How I measure my Social Network ROI using Web Analytics

posted Jun 30, 2011, 10:48 AM by christy bergman   [ updated Aug 4, 2011, 10:52 AM ]
Since there's a lot of interest in social networking these days, I thought I'd share an example how I use web analytics to measure my social network effectiveness.  I have a tiny business, The Secret Wine Shop, with a website (hosted (in the cloud) by Google Sites).  I have Google Analytics set up on my Google Sites.  For social networks, I've got the usual array: Twitter, Facebook, Yelp, Chowhound, Google Maps Place.

The first step was to define a couple monetary goals for my website and customize Google Analytics to measure them.  For me, the goals were 1) sell Wine Club memberships, 2) sell special event tickets, 3) contact us.  Next, I decide to analyze what's going on with one of those goals, in this case a special event.  The screen shot below shows Unique Visitors by page, and it just so happens the Goal Conversion Page for purchasing tickets to my upcoming special event was the top traffic site.


Next, I drill down into that goal conversion page.  Using a custom report with custom dimensions, I can see the top referring traffic and bounce rates.  Given the numbers I can estimate (next step is put this calculated field directly in my custom report) the number of truly interested people in my event as #UniqueVisits * (1-BounceRate).  Direct traffic (people who received the link directly in their non-browser email such as Outlook & clicked on it, I know this because I administered the campaign) was the highest source of truly interested people.  Average time on page for all 3 top referrers is close enough to the average 2'46" I'll count them all equal after bounce rate. To give an economic value to this, I use avg $ price of ticket * #trulyInterestedPeople. 

Referrer        #unique visits  BounceRate #Truly Interested Peoople   Value  %Total
Direct           188               85.65%         27                                 $972      61%
Facebook        32               82.35           5.6                                $202      13%
Twitter           23               78.26           5.0                                $180      11%
Google            18               62.50           6.7                                $241      15%


So let me drill down into the Facebook traffic and see what's going on.  In the next screen shot, you see I had a distinct (small because my business is small) peak in my facebook traffic on May 16.  The 2nd peak was on May 23, the day of my event, so I'm ignoring that, social chatter that's too late to be of any use for this particular campaign.


So now comes the juicy part.  Now I know exactly what day facebook activity peaked.  I ask myself what did I (or someone else) do on facebook that day that triggered a peak in interest?  My search is narrowed now to just 1 day and just 1 campaign.  I start by looking at my own activity on that day, May 16th (fortunately facebook time stamps are in the same timezone as google analytics).  I made one post just before midnight before that day, followed up during that day with comments.


But generally my page isn't such a big deal. I'm sure there's more to it than this.  I suspect it's Randall Grahm who commented on my post that day.  So next, I click on his profile.  This is a harder task, he's a much more prolific communicator than I am.  Fortunately I'm only looking for activity related to my event on 1 specific day.  After much scrolling down, I see something there on his page!  I forgot that I made the same post at the same time on both our Walls.  I also take a look at who "liked" it, among them the one with the most followers is Meg Houston Maker.  I look at her wall to see if she also made a comment, I don't see one.  Just to make sure, I do the facebook Search > "Secret Wine Shop" > posts by everyone.  Nada, no one else talked about it.  So I give myself a pat on the back, I actually did something to trigger that traffic. 

So I make the following conclusion about effectively using Facebook to promote an event and measuring its impact.  On the Google Analytics side, make sure you're set up to measure the effects.  On facebook:
  • Give relevant info about your event
  • URL link to your conversion page
  • ...and also post on that influential person's wall!
I wouldn't have learned this last tip without the aid of Web Analytics!  It's the posting on someone else's wall that is the secret sauce!



The next step is repeat this process, segmenting by Twitter traffic.  Here I see 2 peaks: May 12, May 16.  I also see a smaller spike in activity on May 20-21.


Again, I start by looking at my own tweets on May 12th and May 16th.  I see I made a few tweets about the event and replied to others about the event, but I'm a small business with a small twitter following. I suspect it's not solely my amazing tweets bringing in more traffic on those days. 


Twitter gives us some handy tools: @mentions and retweets.  I look to see what mentions I have on those 2 days, and lo and behold, I have mentions and a retweet from Randall Grahm and his crew.


The insight on twitter is similar to what I saw on facebook - on twitter it's important to @mention an influential person/group of people; otherwise, if you're small like me your tweets don't have much impact on their own to your bottom line.

Bottom line: There's more potential ROI if I focus on improving my direct email campaigns.  Social media is worth something, just not as much. So when I do spend time on social media for a campaign, I need to be more efficent.

What I've shared is just one iteration.  The good thing about all this is I now have a way to see what works in social media & make myself follow my own "best practices", which in turn may lead to revised "best practices", and so on, to be repeated.



Comments