A few weeks ago, Google unveiled some new changes to AdWords’ system of matching keywords to search queries. According to the official Google blog, these new campaign settings will automatically use your existing keyword lists to match for misspellings, acronyms, and other close variants in existing phrase and exact match terms.

Search Engine Land summed up the occasions where Google’s new system may kick in:

• Misspellings (“waterprof sunblock” instead of “waterproof sunblock”)
• Singular/plural forms (“beach balls” and “beach ball”)
• Stemming (“single serve” and “single serving”)
• Accents (“hotel” and “hôtel”)
• Abbreviations (“Dr.” versus “Doctor”)
• Acronyms (“NYC” versus “New York City”)

So, this is a noble concept. I’ll give Google credit for that. Essentially, Google is taking out the blood, sweat, and tears out of creating keyword clusters around brand or other misspellings, and taking the guesswork out of matching for every possible combination of keywords a user could describe a question or product they are searching for. The focus is user-intent, they say.

According to Google, 7% of all searches contain misspellings and the longer the query, the higher the rate of misspelling. The new match type setting will allow marketers to focus on core keywords rather than mine Search Term Reports for long-tail phrases or misspellings that should be captured and bid on. In this way, your accounts will be less cluttered with phrase and exact match long-tail terms and other variants–all of which tend to carry low CTR. And we all know that a great campaign starts and ends with a great CTR. (Think about how this plays into Quality Score and Ad Rank!).

But, let’s take a step back and critically examine what this change means for SEM/PPC professionals. If history has taught us anything, it’s to beware of Google! Always question what new options are in front of you. How can this benefit my account and my client…but more importantly: how could this new option go wrong?

Here are my thoughts on this new (optional!) setting for phrase and exact match types:

PRO:

1. Helps Advertisers with Limited Time to Manage Campaigns

If you’re a small business with little time to dedicate to paid search programs, this setting is for you. Sometimes mom and pop shops just can’t put 8 hours a day into optimizing their campaigns and neither can they afford to pay premium agency prices for someone else to handle it. In this case, it’s easy for anyone to set up a campaign and a couple of ad groups focused around their product or service and let AdWords do its magic.

The setting will catch relevant variations on phrases and exact matches such as “clubbing in cambridge,” “handmade wedding dresses boston,” or [gourmet cupcakes]. Think of all the combinations of the above keywords that could be relevant for a search such as the aforementioned. This new setting would capture a search for “wedding dress boston” in the singular, [gourmet cupcakes] in the plural, and so forth. “Clubbing” could possibly be matched by the phrase “club in cambridge.” Plus, any acronyms that Google is smart enough to catch will enjoy some extra traffic by enabling this setting.

AdWords beginners don’t have to bother as much with match type settings, tests, or advanced optimization. Combine this with a good geo-targeted strategy, and your local marketer is good to go.

2. Insight into New Keyword Opportunities

Looking for keyword opportunities that you haven’t yet thought of? This is for you. So let’s say that you have a product that’s new-to-market. And no one really knows what to call this thing you’ve built–but you know it satisfies a consumer demand. How can you use this setting to build out your existing keyword lists? Looking at Search Term Reports is one way, but having AdWords do the work for you may be a less time-consuming strategy. Operative terms like “contraption for oracle server” or “contraption for doing this to this product” are very difficult to build out keyword lists for. The difficulty is not in the operative keywords, but in the phrasing. “A product is used for B” is very different from “A compliments product B”. The direction of usage, of ownership–and this goes back to the basics of the English language–really do affect what keywords you should be bidding on.

3. New Setting Won’t Affect Quality Scores (Sort of…)

According to Google, your quality scores are safe with the new match type setting. But, this does come with a catch. If your account uses the new match type setting, any derivations that the algorithm may serve ads for will follow the quality score of the root keyword that triggered the relevant variation…that triggered the ad.

While we could get into the concept of how this new feature is taking power away from modified broad keywords–

CON:

*I guess I should take the time to explain that my views reflect ONLY my views and not the views of Overdrive Interactive or anyone else! Hah!

1. Wreaking Havoc on Dynamic Keyword Insertion Ads

Firstly, you should always be cautious when setting up keyword lists that will accompany Dynamic Keyword Insertion ads. If you don’t know what these are, look them up! It’s advisable that you do not use broad keyword lists for these types of ads because any individual keyword in a broad-matched term can trigger an ad. This sometimes produces headlines that are nonsensical or are just flat-out wrong.

Using the new keyword match type settings will essentially broaden your exact and phrase keywords and therefore broaden your Dynamic Keyword Insertion ads’ probability of displaying completely incorrect (or even illegal) ad headlines. Something to think about!

2. Blurs/Overshadows the Modified Broad Keyword Match Type

This was my first reaction when I first heard about this new campaign setting. What about my broadified keywords? If you don’t know what the modified broad match type is, check it out here.

Essentially, +this +keyword (yes, with a + in front of each word [called a fully modified broad match]) tells Google “my ad can trigger for a query that contains every single one of these words, plus similar spellings and other relevant variations.” Although this is NOT what the new match type settings are turning your exact and phrase keywords into, it DOES take a little bit of power away from fully modified broad keywords.

Of course, this isn’t confirmed–yet. But, this is how I suspect the situation will play out. If AdWords favors the un-modified broad keyword, you may find your impression levels spike. Depending on your campaign settings, you may find that Google will favor root keywords that tend to trigger a lot of irrelevant impressions rather than your extremely targeted modified broad keyword that generally doesn’t cause Search Term Report headaches.

3. CTR Will Suffer

I won’t preach too much about this one. More possible combinations in keywords that can trigger your ad = impression increase. And not all of that traffic will be qualified traffic! Do not expect that your clicks increase unless your ad is poorly written and/or confusing to the user. This deadly combo will cause your CTR to decline.

4. Google Wants to Increase Q2 and Q3 and Subsequent Qs Revenue

The conspiracy theorist in me is having a field day with this. Irrelevant clicks = wasted money for marketers but PROFIT for Google. And since their profits in Q1 were only up 12%, they need to increase the number of irrelevant clicks if they want to have a similar Q2 and Q3! </conspiracy>

Google has also let marketers to believe that they are missing out on monetizing the misspellings, synonyms, and variation-type keywords. Bidding on every single combination of keywords is like trying to pick up a handful of sand and expecting to not lose a single grain. It’s just impossible and inefficient.

5. Double-Triple-Quadruple Check Your Search Terms Report!

More possible combinations in keywords = more irrelevant queries that trigger your ad. The evidence of what causes someone to click an ad will be apparent in your Search Term Report. I expect the *expected* increase in impressions will increase the amount of data in your STR. This causes several complications.

First, you’ll have to spend quite some time surfing through your STR to make sure those queries are at least somewhat relevant. Secondly, you’ll use that analysis to make a pretty comprehensive negative keyword list. This is very important!

6. Wasted Spend on Irrelevant Clicks

There is also a strong possibility that your keyword list, if not properly optimized, will trigger ads for irrelevant queries. An increase in irrelevant impressions will also increase the probability of irrelevant clicks, which will result in wasted money.

7. CPCs will Increase AdWords-wide

An increase in the number of possible keyword combinations that could win the AdWords lottery artifically inflates the compeitition for keywords. Think of the keyword space as the bullseye at the center of a dartboard. Lets say two competitors have, each, 50% of their keywords overlapping. So this would look sort of like a Venn Diagram. (follow me for a second…)

Adding the keyword match settings will increase your reach or amplitude by one ring on the dartboard. Think about the range of opportunites (and increased overlapping opportunites) which now present themselves. Look at how much more area is “overlapping” on your dartboard. I should also point out that this “extra ring expansion” was not due to a budget increase, keyword list increase, or anything else but a simple campaign setting. That in itself defies basic economic laws.

I think this will give the illusion that there are more competitors in a space than there actually are. I think that increased impressions and number of qualified queries for an ad will absolutely increase the competition for a lottery and lead to increased CPCs. Again, this hasn’t been tested on any real PPC accounts yet; we may find that this accusation is false.

I’ve rambled enough. What are your thoughts?

I write for Overdrive Interactive‘s blog. Check out my new post on Enhanced Sitelinks in Google’s SERPs.

http://blog.ovrdrv.com/a-new-look-for-google-sitelinks/

Among the frenzy of music-related social sharing applications, a few major players have set their sights on becoming the go-to app for sharing and experiencing music on the web.

I will start with a few you already know:

Spotify, launched in the United States last year to an initial audience of invite-only users, quickly became a favorite among music connoisseurs. The platform-based application allows users access to over 15 million songs within its database. Users can set playlists, listen to Radio, a random playlist of genre-specific groupings of songs, integrate accounts with Facebook and Twitter, and more recently, access 3rd party applications from Last.fm, Pitchfork, and other music tastemakers.

Even before Spotify launched in the US, Turntable.fm held our attention throughout the spring and summer of 2011. The service allowed users to join virtual “rooms” where individual users could select songs from Turntable.fm’s database or use songs from their own libraries to play for an audience of other users. Rooms were set up around themes–usually genre-based themes or moods–to create a “concert” feel. Some of the most popular rooms included “Indie While You Work” (implying that many users were listening to music passively while at work), “Coding Soundtrack” (implying that many users were also computer programmers in some capacity), and “Dubstep” (a genre that, in the summer of 2011, was on a path to mainstream success).

In both platforms, and in fact, all recent web 3.0 applications, there is an essential emphasis on the sharing of one’s tastes. And in theory, the expression and sharing of tastes among a considerably large audience will lead to music discovery.  (As I write and re-read that statement, I cannot help but think of what Theodor W. Adorno would say about “sharing one’s tastes”…but, that is for another, far more lengthy discussion).

The whole point of Turntable.fm or Spotify is not to share one’s taste in music–that is secondary. However, a new web application, This is my Jam, aims to fill the gap between actively and passively sharing music.

TIMJ captures the song that has been stuck in your head (aka: your current “jam”) and use your song, along with the jams of your friends, to enable friend-based music discovery.

The whole process takes just two steps: Pick your “song of the moment” and post it to your profile, then follow your friends or people who have favorable tastes in music. You can even create a playlist of your friends’ picks.  Each week, you’re reminded to change your Jam. This is so a constant stream of like-minded Jammers can be introduced as being similar to your tastes.

In essence, TIMJ does what many other music platforms cannot–aggregate music from the friends that actually influence your music discovery and subsequently your taste for new music.

However, in the age of platforms-that-do-it-all, This Is My Jam just isn’t my jam. After using the app, I found that many people that were suggested to me just did not share my taste in music. Yes,  X-user and Y-user posted my current Jam only one week ago and a few other songs I posted awhile back, but our tastes in music still do not align well. I also could not find the time to sift through my Facebook and Twitter leads to find people who actually did have a good playlist. Nope. The best playlists I knew were of my own creation.

In applications meant for mass consumption, developers need only to be concerned with one thing: keeping users enticed and using the application daily. TIMJ fails to integrate with the players I use the most, like Spotify, Turntable.fm, or Google Music. TIMJ can’t even tell me my real jam–the song I have played most on Spotify within the last week.

If TIMJ could integrate with my usual music players with the same goal–to connect me to new music through similarly-aligned users–I would use it every day.

On another point, I wonder to what degree the TIMJ algorithm does and doesn’t return results that are of interest to you. Will users stick around long enough to get desirable results? From what I can tell so far, TIMJ has not returned any users that really give me a good balance between the good and the new. However, this is more than likely because TIMJ just does not learn the way Spotify and other applications are capable of learning about our taste in music. I don’t post songs as often as I listen to them on Spotify. There are simply huge gaps in the data collection funnel.

TIMJ also requires more effort from the user–a lot of effort. I cannot foresee this application gaining much momentum if the integration is not there. I would love to see some automation in the friending and playlisting features, for example. This would minimize my level of frustration when trying to find different users with similar (or different!) tastes.

Though I’m skeptical of the longevity of TIMJ, I look forward to the multitude of directions that could be beneficial for this application.

PS: Throughout this post, every time I wanted to mention “posting a song,” I defaulted to saying “scrobbling a song” in my head. One thing I really like about Last.fm is their ability to scrobble songs their database silently and with no interaction from me whatsoever. Quite nice. To this day, I still believe the best recommendation engine for music has been Last.fm. They were the first on the social-music scene and because of that they’ve built up a huge repertoire of data on users at the individual and group level. I’d love to see them launch something more than an app for spotify. I’m sure they’d be able to utilize data from the Echo Nest and Hype Machine rather effectively.