Tip:
Highlight text to annotate it
X
[Intro] Hi I’m John Nicholson with Marketade. We’re
an internet marketing company in Washington DC. We hear from a lot of companies who’ve
done some search engine optimization - or SEO - for their website and are frustrated
because they have no idea whether it’s helping their business.
If they’re doing anything at all to measure SEO, it usually consists typing their favorite
industry keywords into Google and seeing where they pop up. Or if they’re a big company,
maybe they use software to track their rankings automatically. But it doesn’t matter how
sophisticated you get ... if you’re focused on rankings, you’re focused on the wrong
thing. A much better approach is to use web analytics
data and today I’ll walk you through a 5-step process for measuring SEO impact using Google
Analytics. As an example I’ll use data from the website for Affinity Lab -- a shared office
space in DC for small businesses (and where I am now).
Now in a minute I’ll take you into their GA account and show you the process but first
let me give you the quick story behind our SEO work for the Lab ... because if you don’t
understand our goals you won’t really understand the steps.
[The Story] So the Lab has a great network and brand in
DC and a lot of people who go onto become their customers first hear about it from a
friend and then Google their name ... And as with most companies, the Lab has always
been at the top of Google on these brand searches. But there are a lot of people in DC who have
not heard of the Lab -- or who forget about it -- and are online searching for something
just like it. They’re Googling things like “dc office space for small business” or
“coworking space in dc” or “shared workspace”. A year ago the Lab did not rank well on Google
on a lot of these non-brand searches and they were missing out on this great source of leads.
So earlier this year we did some research and identified about 15 of these target keyword
phrases and optimized the site around them. We did most of the SEO in Feb and a little
more in June. Now I want to see if it worked. Are we getting more traffic from these target
keywords? And is it quality traffic? To figure this out, I need to pick a period after the
SEO and compare its results to a similar period before the SEO.
For the after period I’ll use July thru Sep (last full month we have data for). Ideally
my before period would be those same months last year -- to remove seasonality as a factor
-- but the Lab didn’t start using GoAn until mid-Oct last year, so we’ll use Nov
thru Jan. It’s not a perfect comparison, but it’s good enough to see if the SEO is
working.
[1. Select Before and After Periods] OK so here I am in the Lab’s GA account
and I start by going to the date range section ... and selecting the “after” time period
... and then I check “compare to past” and select the “before” period. Now we
can see total traffic to the site for these 2 periods side by side. (11/01/2009)
[2. Select Non-Paid Keywords] Next I start narrowing down the traffic. First
I go to Traffic Sources and select the Keyword report which gives me search engine traffic
broken out by keyword. Then I select non-paid. And this gets rid of any traffic from search
advertising, like AdWords or other sponsored links.
[3. Filter Keywords] And here I can see they’ve had 2800 non-paid
-- or “organic -- search visits in the after period, a big jump over 1300. That sounds
great, but a lot of this traffic is from keywords that have nothing to do with our SEO efforts
-- in particular the brand searches I talked about earlier.So I need to focus on traffic
from our 15 target keywords. And rather than scanning through 1000+ keywords I go to the
bottom and use filters to isolate our target traffic.
And here's the right way to think about filters. So far we’ve been narrowing down our traffic
at each step. But given the way people search, if I filter on exact keyword phrases, my results
will be too narrow. So for filters I broaden our scope a bit, first by identifying common
threads or themes among my target keywords, and then by using the smallest possible root
for each of those theme words. For example we had couple target phrases with
the word “shared” -- so I use "shared" as my theme and then I use “shar” as my
root so I can get as many relevant variations and misspellings as possible.
So I click Advanced Filter and set one filter to contain “shar”. And then I set another
to exclude “aff” which will remove any searches with “affinity lab” or misspellings.
Back up top I can see the result: in the pre-SEO period we only had 17 visits from people searching
on words like “shared office”. In the post-SEO period that number jumped to 119.
And I just take these numbers and enter them in a spreadsheet.
[Detour: Why I Love Filters] And that’s all you have to do for this step
... but before moving on let me show you why filters are so important - because a lot
of people don’t really get this and they skip this step. If I scroll down I can see
the before-and-after numbers on each of the 60 or so individual keywords in this bucket
that sent traffic. And of course we notice these big jumps on the phrases we had on our
target list. But just as important is this big group of
keywords further down that went from 0 to 1 visit. These tend to be longer, more specific
phrases we never would have thought of in advance -- but a lot of them are a great match
for the Lab. Whoever did this search visited 15 pages and spent 29 minutes on their site.
Here’s the same report in another tab except I’ve removed the “before” period and
I’ve selected the graph view. Notice how long the tail is on this graph -- all these
keywords sending 1 click that don’t seem like much until you total them up. In this
case they make up about 50% of all traffic for this KW group.
This is something we see over and over again when SEO is working and it’s why filters
are so critical. They let you capture and aggregate this long tail traffic -- traffic
that most people completely miss because they’re just scanning keyword lists (or rankings)
for the obvious, high volume terms.
[4. Measure Quality] OK back to our steps ... so for this keyword
bucket, we can declare our SEO a success in terms of driving more traffic. But what we
really want is more quality traffic -- people who are likely to join the Lab. So in addition
to looking at visits for this keyword, I’m going to look at 2 simple yet powerful quality
metrics: Avg. Time on Site and something called Bounce Rate -- which is the % of people who
bounce off your site without doing anything. [point with arrows] We want TOS to be high
and BR to be low. And the relevant comparison here is not “after”
vs. “before” but rather this keyword vs. average site traffic -- which if I just have
the after period selected I can see these 2 metrics right next to each other.
So “shared” keyword traffic is much higher quality traffic than average traffic. It’s
bouncing at 1/2 the avg. rate and staying on the site over 50% longer. Now there are
other ways to measure traffic quality -- some of them, like using conversion goals,
get you much closer to seeing SEO’s impact on the business. But ATOS and BR are great
ones to start with because they apply to nearly every type of business and they’re easy
to find. So once I have these numbers I just add them to my spreadsheet
[5. Repeat Steps 3 and 4 for Each Keyword] And then I just repeat these last 2 steps
for each of my target keyword themes. And once you get the hang of it it’s a pretty
easy process. So if my next keyword theme “coworking” ... I just switch the contain
filter to “cowork” In this case I get an even more extreme result: only 1 visit
before and 117 after. Quality isn’t as good as “share” but it’s still a lot better
than average. And that’s it - I write down my results and go back and do it again for
the rest of my keywords.
[Bonus Step: Package It]
The last thing I do is compile the results in 1 or 2 graphs. This is not a required step
but it’s one worth doing esp if you’re sharing this other people. You’ve done all
this great analysis -- take the time to present it in a way that others can understand it
and act on it. So here’s the spreadsheet I use to collect
the numbers we just looked at. Now I don’t want to give away the Lab’s entire keyword
list but let’s pretend that I’ve entered these same metrics for the rest of their core
words. Now I can graph it in a way that clearly shows
where our SEO was successful. More importantly -- and this is something I’ll cover more
in another video -- it shows where you should focus your future optimization efforts.
Again this is not a perfect comparison. But it is much more scientific and actionable
than most approaches - esp rankings
[Wrap Up] So that’s the process. Now there are a number
of important caveats and alternatives and further steps in this process that I can’t
go into here. So before using this approach, read our companion article which we link to
at the end of the video. Thanks to Affinity Lab for letting us share
their story. If you’re in the DC area tell your entrepreneur friends about the Lab -
it’s a great community. And if you enjoyed the video, sign up for
our newsletter so you can watch more of these in the future. Thanks!