Editor’s Note: This post was originally published on Seattle 2.0, and imported to GeekWire as part of our acquisition of Seattle 2.0 and its archival content. For more background, see this post.
By Sasha Pasulka
This is such a common mistake, and it’s one I made over andover and over again before I had a light bulb moment. I pored over my GoogleAnalytics data and assumed that the keywords that were popping up in my searchengine referral log were the only keywords for which my sites were ranking, orat least the majority of the keywords for which they were ranking.
This was not at all the case. These were the keywords thatproduced the search engine results thatpeople clicked on.
I finally sat down and compared what Google Webmaster toolslisted as my search queries to what Google Analytics listed as my referralkeywords. There were some major discrepancies.
For instance, a category page on my celebrity gossip blogwas ranking well for “Katy Perry,” but the title and description it wasdisplaying was about Katy’s long-ago fling, Benji Madden. As a result, theclick-through rate was below 10%. Similarly, a page on my women’s issues sitewas ranking well for “Bristol Palin,” but the title and description were aboutabstinence. When I tweaked them so they focused on Bristol’s televisioncampaigns and print ads for abstinence, the click-through rate shot up. Peoplewanted to see and hear Bristol talk about her pet causes – they didn’t actuallywant to read about abstinence.
I had to think about how my page titles were phrased, whatmy meta descriptions said, and, most importantly, what searchers wanted when they typed in thosekeywords. If I was providing theinformation they wanted, but I wasn’t making that clear in the title and thedescription, some tweaks were in order.
This leads me to my second point.
2. Use organic search terms to segment your market.
Your website doesn’t have to serve only one market. Thisis true if you’re an ad-supported website (like the ones I run) or if you run awebsite with the goal of driving a purchase. An analysis of the search termsfor which you rank – and for which your competitors rank – can help you furthersegment the market and determine where you fit in.
Let’s say you run a website that sells widgets. You assumedyour target market would be women ages 30-45 searching to buy widgets forthemselves, but when you analyze your organic search terms, you realize thatyou’re ranking for “anniversary gift” and “best widget for my girlfriend.” Thenyou analyze your main competition and realize that they’ve targeted their SEOfor “graduation gift” and “widget for college student.” They also seem to betargeting “fubars,” a potential substitute for the widget.
(You can often get a sense of your competition’s SEO targetsjust by scanning the tags on their page source, but if you want a more detailedanalysis, you can invest in a professional SEO tool.)
By looking at organic search data, you’ve just uncovered threenew active markets without investing very much time or money in marketresearch.
Now, not only can you tweak your SEO for these markets, butyou can change the experience the user has on your site based on the searchterm they’ve come from, regardless of the page they land on.
We spend a lot of time optimizing the navigation of websites(“For Grads”, “Anniversaries”, and “Just Like a Fubar!“), assuming searcherswill land where we expect they’ll land. This is a very, very useful practice,and most definitely something you should be doing. But if your site gets a lotof traffic from search, it makes sense to take it a step further and actuallyuse the search term to determine what content the site displays.
For instance, if a user comes in to your home page using a search termcontaining the word “girlfriend” or “wife,” you might want to make sure youdisplay a picture of a woman gratefully kissing her male companion, widget inhand. If a user comes in using a search term containing the words “college” or“graduation,” perhaps a college kid high-fiving his awesome widget-bearing dadwould be more appropriate. If your user comes in on any variation of “fubar,” considerusing a portion of the site to do a simple and favorable comparison of fubarsto widgets.
3. Always Be Testing
I stole this from REI’s Jonathon Coleman, who gave afantastic presentation on organic search at the Online Marketing Summit earlierin the week. (You can check out the slides here.)
It’s a good mantra: Our SEO failures and successes should bequantifiable. It helps to have more specific goals than “I want to sell moreproduct” or “I want more traffic.” A clearer goal would be “Over the next ninemonths, I want to triple the number of widgets I sell to the anniversary-giftmarket as a result of search traffic.”
Now that you have a clear goal, you can start makingactionable, testable decisions. We like to overlook the “actionable” and“testable” parts and just make decisions. I know I do. I make a lot of decisions. I make them quicklyand I make them all the time, and, asa result, they are not always good decisions. More tellingly, though, I usuallydon’t have any idea if they were good decisions or not, because I’ve set up noclear path for action and test.
For the goal above, I could break it into two portions:
a) I want to double the amount of search traffic Iget for anniversary-related searches
b) I want to improve conversion rates fromanniversary-related search by 20%.
Okay, cool. Those are actionable goals that you can testagainst. Now, take a break from being proud of yourself for creating actionablegoals that you can test against, and actually set up the tests. (This isanother part where I have trouble.)
You can use Google Analytics and multivariate testing totest against both of these goals, or, if you want to get fancier, try aprofessional SEO solution like SEOMoz or Optify or any other one of the eightbillion SEO analysis companies that multiply in Seattle every time it rains. Knowyour baseline and track your actions against changes to your baseline. If youwant to show off, make a bunch of changes over a span of time and run aregression on the data. (Seriously, though, don’t do that.)
We all know how todo this kind of data analysis – everyone here took a lot of math in school –but I’m always surprised at how few entrepreneurs do. We like to be making decisions and doing things and changing the whole world asap, notmeasuring boring results. But there are so many ways to quantify search-marketingdecisions, to see what works and what doesn’t. There’s just no excuse not to doso when you consider the relative payoff.
Now, will somebody please volunteer to call me every morningand remind me to do all the things I just told you to do?
