Microsoft and Information Retrieval | Science for SEO

Tracing the tragetory of Bing from the path of Microsoft research.

Posted via web from Charles Thrasher's Lifestream



Social Bookmarking

Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

July 2, 2009 | Permalink | Comments (2) | TrackBack (0)

What's the value of an agency to a search engine?

I work for Microsoft as project manager in search marketing for SMB (small, medium business) clients. One of my projects is to build a value model for agencies representing SMB advertisers. It's a lot harder than you might think.
 
The basic formula is, well, basic -- revenue less cost to serve equals gross margin. But what about the intangible values? What about an agency's strategic influence on the market or the lost opportunity represented by prospects available only through agencies? On the flip side, how does the addition of an agency complicate the conversation between search engine and advertiser?
 
These questions become more cogent in the SMB space where margins are thin and efficiency a prerequisite for profitability. In SMB, processes need to be scalable and lean. It's no coincidence that I've chosen this project for my Six Sigma Greenbelt certification.
 
Over the next few weeks I'll explore the agency value proposition for search engines. This is a hotly debated topic and not just at Microsoft. I'd be interested in hearing the value prop stated from the agency side, too.

Posted via email from Charles Thrasher's Lifestream



Social Bookmarking

Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

July 2, 2009 | Permalink | Comments (0) | TrackBack (0)

Microsoft JobsBlog

Posted via web from Charles Thrasher's Lifestream



Social Bookmarking

Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

July 2, 2009 | Permalink | Comments (0) | TrackBack (0)

Finding a job at Microsoft

My nephew recently asked if I knew of job opportunities with Microsoft in Southern California. He's a Marine recently wounded in Iraq by a roadside bomb. He's recovering; others in his unit weren't so lucky. He'll soon muster out of the Marines at Camp Pendleton and looking toward a career in networking. Now, I know next to nothing about networking. Years ago I didn't finish a class full of star and ring diagrams. That's the extent of knowledge. But I wanted to offer something of value. He deserves it.
 
So, I thought I'd document what I discover about looking for a job at Microsoft, resources available from outside the company and my own personal perspective from the inside. With the usual disclaimers, of course. My opinion is my own. It's worth what you pay for it. And my horizon is limited. I work as a project manager for Microsoft's search marketing clients. It's a far cry from engineering.
 
So, my first bit of free advice. Subscribe to the RSS feed for Microsoft's JobsBlog. Microsoft is a big company. Like any big company, it has a corporate culture. The JobsBlog offers advice on that culture. As well, hunting for a job across the wide landscape of Microsoft, building your resume, interviewing and working at Microsoft.
 
And my second bit. Don't get discouraged quickly. Microsoft is a great company but it's big and complex. Finding your way takes time.

Posted via email from Charles Thrasher's Lifestream



Social Bookmarking

Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

July 2, 2009 | Permalink | Comments (0) | TrackBack (0)

The Dirty Bits & Bing

The reason for the cry of alarm raised by the righteously indignant about Bing’s inline preview of pornographic video escapes me. Of course, pornographic video is bad. Categorically, irredeemably bad. Which begs the question why unambiguously pornographic searches are among the most popular on search engines. A whiff of hypocrisy lingers in the air rather like the smell of ozone after an electrical storm.

But the children. Certainly pornography is inappropriate for children? I agree that children are an inappropriate audience for pornography. It certainly stunted my growth. But why have the children disabled Bing’s Safe Search protections in order to preview porn on the SERP? (And what were they searching for in the first place?)

It’s a rhetorical question. We all know the reason why. It’s the same reason children lie when required to confirm their age on actual porn sites.

My point? Safeguards imposed by web publishers on adult content are ineffective. Other than parents fostering proper values in their children, there is no effective prevention short of removing adult content from the web entirely or secreting it behind paid admission. Given the traffic and the revenue generated by porn publishers, I don’t expect adult publishers to strike their tents and steal away anytime soon.

Although the dirty bits seem to have captured much of the media’s attention, the inline preview feature of Bing’s video search is impressive even when the content isn’t salacious. Consistent with Bing’s goal of simplifying search, it offers an efficient means of identifying content that’s most relevant to your search query.

And frankly, it’s a lot of fun.



Social Bookmarking

Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

June 11, 2009 in Search Engines | Permalink | Comments (0) | TrackBack (0)

Bing, SEO and Internal Search

 Bing_Logo One of the interesting features of Bing, Microsoft’s new search engine, is its employment of the wisdom of crowds. When searchers consistently express satisfaction with a specific result for a query, there’s a high probability it’s the right result. What the measure of customer satisfaction in the context of a SERP (Search Engine Return Page)? A bounce.

Bounce rate is a measure of the signal to noise ratio of a SERP. A high bounce rate indicates more noise and lack of relevance. A low bounce rate more signal or information. The higher the information value of a particular link for a particular query, the higher Bing ranks the link on the SERP until at some mysterious point it becomes the “Best Match.”

Bing_BestMatch

Above is an example of best match on a query for Microsoft. Achieving best match status has several advantages but the one that first caught my attention was the inline search using the site’s own internal search function.

If a site warrants best match for a query, Bind sends a robot to crawl the site and determine if there’s a simple search interface. If it finds one, it incorporates the interface into the SERP. It’s an interesting innovation that extends the value of search by making it more efficient, or at least as efficient as your internal search function.

Having good internal search isn’t an option. For some people it’s their preferred method of navigation. And now it’s a doorway to your site, an even greater impetus to perfect the relevance of your site’s search function.

Given the sad state of search on most name brand sites, there’s a real risk of being caught with your pants down in public.



Social Bookmarking

Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

June 3, 2009 in SEO | Permalink | Comments (4) | TrackBack (0)

Actionable Web Analytics

A Review

ActionableWebAnalytics Actionable Web Analytics is one of the most useful books I own about web analytics, and I own quite a few. It isn’t an introduction to web analytics — for that I recommend Avinash Kaushik’s Web Analytics: An Hour a Day — but for practical application in a business environment large or small, this is a book you need to own. I may be needier than most; I own two copies. The chapter “Monetizing Site Behaviors” alone is worth the price of the book.

Corporations Large & Small

I’ve worked for small companies most of my life, some vanishingly small. Now I work for Microsoft. I’m more impressed by the similarities between companies large and small than their differences. Whatever their size, most companies have yet to fully exploit one of the most significant advantages offered by web analytics — prioritizing their development projects.

Monetizing Site Behaviors

Monetization is simply assigning a dollar value upon desirable site behaviors — the things you want people to do on your site. It doesn’t mean you’ll earn that amount if someone does what you want them to do; it’s an approximation of that behavior’s value in achieving your goal. How you assign a value to a behavior is less important than consistency in your method. (Burby and Atkinson do provide details even if I don’t.) The ultimate value of a monetization strategy is guidance where to invest your development resources.

Everybody’s hard pressed to choose what to prioritize for their next web development project. Investment is risk. Usually it comes down to the HiPPO’s decision — the Highest Paid Person in the Organization. Monetization is a method of managing risk by determining which projects are likely to pay higher returns on investment. It takes the decision out of the hands of the HiPPO and bases it upon data. Making data-driven decision is what web analytics is about. And, obviously, actionable web analytics is what this book is about.

Actionable Web Analytics
Jason Burby & Shane Atchison
Wiley Publishing, 2007
ISBN 978-8-470-12474-1

Reblog this post [with Zemanta]


Social Bookmarking

Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

March 8, 2009 in Analytics | Permalink | Comments (3) | TrackBack (0)

Web Analytics and Monetization: Prioritizing Change

You have a web analytics tool. You’re generating cubes of data, reams of reports. There are so many things you could change on your site, so many tests you could run that you’re paralyzed by indecision, a deer in the headlights. More likely, it’s your boss who’s paralyzed. How do you set priorities, evaluate opportunities, and convince your boss? Monetization.

Monetization allows you to compare the value of dissimilar behaviors and prioritize your actions. Where are changes likely to earn the greatest return? What’s the opportunity? How do you measure improvement?

What is monetization?

Monetization is an imposing name for a simple concept—attaching a dollar value to desirable behaviors on your site, whether those behaviors generate revenue or reduce expenses. The obvious example is the average sale amount for ecommerce sites. It gets a little less obvious from there. For example, assigning an average value for a referral to another site or attributing your web site’s contribution to offline sales. The idea, however, is to assign a value not only for conversion events but to those behaviors that lead to a conversion.

Once you’ve assigned a dollar value to behaviors as dissimilar as a visit to a product detail page or a customer service page, you can estimate whether an incremental improvement to the product detail page will likely earn you more money than the saving likely from an improvement to a customer service page. Of course, that begs the question of how to assess the average value of a visit to a product detail or customer service page. That’s where Burby and Atchison come in.

Actionable web analytics

Jason Burby and Shane Atchison have written a wonderfully practical book called Actionable Web Analytics.Burby and Atchison work at Zaaz, Inc., a web design and analytics consulting firm that focuses on monetization as a key strategy. They use monetization to help their clients:

  • Understand the impact of acting on opportunities.
  • Prioritize opportunities based upon that impact.
  • Eliminate time-wasting projects.
  • Focus on projects that drive the business.
  • Provide a common framework for evaluating the impact of change.

Speak to the dog,
in the language of the dog,
about what matters to the heart of a dog.

As well, monetization gives you the vocabulary to talk to management in the language they understand best. Borrowing a graphic phrase from the Eisenberg brother’s book Persuasive Online Copywriting, “Speak to the dog, in the language of the dog, about what matters to the heart of a dog.” To this dog, money matters. Providing an estimated dollar value for proposed changes is highly effective in a conversation about what to change and when.

    Actionable Web Analytics includes detailed instructions on how to build a monetization model for your site. It’s a fascinating subject with tremendous potential that Burby and Atchison address capably.

    Actionable Web Analytics
    Using Data to Make Smart Business Decisions

    Jason Burby & Shane Atchison
    Wiley Publishing, Inc.
    ISBN: 978-0-470-12474-1



    Social Bookmarking

    Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

    March 1, 2009 in Analytics | Permalink | Comments (1) | TrackBack (0)

    adCenter Conversion Tracking

    What I’d like to see next.

    One of the most valuable features a search engine can offer advertisers is conversion tracking. Mapping conversions to keywords and adds can differentiate the performers from the dogs. It’s the first step in optimizing for ROI.

    Basically, conversion tracking is a bit of JavaScript placed on your site’s confirmation page—the page that loads when a prospect does what you hoped they would. The JavaScript fires on page load and sends data to a remoter server. (In reality it can be more complicated, including the download of additional JavaScript, but let’s keep it simple.) Typically, some conditions must be met for the script to increment the conversion count.

    A Microsoft adCenter promotional graphicIn the case of Microsoft’s adCenter—the tool that manages ads for Live Search—those conditions include a browser cookie set by Live Search within the previous 7 days and successfully mapping the cookie to an adCenter campaign. If any of those conditions aren’t met, if the cookie is older than 7 days or was deleted entirely, the conversion won’t be counted.

    A conversion mapped to a keyword in adCenter’s reports is useful but begs the question—what was sold, how many, how much and what was it worth to me?

    Here’s my wish list of additional information captured by adCenter’s conversion tracking script.

    Product SKU

    First you need to know which products are selling. The fact that a keyword resulted in a conversion is interesting, but for ecommerce, you really want to know what product sold—tennis ball or bowling ball?

    Sale Price

    What price did you sell the ball for?

    Quantity

    How many balls did you sell the customer?

    Product Margin

    A ten-pin bowling ball and two pins.From the perspective of ROI, this is where the rubber meets the road—or the ball meets the pin. Obviously, not all products are equally profitable. The sale of a tennis ball will make you less money than the sale of a bowling ball, presumably. A keyword that results in the sale of a high margin product is worth more to you than one with little profit. Your keyword bidding strategy ultimately depends upon the keyword’s margin—the profit earned from conversions associated with the keyword.

    Optimizing for margin requires that you know the margin for each product you sell. That can be a taxing effort for many companies. Often the margin is only roughly approximated or generalized across products. The closer your approximation approaches reality, however, the more accurately you can tune your search advertising strategy.

    Discounts

    Were any discounts applied to the sale? Discounts affect the transaction’s overall profitability.

    Shipping

    If your site offers free shipping—and that’s becoming the cost of entry for most ecommerce sites—then you’ll need to deduct the cost of shipping from the transaction’s profit.

    Tax

    Just so can break out the amount from cost and profit.

    Bottom Line

    If adCenter’s conversion tracking incorporated my entire wish list, it could accurately report not only the margin for each transaction but the margin for each keyword. No more wondering about the ultimate value of a keyword, an ad group, or even a campaign.

    Reblog this post [with Zemanta]


    Social Bookmarking

    Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

    February 25, 2009 in Bidding, Optimization, PPC, Search Marketing | Permalink | Comments (0) | TrackBack (0)

    Search Marketing, Even Harder

    Image representing Microsoft as depicted in Cr...

    I recently wrote a post for the adCenter Analytics blog titled Search Marketing is Hard. In that post I argued that search advertising was especially hard for small business; it requires a paradigm shift from marketing by fiat and intuition, often dictated by the highest paid opinion, to a data-driven process integrating web analytics and content testing, a process designed to wring every last bit of value from every advertising dollar. What’s true for small business is even more true for agencies.

    The Agency Disadvantage

    Agencies who manage paid search are at a disadvantage. Much of what they need to know to maximize return on their clients’ advertising investment in invisible to them. Their line of sight typically ends at the click between search page and landing page. Beyond that, they’re blind.

    In these lean times advertisers themselves have a fierce focus on what most matters to them—ROI. They want conversions, sales, profit. That means optimizing each step of the long, circuitous path from search ad to landing page to check-out funnel to sale. If anywhere along that path the prospect stumbles, they won’t become a customer. No conversion, no sale, no profit.

    How does an agency gain access
    to their client’s web analytic data?
    Save them money!

    It does no good for an agency’s account manager to throw up his hands and complain he has no access to the client’s web analytics and no influence on their web site if he did. It’s true but irrelevant. The only thing that matters is results and better results are achieved by managing the entire sales cycle from first to last. If not managing, then at least influencing.

    75px-MSFT_5years[1] How does an agency make the case for access to their clients’ web analytics? Save them money on worthless clicks. A paid search referral that hits a landing page and then bounces back less than 10 seconds later is a wasted expense. They were never a viable prospect. They were mistargeted by the search campaign. An agency with access to the bounce rate of that page could use the data to refine the campaign targeting.

    • Modify ad text
    • Add negative keywords
    • Demographic targeting
    • Geographic targeting
    • Day parting

    Moving the Needle

    The trending of the bounce rate acts as a needle. Move the needle in the right direction, you’ve succeeded in saving your client money. Don’t forget the ultimate focus: conversions. Reducing bounce rate while reducing conversions is counter-productive. Segmenting search traffic by conversion is a critical function of managing by analytics.

    The practice of web analytics
    is basically the scientific method
    with self-interest.

    Web analytics isn’t an arcane art. Basically, it’s the scientific method with self-interest. Data is a proxy for behavior. Observe the behavior, create a hypothesis to explain the behavior, then test the hypothesis to see if the behavior changes in the desired direction. If it works, repeat; if not, revert.

    Even in my superficial explanation, it’s obvious that testing is essential to the practice of web analytics. Learning the methodology of content testing is another challenge for most agencies. Data-driven decision making is no more a part of agency DNA than their clients'.

    Life isn’t fair.
    It’s Darwinian.
    Get on with it.

    Trusted Advisor

    If life were fair, clients would take responsibility for expertly managing search engine traffic on their site, analyzing data, testing their analysis, and optimizing conversions. Life isn’t fair, it’s Darwinian.

    It’s my opinion that agencies must become expert in analytics and testing in order to optimize for the best results at the least expense. That skill set requires agencies reach beyond their comfort zone and account managers gain new skills.

    • Familiar with major web analytics solutions currently on the market.
    • Expert in analyzing data collected by analytic solutions.
    • Capable of providing guidance to their clients on KPIs and analytic best practices related to PPC advertising.
    • Comfortable with defining and interpreting A/B and multivariate tests.

    Search marketing is even harder for agencies than advertisers because neither are likely to be comfortable making decisions based upon data. Agencies not only have to become expert at data-driven decision making, they need to help their clients become expert as well. They need to become their clients’ trusted advisor. They need to master the complicated integration of testing and analytics because it’s immensely powerful and produces results. And results are what’s going to keep them in business when others fail.

    Reblog this post [with Zemanta]


    Social Bookmarking

    Add to: Yigg Add to: Digg Add to: Del.icio.us Add to: Reddit Add to: Simpy Add to: StumbleUpon Add to: Slashdot Add to: Furl Add to: Yahoo Add to: Spurl Add to: Google Add to: Blogmarks Add to: Technorati

    February 3, 2009 in Analytics, PPC, Search Marketing | Permalink | Comments (4) | TrackBack (0)