How to measure if social media marketing is working for you

Written by

Advertising, Marketing, Metrics, Social media

Warning: this content is older than 365 days. It may be out of date and no longer relevant.

If you’ve got Google Analytics goals and goal values set up and working correctly on your web site, there’s a fast, relatively easy way to tell if social media marketing is working for you in Google Analytics thanks to a new feature called Social Value.

Go into Google Analytics and find Traffic Sources, Social, Overview:

Here we’re looking at the Social Value graph. This graph shows you overall conversions on your web site, conversions that were influenced by social, and conversions where the last-touch was social. The question remains, however: how do you know social is working for you? Just looking at this chart isn’t necessarily all that helpful.

Recall that on any given site, you want conversions to be coming from four major sources: direct, referral, campaign, and search. We discussed this previously in a balanced pie. As such, what you want to see here is that balance reflected in the Social Value. Let’s look at some common scenarios that will give you a quick health check of your social media marketing. You’ll note in the chart below that red corresponds to last-touch social conversions, yellow corresponds to assisted conversions, and green corresponds to your overall conversions.

Download this chart as a PDF here.

Balanced: Great job. Social is responsible for 25% of your last-touch conversions, which means you’re asking the right amount from the channel and presumably giving as much. It’s also responsible for 25% additional conversions in the form of assisted conversions, which means that your social efforts are bolstering search, direct, and campaigns. People recognize or know to look for you from your social media marketing. Good work.

Socially Broke: Whatever you’re doing, it’s not working. If you’re not even getting 5% of your conversions and 10% of your assists from social, you’re not participating enough. Add focus, add calls to action, and give more to your community in order to start deriving benefit.

Over-Reliant Social: On the surface, your social media marketing seems like it’s on fire. Conversions are rolling in, business is being done, and nearly every single sale you’re making feels like it’s coming from social. The hidden danger here is that you’re over-reliant on the channel, and if you’ve got one or two people working it for you, you’re basically at their mercy for your business. If they quit or your accounts get banned/changed/moved, you’re in a world of trouble.

These three scenarios give you the overall health of your social media marketing. There are two additional considerations to look at, the ratios between assisted and last-touch.

Chatty: When the ratio of assisted to last-touch is greater than 2:1, social is working for you in the sense that it’s helping push conversions, but you might not be asking frequently enough. This isn’t necessarily a bad thing but it’s something to think about – you could ask a little more frequently.

The I in Team: This is a case where assisted conversion is less than 2:1 with last-touch. You could be asking too often and/or not participating enough as a human being in your community – whatever the case may be, social isn’t powering enough other conversions from search, direct, and campaigns.

The Social Value chart isn’t the end word in social media marketing or measurement, particularly if you’ve got a scenario where direct revenue value isn’t an end goal. That said, if you can ascribe any kind of value whatsoever to your online goals, the chart will give you a good starting place to ask more questions and dig more deeply into your social media efforts.

What do you think of this assessment? Does it accurately correlated to other ways you’re measuring your social media marketing?

Comments

13 responses to “How to measure if social media marketing is working for you”

April 16, 2012

Ashley Porter

This is definitely worth implementing…I must say that although I have Google Analytics installed I hardly know much about this Social Value tool and will have to check into it asap.
Thanks for sharing Christopher

Reply
1. April 16, 2012
  
  Christopher S. Penn
  
  Glad it was helpful!
  
  Reply
April 16, 2012

Edward Smith

Thanks for that, this is real cutting edge stuff. Thanks again, Edward Smith.

Reply
April 17, 2012

AdamSinger

Love the thinking here Christopher, glad you’re diving into the new reports …and even happier you have an opinion on what effective results look like.

Reply
April 17, 2012

Gnosis Media Group

I see what you’re getting at here, Christopher, but I think your methodology of measuring Social Value is a bit narrow. For example, we routinely generate appointment signups from Twitter to our online appointment setter. We use Genbook, and obviously we cannot install Google Analytics onto the Genbook CMS in order to track a signup from Twitter to Genbook. This wouldn’t be reported on your model at all. Yet, it is still a conversion generated by social, which also has a definite value.

We also utilize social for customer service. Answering a customer’s question via Twitter or Facebook, sending them away more satisfied than they were when they first came, provides a definite business value. However, no conversion in the conventional sense would have taken place. Again, your model doesn’t measure this at all.

Or, how about this one. As a PR firm, we sometimes run into journalists on Twitter who cover a beat of one of our clients. We develop a relationship with that journalist, which over time leads to their writing a story about one of our clients’ products. The story creates significant buzz for the client, and produces incremental sales as well as increased brand awareness. Again, not something straightforward enough to be tracked with your model.

In short, there are many uses of social, which deliver tangible business value, which cannot really be tracked on your model. I think you offer some interesting insight but your model here is much too simplistic.

Eric Bryant, Director
Gnosis Media Group

Reply
1. April 18, 2012
  
  Christopher S. Penn
  
  I probably should have stated that this is within digital marketing. Also, you could do a valuation with Genbook as long as it was assigned a goal and goal value, even if it was just an event-based click. You’d have to put an intermediate URL in between that would record the data, but it would be possible.
  
  Where you run into trouble is outside the clickstream, which you point out. It’s outside the limitations of Google Analytics and any web analytics package.
  
  Thanks for stopping by.
  
  Reply
April 17, 2012

Anand Patel

Thanks for breaking down the different possible scenarios! I am pretty excited about this new feature by Google because it can really help show you how effective your social media efforts are. It will also be interesting to see how much further Google tries to push social analytics. Either way, thanks for the graphs…I will definitely use those to evaluate my social efforts.

Reply
April 18, 2012

Tim Leighton-Boyce

This is a great introduction and I really like the way you break the different levels of interaction into chunks. But I also think it’s important to stress that if people are using multiple devices and not actually clicking through to your site then GA will not make the connection.

That’s a very common scenario on sites I work on. People are very active on the Facebook page for example and are probably using mobile devices for that. They seldom click through on actual links from FB to the site, probably because they’re in a mobile situation and not in the right context to buy.

On a separate occasion (judging from circumstantial evidence from specific promotions) they then visit again using a non-mobile browser via something like navigational search. Social was an ‘assist’ but will not be reported as such.

So these reports are great for trending against themselves — are new activities shifting us from ‘socially broke’ to ‘chatty’ for example — but they are samples. They still can’t give an accurate picture of the total value of social.

Reply
1. April 18, 2012
  
  Christopher S. Penn
  
  Quite correct – and it also doesn’t take data outside the clickstream into account either, but nothing short of primary research can.
  
  Thanks for stopping by.
  
  Reply
April 18, 2012

Brian Clifton

Nice post Christopher. I can’t help thinking why the other mediums (channels) are not treated in the same way by Google Analytics….

“Social” is just an aggregate of specific referral sites – you can achieve the same using a profile filter or advanced segment. The new social reports that show its impact would be just as valuable for organic, paid, referral, direct, email, etc…

May be its on Google’s list for enhancements to the Traffic Sources area 😉

Best regards, Brian
Author, Advanced Web Metrics with Google Analytics

Reply
1. April 18, 2012
  
  Christopher S. Penn
  
  Unquestionably, Brian. I’ve used my own regex for an advanced traffic segment for quite some time now, and while it’s easy to get these numbers separately, for the average user, having the three-circle chart is what sets it apart from yet another segment of numbers.
  
  Creating that kind of chart would be very helpful for all the other channels and advanced segments. I did the slide in Keynote, and basically did the proportions scaled to the data. Manual, but better than nothing for now.
  
  Thanks for stopping by.
  
  Reply
April 19, 2012

aaroneden

I haven’t checked out this new feature of Google Analytics so I’m saving your tips here, thanks! I think there’s just too much data these days to interpret correctly and finding that one metric that you can focus on to help you reach your goals has become tougher.

Reply
July 18, 2012

Moses

Thanks for this!! I use Bitly for my social media posts and based on your graph, I’m Socially Broke. Does GA recognize that a referral is coming from FB even though it went through Bitly first? From what I gather online, GA doesn’t recognize the short link unless I add query parameters. What are your thoughts?

Reply