4 Ways to Link Facebook Posts to Business Outcomes
One of the most frustrating things about Facebook is to understand the potential of the network and not to be able to tie network activity back to business outcomes.
We all understand the unique massiveness that is Facebook:
- 936 million average daily active users
- 798 million average mobile daily active users
- 1.44 billion monthly active users
- 1.25 billion mobile monthly active users
Nearly everyone uses Facebook for some purpose and there are many different tactics to capitalize on that audience. The challenge has always been to (meaningfully) measure the impact of Facebook activity to your business objectives.
What I want to do in this post is explore four ways that you can link Facebook posts to business outcomes.
1. Referral Codes / Referral Pages
I know what you’re probably thinking: referral codes and pages are pretty rudimentary solutions to an analytics problem. They’re used on radio programs and podcasts, but have been around forever… and surely there are much better technological solutions (multi-channel attribution models, anyone?). Let’s explore what they are and how they can help to measure Facebook’s impact to a business.
Referral codes are unique codes that attribute referrals to a particular source (usually for a discount of some sort). An example might be to use the referral code “cision” to get a 10 percent discount from a particular vendor. Referral pages are unique pages that are linked to from referral sources. An example might be a unique Facebook landing page. There are all sorts of variations that you could do from these tactics, but in a broad sense that’s what they are.
The ubiquity of these tactics are probably indicative of their effectiveness. A helpful way to think of why these are such effective is to think about why we we want to measure things. In a recent talk (embedded below), Andrew Gelman, Director of the Applied Statistics Center at Columbia University, discussed the importance of post stratification to gain better understanding of why statistics give us certain results. He uses national election data and shows the importance of party affiliation, age, and race to understand polling data.
Referral codes and pages give us a means to stratify conversions by a unique source (in this case Facebook in some capacity). Rather than looking at overall sales and estimating the effect of Facebook on these, referral mechanisms give us a tangible measure of the effectiveness of a specific piece or group of content on a channel. Measurement of this kind is oftentimes referred to as “last-touch attribution.” I’ll touch on other attribution models a little further on.
2. Facebook Conversion Pixels
Facebook conversion pixels are intended for retargeting customers that visited your website with Facebook Ads. For example, if a person reads a piece of content about widgets on your website, a conversion pixel would fire to Facebook tagging the user for any of six activities that they performed on your website:
Not only can you can see that this data reported back to you in your Facebook Analytics but Paolo Margari writes that the Facebook Conversion Pixel script can be implemented using a Google Tag Manager Custom HTML tag, which means that you should be able to capture the Facebook Conversion Pixel Data with Google Tag Manager as well.
One caveat to this data is that Facebook’s conversion data is observably higher than Google Analytics, a fact that marketer Jon Loomer attributes to different methods of collection. Incidentally, this is a phenomenon that I’ve experienced with server hits where Google Analytics filters far more out than other analytics tools do.
In any event, Facebook Conversion Pixels could be a very useful integrated marketing tool to attribute sales or conversion events back to different communications events.
3. URL shorteners
URL shorteners have been around forever – the advantages for a Twitter post might be obvious, and for a Facebook post less so. But a tool like this could help to track content at a more granular level than you are currently.
Of course Facebook has its own native link shortening service, which probably won’t be as useful to you as an external one (meaning that you can get more specific data from a third party). I’m going to use bit.ly as an example (although there are plenty of newer, shinier alternatives). I like bit.ly primarily because you can create multiple links to the same content as see the individual click-through and the aggregate click-through. If you create a link through Google’s goo.gl service, shortlinks for every unique page are always the same.
By using bitly links to post to Facebook, you can measure click-through by post instead of click-through by content. You can do A/B testing rather simply as well as other experiments.
The downside to using URL shorteners is that less people click them – this is a survival mechanism for social users who have probably been duped multiple times in the past. (I was scammed by a company on Facebook a couple of weeks ago.) If URL shortening is a helpful analytics tool, it may make sense to use these with representative samples of your customers, rather getting diminished click-through rates from the entire population.
4. Conversion Lift
If you’ve been tantalized by my earlier tease of multi-channel attribution models, we’ve come to that time. Google lists some of the popular attribution models on its support site:
- Last Interaction – 100 percent attribution to last action (this is what I referred to as “last-touch”)
- Last Non-Direct Click – 100 percent attribution to the second-to-the-last action (which is never referred to as “last-touch-doesn’t-count”)
- Last AdWords Click – self-explanatory
- First Interaction – 100 percent attribution to the first action
- Linear – Equal attribution to each touchpoint
- Time Decay – Diminishing attribution for each touchpoint relative to point of sale / conversion
- Position Based – 40 percent to the first action, 40 percent to the last action, 20 percent equally attributed to everything else
There are all kinds of proprietary models to allocate attribution, but you can see the challenge to trying to accurately gauge the impact that each customer touchpoint has to a conversion event. One of the more thoughtful ways to do this is by random sampling and tracking of the behavior of targeted customer populations. You can do this on Facebook.
Facebook calls this sampling “conversion lift” and what it does is uses the Conversion Pixel tool to track data and randomly assigns two sample populations (a test group and control group). The test group sees ads, the control group doesn’t and the impact of the ad is measured by the difference in conversion rates between the two populations. Of course this is self-serving to Facebook, but is also rather thoughtful – they understand that they need to show their value to businesses.
Attribution for any PR activity is hard (I assume I’m preaching to the choir on this point). Granular attribution for Facebook activity is a pretty big challenge, and metrics from the native Facebook analytics dashboard aren’t as useful as they could be to show Facebook’s value to PR practitioners.
Between referral codes, Conversion Pixels, URL shorteners and Conversion Lift, hopefully I gave you an idea of some tools that can help you to find the measurement that you need to properly attribute Facebook activity to your business objectives. And remember…..
“If your only tool is a hammer, every problem looks like a nail.” – Abraham Maslow
Communications Best Practices
Get the latest updates on PR, communications and marketing best practices.
Cision Product News
Keep up with everything Cision. Check here for the most current product news.
Thought leadership and communications strategy for the C-suite written by the C-suite.
A blog for and about the media featuring trends, tips, tools, media moves and more.