This article is Part 10 of an ongoing series, “Rescuing Ourselves from Social Media Measurement Dinosaur-dom,” based on Angela Jeffrey’s paper Social Media Measurement: a Step by Step Approach … using the AMEC Valid Metrics Framework. The complete paper is a free download and includes much more detail than this series does.
Finally, we focus on the most exciting part of social media measurement, “Target Audience Effect” or “Outcomes.” This post covers the second half of Step Five of my Eight-Step Social Media Measurement Process: Choose Tools and Benchmark using the AMEC Valid Metrics Framework, (see the first post here). As always, for the whole reference I suggest you download the complete paper, which covers the entire Process. As with the last installment, this post is full of tools from Appendix D of the paper.
As a reminder, here’s my entire Eight-Step Social Media Measurement Process:
- Identify organizational and departmental goals.
- Research stakeholders for each and prioritize.
- Set specific objectives for each prioritized stakeholder group.
- Set social media Key Performance Indicators (KPIs) against each stakeholder objective.
- Choose tools and benchmark (using the AMEC Matrix).
- Analyze the results and compare to costs.
- Present to management.
- Measure continuously and improve performance.
Now, on with the last part of Step 5…
As you’ll recall, we are working our way through the AMEC Valid Metrics Framework (for an introduction to the Framework, see this post) for the ultimate purpose of linking our PR Activities and Intermediary Effects with Target Audience Effect (see the left side of the sample below). In today’s networked world, this is easier than it has ever been, with rich toolsets available through both the communications and marketing departments. The complete paper covers several ways of doing this, including surveys, e-commerce measures, cross-silo measurement, statistical modeling, and web analytics. To keep this post to a manageable size, today I will just touch on two of my favorites: Statistical modeling and web analytics.
The AMEC Valid Metrics Framework for Social/Community Engagement
One way to link Public Relations Activity and Intermediary Effects to Target Audience Effect is through statistical modeling, such as market mix models, advanced statistics, and even simple correlations.
If you have access to a market mix model, provide the program’s KPIs to the modeler to determine whether there is a causal relationship between those metrics and business objectives. Market mix models are extremely sophisticated and study all parts of a marketing program to determine cause and effect. They are also very expensive, and social media reach numbers may not be high enough to register amidst all the other marketing metrics. For more information, download Representing PR in the Marketing Mix: A Study on PR Variables in Marketing Mix Modeling, by Brian G. Smith, (and winner of the 2007 Ketchum Award).
A more accessible method is to use correlation (Pearson product-moment correlation), which is a measure of the association between two variables X and Y, giving a value between +1 and −1 inclusive. This correlation coefficient (sometimes referred to as Pearson’s r) can be calculated right in Excel (see below), or with the help of your marketing department.
If campaign objectives were to increase sales, be sure to collect sales and revenue data from the finance or sales departments along with the average sales price, so hard Target Audience Effect data can be correlated against it. While correlations are fairly simple, it is best to seek help from an expert, since factoring in lead-time to lag ratios can be challenging. In other words, Target Audience Activity should result after Activity and Intermediary Effects have been conducted, and often reflect a product’s sales cycle in overall timing. For example, if a campaign is run in January, Intermediary Effects may be seen immediately, but Target Audience Effects may not show up until June if the sales cycle is six months.
A very simplified example of a correlation is shown below. In this case, the goal is to see if there is a relationship between a campaign’s share of voice and sales leads.
For a positive relationship to exist, a correlation should have an r value of at least .7 out of a possible perfect 1.0. The example below shows a very high r value of .80, which indicates that the number of leads are increasing as the campaign’s share of voice is increasing. Again, this does not prove that share of voice is driving leads (causality), but does indicate that the two variables are related (a correlation).
Easy correlations calculation with Excel: Simple correlations can be calculated in Excel using one of these commands:
=CORREL or =PEARSON
Step 1: Set up an Excel spreadsheet using the first row for time periods, the second for media analysis scores (ideally share of voice), and the third for the desired business results:
Step 2: In an empty cell in the spreadsheet, enter the cell numbers of the starting and ending values in each row into the formulas above, like this:
Step 3: Hit “enter” and the correlation appears, in this case an r = .80.
Without a doubt, web analytics is one of the most exciting new frontiers in measurement, since tying PR Activity to both Intermediary Effects and Target Audience Effects (see the left side of the Valid Metrics table above) can be done seamlessly. The PR industry has never had such an opportunity to track its work so easily against KPIs at such little cost.
However, it is critical to remember that online conversation accounts for a relatively small percentage of a target audience, so web analytics do not tell the full story. Regardless, spend half a day and learn the ins and outs of a free system like Google Analytics. An excellent book to read is Michael Miller’s Sams Teach Yourself Google Analytics in 10 Minutes. The book is comprised of a series of 10-minute lessons which can be mastered in less than a day. Additionally, Google has developed a great series of how-to videos for marketing agencies that explain the new features in Google Analytics. For additional excellent and in-depth resources into the many marketing uses of web analytics, see the following books by Avinash Kaushik: Web Analytics: An Hour a Day and Web Analytics 2.0.
Web analytics can measure both hard and soft goals. For some companies, sales or e-commerce is not an option, so they have to look at indications of interest and loyalty. The most important metrics will be conversion, which is defined in Google Analytics by the user as any web-trackable activity one wants a prospect or customer to engage in. This may be as simple as a click-through to a particular web page, registration to a special community, etc.
There are endless metrics available to look at in web analytics platforms, but many suffer from being too tactical. Avinash Kaushik says many of them – visits, page views, time on site, impressions, clicks, emails sent, followers, likes, video views, etc. –simply imply that bigger is better. He goes on to say that it’s more important to view metrics that span several sessions to look at a potential target’s true interest over time. Tips from two posts on his Occam’s Razor blog, I Got No Ecommerce. How Do I Measure Success? (Kaushik, 2007a) and Your Web Metrics: Super Lame or Super Awesome? (Kaushik, 2011b) are included in Appendix D in the complete paper.
Don’t forget to check out the other Target Audience Effect measures (Surveys, e-Commerce and Cross-Silo) from Appendix D of the complete paper.
Meanwhile, if you have questions or measurement needs, I’d love to speak with you! Send me a note at firstname.lastname@example.org.
Latest posts by Angela Jeffrey (see all)
- Advertising Effectiveness: Why Are So Many Likable Ads Ineffective? - May 3, 2016
- The Truth About the Top Five 2016 Super Bowl Ads - March 8, 2016
- The Microsoft-Apple “Bear Hug:” Winner or Loser for Microsoft? - January 16, 2016