Graduate students are even crowdsourcing their research costs. This office, as a way to gently tease a colleague with an aversion to birds, “adopted” a kestrel hatchling in a Boise State graduate research project on Rockethub.

Downtown Detroit from Belle Isle

Now, charitable giving can be crowdsourced. I’ve heard that community projects in Detroit have trouble competing for traditional grants. Detroit4Detroit is fostering 150 community projects in 2012 through asking Detroiters (including all of us here in Ann Arbor) to help these projects get off the ground.

Detroit4Detroit will host an event this Thursday to raise awareness and funds. The event, Crank Up the Cause, will take place at Eastern Market 6:00 p.m. to 10:00 p.m. on Thursday, May 17. It has a $10 entrance fee, but you’ll get your $10 back as a voucher to give to a Detroit nonprofit on-site. More information and tickets available via Eventbrite.

If you cannot join Thursday evening, maybe you can organize a group to support one of the causes. My local university alumni club is going to support one of the 150 projects. What a great way to get involved and make a small contribution to the larger good.

Walled garden, made available via creative commons license by recursion_see_recursion on flickr

1. a closed one where the freedom of information exchange promised by the 1990s and 2000s Internet gets subdivided or gerrymandered into separate walled gardens (e.g. Facebook), proprietary technologies  (e.g. iPad or specific apps), in a quest for control of our information, attention, and money, or
2. an open one in which consumer demand for clean usable design and access forces publishers and technology providers to offer more accessible and more open platforms.

Count me on the side of the open future.

With this in mind, I appreciated a recent story in MIT’s Technology Review by its Editor-in-Chief Jason Pontin: “Why Publishers Don’t Like Apps: The future of media on mobile devices isn’t with applications but with the Web.”

From Pontin’s article:

I hated every moment of our experiment with apps, because it tried to impose something closed, old, and printlike on something open, new, and digital.

Score one for freedom, open standards, and usability.

Hat tip to Kottke.org.

• Optimize for clicks, the default setting which shows ads that generate the most impressions and clicks
• Optimize for conversions, which takes the conversion and clickthrough rate into consideration
• Rotate evenly, a popular setting which tries to give each ad similar impressions

The Weasel

The rotate evenly setting is often used to perform A/B tests on ads. The change of this rotation setting is why PPC’ers are so edged. But I think everyone needs to not tax Google’s gig so hard-core.

First off, you can still rotate ads evenly for up to 30 days. After this time period, if there have been no creative changes made to ads the entire adgroup will begin to show ads that are generating the most clicks. Most ad tests should collect enough data within 30 days to determine a statistically significant winner. Some smaller, niche campaigns may take longer but if you make a small change in creative you can give your adgroup an extra 30 days of even rotation.

Test Your Testing

Sometimes rotating ads evenly does not give you the ad with the greatest performance. See the following data set we collected over several weeks of testing.

We tested three sets of display URLs. Original had just the domain (www.mtvbeachhouse.com), Location added a location to the URL (www.mtvbeachhouse.com/daytona-beach), and Keyword referenced keywords from the adgroup (www.mtvbeachhouse.com/spring-break).

We ran the test using the optimizing for clicks setting. Notice that while the Location test saw a higher CTR and conversion rate; Keyword generated more conversions – because AdWords thought the Keyword ads were more relavant and gave the ad significantly more impressions.

You may say that if we rotated the ads evenly then Location would have generated more conversions due to the higher conversion rate. With this highly competitive industry these ads are targeting, visibility is key. We try to generate the most impressions out of our ads. Since all the ads are being directed to the same landing pages we depend on our user experience testing to increase the conversion rate, not Google’s choice of ads. Our confidence in the landing page allows us to show ads that generate high amounts of traffic. Until our client’s CRM system shows that the leads are becoming less qualified, we will continue to optimize for clicks.

In closing, PPC’ers should even test their method of testing. The SEO’ers will tell you not to put all your efforts behind one thing because Google is known to change things suddenly (see the Panda update).

Finally, don’t wheeze the juice:

It isn’t often we get to mix philosophy and SEO, but this question sparked my interest. Unlike many news outlets, we assume an unbiased and unfiltered results page through the major search engines. But that is hardly the case at all, especially as SEO expands as an essential piece to a company’s web strategy and as a profession in its own right.

Especially for large and competitive concepts, the websites with the time, money, and expertise are ultimately going to rank well for the set of keywords for which they optimize. Moreover, to gain any online visibility at all, many websites without a high Domain Authority are forced to optimize for very specific keywords that they believe will allow them to rank well.

Google’s intention is to serve up the best website possible for the intent of each query, but this is an inherently impossible task. Each algorithm update is meant to improve the quality of the results page. After all, their credibility depends on doing this better than their competitors! But it is a constant game of catch-up as clever people find more clever ways to game the system, even if only temporarily.

I asked a few other members of our SEO team to chime in on this topic…

—————————————————————————————–

Scott DeRoche: Pure Visibility SEO Analyst

“Google is the medium for the organic message; the hierarchy of its results influences searchers who largely assume that the first few results returned are all of the information that they need. Pages that do not receive top ranking largely fall into the wayside of ignored results.

It is difficult to say that ‘history’ is defined by SEO, but certainly success on Google makes or breaks websites and businesses. Sites that do not have optimized content often fail to succeed at ranking competitively on Google.” -Scott DeRoche

—————————————————————————————–

Taylor Caldron: Pure Visibility SEO Analyst

“Ultimately, SEO writes history by controlling which ideas ever make it into the public eye. It influences what we buy, how we vote, and what we believe (which, for Google users, is generally whatever the latest editor on Wikipedia believes).” -Taylor Caldron

—————————————————————————————–

I also included a powerful TED talk about search filters by Eli Pariser. If you’ve liked this post thus far, I highly recommend spending 10 minutes to listen to his talk:

“As web companies strive to tailor their services (including news and search results) to our personal tastes, there’s a dangerous unintended consequence: we get trapped in a ‘filter bubble’ and don’t get exposed to information that could challenge or broaden our world view.”

Eli Pariser argues powerfully that this will ultimately prove to be bad for us and bad for democracy.

Then, one day, it changed. It looks completely different, and a notable change is that it now promotes newer, exciting features. One detail stood out to me: the “Create an account” button logs you in.

Wait, what? The button doesn’t do the thing that it says it will do? Well, it works that way if you’re already logged in to your Google account in that browser.

There are two links in that part of the page: The big orange “Create an account” button and next to it, “Sign in” in plain text. A big, jaunty button that looks clickable and a bit of plain text.

If you’re in a hurry and used to just clicking through this page on the way to your account, guess what happens? You accidentally click on the button.

Having the “Create an account” button sign me in is a nice touch. I only noticed it one day when I sat there and watched the page load after clicking the button. In that moment, I actually read the button’s label for the first time and realized what I’d clicked.

Would I have to waste time backing out of some sort of account creation form? I thought for sure that I’d clicked on that button several times already since the redesign! It turns out I had, and every time things worked out well for me because I was taken to the right place.

Over the next few days, I trained myself to click on the “Sign In” link just in case I’m using a strange browser, but I still appreciate Google anticipating that people would accidentally click on the button.

It’s hard to say how this kind of attention to detail affects your bottom line… how much does your conversion rate go up because you help a portion of your users avoid wasting a few seconds by preventing an error?

While it can be challenging to pick out the effects of these minor touches, as part of a larger program of detail-oriented user-centered design, it can pay off for you.

xkcd's cartoon "ablogalypse"

In response to this news, you might have the question: “What is Tumblr?”

Tumblr is a blogging platform that is more expansive than a microblog like Twitter (Twitter restricts updates to 140 text characters). Tumblr includes photos and video seamlessly in the body of the update, as opposed to by link. Sharing photos, videos, and discussions is what it seems to do best. The very visual nature of Tumblr can be seen in the recently famous “Texts from Hillary” Tumblr, which actually got a submission from Madame Secretary herself, Hillary Clinton (or her staff).

The growth and reach of Tumblr to me was underscored when longtime A-list blogger Jason Kottke released a version of his kottke.org “remaindered links” blog on Tumblr: bonus.kottke.org.

Like Facebook or Twitter, yet unlike WordPress, I can post to Tumblr using an app on my phone, which lowers the barrier to posting. (Bored while waiting? Amuse yourself by taking a photo, dictating a message, or reblogging someone else’s entry via the Tumblr app on your smartphone.)

I’m a latecomer to Tumblr, I was inspired to play with the platform by the back and forth in the comments of this Google + Post by Jeremiah Owyang on the acquisition of Posterous by Twitter. I began to feel a little old-fashioned about my WordPress reliance. So far, it has been fun to play on Tumblr. Find me on Tumblr if you have an account.

Yet, so far, Twitter still wins, outpacing both “blog” and “Tumblr” in Google searches.

A graph showing the relative popularity of search terms. Twitter is more popular that Tumblr or Blog.

Twitter’s higher popularity is confirmed by usage statistics.

Tumblr statistics as of noon Eastern Daylight time on April 16 (or 16:00 Universal Time)

As of this writing (noon Eastern Daylight Time 4/16/2012), Tumblr had accumulated 21 billion individual posts on 51 million blogs and 63 million posts so far TODAY (check for recent tumblr stats here). A recent post from Twitter puts the equivalent numbers at 140 million active users, and 340 million tweets/day in March 2012 (up from 170 million in March 2011). Perhaps not surprisingly, Twitter shares these numbers less visually than tumblr does.

Yet, as soon as you add Facebook to the mix, there’s only one winner.

A graph showing the relative popularity of search terms. Facebook dominates Tumblr and Twitter.

Now that the important business is out of the way, on to the Spam.

Occasionally, you might find that there are visits to your site that are spam, and you will want to take those out before doing any analysis. It can be tricky to find the spam though, so this post will take you through a couple main things to look at when determining what to filter out.

Spam Indicators

Time on Site

You want to watch out for time on site that is one second or less. Generally, when a robot hits a site it stays on only long enough to be counted and then leaves. This means that in your analytics, they will show up as 00:00:00 or 00:00:01 for the most part.

Bounce Rate

Because robots usually just hit one page and then leave, you want to look for bounce rates of 100%.

Finding the Source

Finding the source of your spam can me a very difficult and frustrating process. There are some places that you can always go to look, which are outlined below, but you might need to do some searching of your own as well. Here are some common places where we have found spam.

Region

One of the first places we tend to look for spam in Google Analytics is by region. Select Audience -> Demographics -> Location and then add a secondary dimension of region. At a glance, you should be able to see if something looks off.

If you know that you average about 3,000 visits from the United States in a week and when you glance at your regions and you see 3,000 visits from Ohio alone, you are probably looking at spam from that area.

Browser Version

Filtering down to Browser, under the Audience -> Technology Section, you can add Browser Version as a secondary dimension. As with region, this can usually pretty quickly call out anything that is exceedingly high.

Of course, these are just two areas where we have noticed spam hides, there could be more or different dimensions in which your own spam is appearing, so you will want to make sure you do a through search of your own analytics.

Further Analysis

After you have determined where you think the spam is hitting, or from which browser version, you should set up a filter to include only that dimension for further analysis. Once you have done that, you want to look into a few more dimensions to determine which is the culprit so that you do not filter out any good data inadvertently.

A good place to start is with the service provider. This can be found under Audience -> Technology -> Network. We have found that this is frequently the source of bad data. You might also try browser version and region again, or move on to something else, like keyword.

No matter what you find, you are going to have to do a lot of manual work to make sure that you are only filtering spam. This can be a long, difficult process, but clean data is worth it in the end, and hopefully these tips will get you started!

The lunch is the seventh in a series of community executive networking opportunities. Several speakers will be speaking about the importance of customer services for non-profits and small businesses. The event is an effort to discuss the work of the Young Leaders Society and provide opportunities to get involved in the community.

Additional presenters are:

• Rich Sheridan, CEO of Menlo Innovations
• Dave Hile, Owner and CEO of Hile Design
• Dave McCann from Google

The mission of the Young Leaders Society is to promote and encourage the spirit of philanthropy and volunteerism among the young residents of Washtenaw County and other individuals with similar professional and community interests. Members of the United Way Young Leaders Society are committed, compassionate and generous individuals who are dedicated to making a positive impact throughout our community.

Event details can be found online.

The review is reprinted here, with permission from CHOICE.

We are now firmly rooted in the “Internet Age.” Marketing books such as Internet Marketing Start to Finish now suppose that a business already has a Web presence and is functioning in virtual space. This altogether excellent compilation of best practices and practical advice is intended to guide businesses with their Web presence.

The authors, Internet marketing professionals, offer guidance on improving and integrating Web functions into a business, specifically focusing on marketing strategies. While breaking down “silos” is as old as humanity, this quest can be aided by the authors’ evidence-based approach, which attempts to provide a rational base for employing Web strategies and integrating business functions to bring customers into business processes to build efficiency and customer loyalty. This process can be led by marketing, as the authors posit.

This excellent book argues persuasively for the marketing perspective and is must reading for practitioners and students interested in developing Web strategies for business success.

Summing Up: Highly recommended. All business collections.

Copyright by the American Library Association.

While this study is clearly meant to adress this, it could have been titled: “Are you paying for traffic that you could be getting for free?” Instead, Google opted for something about “Ad Click Incrementality” which basically means “Are my ads sending me incremental visitors that organic traffic wouldn’t otherwise provide if my AdWords ads were shut off.”

After reading through the full study, I now appreciate the official title incrementally.

Search Engine Land gave the study a brief summary. The broad-stroke, non-scientific conclusion is that it’s probably worth it to advertise in paid search for these reasons:

1. For sites in the study, most ads showed for terms that had zero organic rankings. And any click that happens on an ad that doesn’t have an organic listing accompanying it, would be sending incremental traffic.

2. Even when there were organic listings which ranked well on a page, paid search ads contributed significant incremental visitors that the website wouldn’t have otherwise gotten.

Another one of the main points of the study is that “even when advertisers show up in the number one organic search result position, 50% of clicks they get on ads are not replaced by clicks on organic search results when the ads don’t appear.” A very interesting conclusion, since (presumably) a large share of number one organic rankings would be covered by advertisers bidding on their own trademark.

This is a glancing reference to another age-old question: “should I bid on my own trademark?” Going strictly by traffic, the Google study indicates “yes.”

From what I’ve seen, using Google Analytics and sales/lead data, none of these conclusions are surprising to me. They’re consistent with my experiences. Although I still come up short on the reasons why, when bidding on your own trademark, it brings in more traffic, even when you are ranked number one organically for trademark queries.

Using Google Analytics, I have measured trademark traffic both before and after showing ads on trademarks, and it does seem to be the case that ads increase overall trademark traffic.

One hypothesis I have to explain this is that people get used to seeing Google’s yellow box full of ads at the top of search results. Maybe someone who is browsing around begins focusing more on top ad spots and stops paying attention to organic listings, as if the yellow box was validation of relevancy for their query.

Another hypothesis I’ve had is that visitors are interested in seeing every listing for a site – maybe because they aren’t easily finding the information they are originally looking for from the first organic listing, they choose to click the ad. It could be variable across sites, but I doubt Google would help find a generalized answer to what’s going on here.

Of course, Google didn’t measure any trademark queries directly in their study, so it’s all still an assumption. And if you look at the full study, Google also didn’t measure directly the incremental clicks gained specifically on ads that rank for terms that have a number one organic listing.

Instead of direct measurement, they normalized data that includes incremental clicks from all ads, then took a limit to exclude incremental clicks for ads that appear on pages that don’t have any organic listings. From an analysis perspective, I find this confusing because it seems like it would put focus on only a portion of their set and also still include values for ads showing on pages with organic listings lower on the page.

The study also doesn’t give numbers for the mean percentage of clicks from ads that appear next to organic listings. Guessing these numbers from box plots, however, everything seems to add up.

Rather than go by any of these studies, it’s better to do your own tests. You can just try a change and compare data across time periods, but Google also outlines a methodology for creating randomized tests by varying geo-targets.

The idea here is to have a test group and a control group, split up by random geographies for making comparisons, and doing what’s called a “randomized, pretest-posttest, control-group design.”

Googling that type of experiment and looking through Google’s geo experiment methodology, you should see statistical methods for analyzing data from that kind of experiment.

But if terms like “linear model” and “analysis of covariance (ANCOVA)” make you uncomfortable, it’s a rather tough read. Analyzing difference in data without statistical models is possible, if you acquire a lot of data and can see large differences, but it’s a good idea to have the math on your side. After all, you might not know what is “a lot” and what is “large.”