s ability to install programs over the network where every Microsoft product is available to me over the intranet.
• Every intranet website knows who I am thanks to Active Directory. I check my mail with Outlook which shows me the calendar and Communicator status of every person I come in contact with.
• Quick team chats are held with Office Communicator which also links up with IT’s bridges to AIM, Messenger and Yahoo!’s chat networks.
• I connect to Windows 2003 servers with Windows Remote Desktop.
• Before I started, I was given directions to the office with a link to Live.com Maps.
• Team documents are shared on a Sharepoint server.
• I access my e-mail from home with Exchange 2007′s Outlook Web Access web client.
• I have access to all of Microsoft’s current beta software — employees use beta software for real world testing before release.
• I’m an administrator for two Microsoft SQL Server instances and a Microsoft Dynamics CRM instance.
• I made modifications to a program written on top of Microsoft Office Groove.
• My default search engine is Live.com.

If that weren’t enough, I figured this was a good time to replace my old Creative Zen Xtra mp3 player with a Microsoft Zune. I like to load it with episodes of Battlestar Galactica to watch in one of the 26 cafeterias here. I’ve joined “The Social” – I can see the other Zune owners in the building with wifi enabled and what they’re listening to.

It’s a good thing

I’m having a good time. I like my job. I’m driving somewhere new pretty much every day. My mailbox is filled with messages from all the groups I’ve joined already — from Zune owners to Warcraft players to keeping in touch with the other interns and alumni from Drexel. I think this will be a very good 6 months.

Hi Daniel,

We would like to thank you for taking the time to interview with us for an Engineering Internship position. After carefully reviewing your experience and qualifications, we have determined that we do not have a position available which is a strong match at this time.

Thanks again for considering Google. We wish you well in your endeavors and hope you might consider us again in the future.

Sincerely,

{removed}
Google~Staffing Specialist
1600 Amphitheater Pkwy.
Mountain View, CA 94043

How do I feel about that? No big deal; I sent my resume to Yahoo! a couple minutes after getting that mail. I still have an interview with Microsoft in a week and a half, and will talk to DuPont again around then as well. Worst case, I finish off the work experience requirement for my degree at DuPont, which was actually a pretty great job, and only a 10 minute drive from where I live. I’m pretty sure they’d hire me back considering I was offered a management position before my last coop there ended.

The idea behind PageRank is that if a page is important, other webpages will link to it. The more important the linking page itself, the more weight that link would have. This system works amazingly well, instantly bringing up relevant pages for almost any search, without giving benefit to “keyword stuffing” and other attempts to trick the search engine into ranking a page higher. It works so well that Google’s no longer the only search engine taking into account links as a major part of their algorithm. All the major players now do so too.

That’s why link building is a key component to any search optimization strategy. However, most webmasters go about it the wrong way. Google and the other search engines need to provide relevant results to searches to stay in business, and it’s a very profitable business they’re willing to put millions into protecting. Google, Yahoo! and Microsoft have the brightest minds in the computer industry working for them, thinking up ways to keep the results relevant and filter out the pages that aren’t. Especially in 2006, Google and Yahoo! have put a lot of work into improving their algorithms to reduce the weight of non-earned links.

Link Building the Wrong Way

That’s why link swapping, mass article submission, purchasing links and even newer techniques such as link baiting will never be good long-term strategies. Initially, websites exchanged links with other websites that were complimentary to provide a useful resource for their own visitors. Nowadays, many webmasters will swap links with any site that will agree, or any automated link swap page. These reciprocal links are easily detected by the search algorithms and discounted as such. Trading links doesn’t show the same vote for the quality of your site as a link that isn’t reciprocated.

As the more advanced search optimizers realized their link exchanges weren’t benefiting them as much, they started purchasing links on high PageRank sites. How could a search engine know a link was purchased and not placed naturally? Not with too much difficulty, actually. There are a couple tell-tale signs that could be used to discount links that are likely being sold: the same links appearing on every page of the site, links appearing in a list near the top or bottom of the page without any text around them, links that change every couple weeks or months, links near words like “sponsors” or “advertisers”. There’s also the easiest method of all — finding the site listed on a site like Text Link Ads.

Google’s Matt Cutts commented back in 2005:

“But can’t you just not count the bad links? On the dailycal.org, I see the words ‘Sponsored Resources’. Can’t search engines detect paid links?” Yes, Google has a variety of algorithmic methods of detecting such links, and they work pretty well.

Another common piece of advice you’ll find is to write articles about the topic of your website and submit them to thousands of “article directories” where other website owners can choose to republish your work with a link back to your site. That’s another case of easily detectable manipulation of incoming links. The search engines have had the ability to detect and filter duplicate content for years, and this is no different. Your article will be listed, word for word, on dozens of article indexes, on hundreds of automated content scrapers that pick up these articles to build “made-for-adsense” sites, and maybe a handful of actual sites looking for interesting content to publish. In the end, those backlinks to your site from the article will be discounted for being part of duplicate content — not a vote for the quality of your site from another site.

Link exchanges, article writing, and even purchased text links are not worthless investments of time. They’re only not very valuable as tools for a long-term search engine optimization strategy. Swapping links with closely related, complementary sites your visitors may find useful can lead to increased traffic for both sites from actual people clicking those links. Writing articles and hand submitting them to editors of respected websites can lead to traffic from readers of that article and increased reputation as an expert for you and your website. Even purchased links on high traffic websites have value for the visitors they’ll bring your site. They can be part of your marketing strategy, which includes more than search optimization.

Link Building the Right Way

Last month, Google’s webmaster blog wrote:

To sum up, even though improved algorithms have promoted a transition away from paid or exchanged links towards earned organic links, there still seems to be some confusion within the market about what the most effective link strategy is. So when taking advice from your SEO consultant, keep in mind that nowadays search engines reward sweat-of-the-brow work on content that bait natural links given by choice.

So how do you attract those natural links that matter?

1. Present unique or insightful information. If you want other sites to link to you without anything in return, you need to provide something that’s worth linking to. For most sites, that’s going to be unique, insightful content. Don’t just regurgitate what can be found everywhere else — do your own research, add your own opinions, or present information in a new way. Novel organization or presentation can be as important as the information — differentiate your site from others on the same topic.
2. Build something useful. Create a tool, service or web app which people will want to use. Solve a problem and make life easier. If what you build is truly useful, those that use it will share it, both through word of mouth and linking to you from their sites and blogs.
3. Make linking to you easy. Perhaps most importantly, capitalize on those that may want to share your website but are inexperienced or lazy. Provide “link to this page” code, with well-chosen link text already filled in. Provide buttons and banners with your site or service logo to use. Encourage your users to make use of these tools to share your site.

These methods are a lot more hands-off than sending hundreds of e-mails begging for link swaps. They won’t work as quickly as buying links or auto-submitting an article to dozens of article directories. However, they’re the core of the only strategy that has lasted the test of time, and will continue to benefit your site for years even as search algorithms change and filters become smarter.

Some advice straight from Google:

Our general advice is: Always focus on the users and not on search engines when developing your optimization strategy. Ask yourself what creates value for your users. Investing in the quality of your content and thereby earning natural backlinks benefits both the users and drives more qualified traffic to your site.

This post was inspired by my realization that one of my oldest websites, Website Goodies, recently reached #1 for the popular search “website tools” on Yahoo! and #2 for that same search on Google. The site ranks well for many searches, but those are nice broad searches with good volume that I started to target about a year go. I honestly don’t work on this site much, only updating it a couple times a year, but it makes use of the strategies outlined above. For several years now, every article and every tool on the site includes a “Link To This” section at the bottom of the page which pops up copy-and-paste linking code. The “link to this” section on all of the website tools generates code linking to that specific page along with text that includes the words “website tools”:

<a href=”http://www.websitegoodies.com/tools/emailguard.php” title=”E-mail Link Guard”>E-mail Link Guard</a>
- a free website tool from <a href=”http://www.websitegoodies.com”>Website Goodies</a>.

The free tools the site offers are quite popular (more than I would’ve thought when putting them together), and by making it easy for users to link to them using my text, the one-way, earned backlinks to the site from related sites have increased steadily and with them, my search ranks.

For consistency, I only used data for ad campaigns for my TargetedVisitors and VisitorBoost sites since I advertise both on all three search engines. They make up more than 90% of my advertising spend.

Google, by far, provides the bulk of my paid traffic. This is in part because my keyword list is longest there, but also because they have more eyes and higher click-through rates. Take away Google’s content partners, including the AdSense program, and the click-through rates on their search ads alone is 8 times higher than Yahoo!’s.

I can see two main reasons for the much higher click-through rates on Google’s ads versus Yahoo!’s. First is use of the {KeyWord} token which allows me to display the searched phrase within my ad. I make use of it in about half my ad titles while using only the ad text to prequalify the visitor. The second reason is the editorial process at Yahoo!. The requirements for ad text are much stricter there, and often Yahoo! editors will rewrite my titles or text. I often end up with less catchy ad titles that I would expect to attract less clicks.

I currently track my advertising results with AdWatcher, although I’ll be replacing it in 2007 with something home-grown. Every ad I place uses a tracking URL that goes through my copy of AdWatcher, recording the click’s source and setting cookies to track if that click eventually results in a sale.

AdWatcher’s click counts for my advertising are shown above. In all cases, more clicks were recorded through the tracking URLs than the three companies charged me for. Some of these are simply duplicate clicks that they filter automatically. Google AdWords is the only service showing a large discrepancy — I was billed for 18,000 less clicks than I received. I attribute this in large part to fraud across their network, especially the AdSense publisher network. I see “adjustments” on my account crediting me part of my charges every month and their documentation shows that adjustments include refunds for click fraud.

Strictly on a per-click basis, Google was the cheapest search marketing I did this year. The average click cost me just $0.20 even though my max bids are nearly double that. My high click-through rates have resulted in high quality scores, which results in my ads showing in a higher position at a lower click-through rate than I’d pay otherwise. This is one of Google’s best qualities, and something Yahoo! is now only starting to copy with its release of the “Panama” ad ranking system.

Yahoo! Search Marketing, going back to when it was Overture and before that Goto.com has historically had high average CPCs. You’ll find some of the highest bids in the search marketing industry on Yahoo! searches. I’ve managed to pay only a penny per click higher than I do at Google, but do so at the expense of traffic. If I want to buy as many clicks from Yahoo! as I do from Google next year, I’ll have to sacrifice that CPC.

Surprisingly MSN came in at the highest average CPC. As a new service this year, I expected bids to start out lower than at Google or Yahoo!, but when I bid around $0.20 I wasn’t able to receive more than a handful of clicks per week. I had to raise my maximum CPCs significantly to get the couple hundred clicks I did.

The higher CPCs were offset by huge conversion rates with MSN. 13% of those clicks resulted in sales, beating out Google’s 4.3% and Yahoo!’s 2.9% by far. It looks to be well worth the extra effort and extra cost to getting traffic from MSN searches.

All of these conversion rates are lower limits. I receive too many more orders per day than AdWatcher tracks for them to all be sales from natural traffic, which means not all sales from advertising are being linked back to the original ad click. That happens when those visitors clear their cookies before making the purchase or make it from a different PC or browser.

That high conversion rate translated into a very low cost-per-sale for MSN AdCenter advertising. Google comes in second place due to the lowest CPC and average conversion rate. Yahoo! advertising costs me the most for each resulting sale, often resulting in very thin margins. I need to be careful not to over-spend when I try to increase my search traffic from that source.

The following charts I produced mainly for my own benefit and won’t have much meaning for anyone else, as they’re based on my average order size and don’t take into account the costs of filling those orders or running the business.

That wraps up my analysis of the search advertising programs at Google, Yahoo! and MSN for 2006. All three provided a good ROI for the year, with MSN AdCenter showing the most promise for even better results in 2007.