SEM - It's Not Rocket Science - Part 2

So last week we talked about paying for your name and the "whys" and "why nots".  I may have confused some people last week when I said that paid services may not have your best interest in mind and then at the end said that you should use a paid service to do your SEM.  Just to verify, you don't have the personnel to monitor your SEM like the paid services do.  But you should track monthly what you're paying for.  So use paid services if you're going to do SEM but inspect what you expect.

This week we are going to go in to the back end of our web hosting service (ours is dealer.com) and pull some keyword reports.  If you don't know how to do this I would suggest that you get some training by the company you use.  These reports are the nucleus of good SEM/SEO.  The nice thing about pulling these reports is you can download them as a csv file and then sort them how ever you want.  This way we can track what's working and what is not.  In the picture below I sorted the keywords buy the number of visitors and then buy pages viewed.  If there was only one visitor and one page view then they weren't interested in staying on your page after they got there so it wasn't relevant to the search that they ran (you can click the picture to see the larger view).

You can see Scat Packs and Columbiana (who's our sister store) as well as just Jeep didn't do well.  But your SEM company doesn't tell you this so they may have you paying for searches that are getting bounced right back out.  You don't want that, which brings me to the second sort.  Bounce Rate.  If the bounce rate is 50% or more you shouldn't be paying for those searches either.  I changed the cell format to percentage on the bounce rate.  See below.

Looks fairly similar to the last report because it's showing all of the pages that had 100% bounce rate.  But if we look at the spreadsheet below where we scanned down a little more it will show the searches at 50% and less.

Our goal is, of course, 0% bounce rate.  As you can see, many of the 50% bounce rates have our cities name in them (if you remember from Part 1, we may not want to pay for those clicks).  Some of the others are nearby cities and zip codes.  Again, may not be the best investment if visitors are going to your site and leaving immediately.  That's like paying for a ad that get people to drive on your lot, circle, and leave.

What is interesting is there are some companies that bid on long tailed searches such as 2015 Chrysler 200. Let me tell you something interesting from my research.  In the year 2014 we had 984 terms searched that we could track.  Of those term an amazing 226 of them were long tailed searches.  That means that 23% of our searches were long tailed!  So do I really know what I'm talking about when I say it doesn't make sense to pay for these?  23% seems like a large amount of searches.  And if that's the only way you monitor it then it would make sense to pay for those.

Let's dig a little deeper though.  Wouldn't you agree that it would make sense to use the search terms that the most customers use? YES (you're supposed to say yes!).  Well then let's take a look and see how many unique visitors we had in 2014.  According to our spreadsheet we had 29255 unique visitors last year.  While we're at it let's check the average time that our visitors stayed on our site.  The average time was 4 minutes and 51 seconds.  

Now let's compare that to our long tailed searches.  

The unique visitors for long tailed searches for the year was 243.  WHAT?!  243?! 243 out of 29255!  What percentage is that? That is 8 tenths of a percent of all unique visitors.  Big difference between 23% and 8 tenths isn't it?  YES (again, you're suppose to say yes! Keep up will ya?).  Now let's check the time on site for the long tailed search.  Huh!  It's 2 minutes and 51 seconds.  A whole 2 minutes below the average.  So not only are we not getting that many unique visitors but they're not spending that much time on the site either.

So the next time someone tries to present you a package that gets its results from long tailed searches (such as SEM or video) you can tell them, I read the article SEM - It's Not Rocket Science - Part 2 and I'm not buying it, Literally!!

Please share with all of your friends in the auto industry and join my blog.  Next week we'll take a break from SEM and discuss the thing that will make or break your dealership.