Seth DuncanThis is David Geddes, chair of the Institute for Public Relations Commission on Measurement and Evaluation.

Today I am talking with Seth Duncan, Research Director at WCG, a global communications company providing integrated marketing, public relations, creative and interactive services.

Seth, to begin, why should web analytics be part of the toolkit of PR professionals today?

Web analytics are important for PR professionals because of two quickly accelerating trends. First, people are starting to rely heavily on their mobile devices for accessing news and information. This means that readers leave a traceable digital footprint every time they read a New York Times story, a Mashableblog post, a review on Yelp, or tips on Foursquare.

Second, people now interact with brands online at multiple points of the sales funnel, buying process, or communications life-cycle (whichever model resonates most). So, audiences are now leaving a traceable digital footprint when they interact with brands as well. The convergence of these trends is exciting for the PR industry because web analytics make it possible to make direct measurement connections between digital news consumption, brand engagement on social networks and, in some cases, actual sales through an e-commerce. This ability to bridge the gap between online media outputs and outcomes using web analytics is already an important part of digital planning and evaluation for PR professionals and I expect it will become ubiquitous over the next few years.

How can companies that do not sell products through their website – or are not selling any products at all – use web analytics to optimize PR campaigns?

Web analytics data has the clearest applications for marketing communications, especially for e-commerce. However, there are plenty of other target audience behaviors, beyond sales, that can be tracked using web analytics. If your organization does not sell products through its website, you can try to measure sales leads through alert subscriptions, white paper downloads, or clicks on a “contact us” button.

Or, if you are trying to build an audience for your corporate blog, you might measure the number of return visitors each month or the length of time people are spending at the blogs. If an organization is trying to facilitate any sort of online interaction with customers, analysts, journalists, or employees, there will always be a valid web metric for measuring the target behavior. You might just have to be a bit more creative in developing the metric if your audience’s target behavior does not occur directly on your website.

Let’s go to a hot topic. What is engagement?

That’s both hot and difficult to answer. Like other concepts that get floated around PR and marketing (e.g., influence), engagement can mean many things to different people. When I hear or say “engagement” in PR or marketing circles, I interpret it as a broad term that is synonymous with “behavior towards the organization or its immediate audience”, usually pre- or post-purchase.

I like to keep our industry-level definition of engagement broad because I think people can engage with an organization in a variety of ways. People might re-Tweet a story about an organization in mainstream media. I would consider that engagement because someone has behaved in a way that demonstrates interest in the organization and wants to share information about the organization with his or her online social network. In other cases, someone might engage directly with the organization by either Tweeting at them, visiting their website, or physically walking into a store or office and talking to a company representative. People can engage with organizations in limitless ways.

So if you see “engagement” in a broad sense, what does that mean for how we should measure engagement?

Since I think that engagement is a very broad term for our industry, I think that PR professionals need to be painstakingly specific when they define engagement for their organization or campaign. It’s not possible to operationally define measurement as “any behavior related to an organization” and start collecting data. The operational definition for engagement must be specific to organizational and campaign goals. This is a simplistic example. If you have a campaign goal of attracting job applicants from elite universities, you might operationally define engagement as number of repeat visitors on your organizations’ blog that originated from the selected schools (which you obtain using web analytics). Defining engagement in a way that is relevant and focused on your goals will save you a lot of potential measurement headaches and budget. If the engagement metrics are focused, they’ll also make a great deal more sense to others in your organization.

It is also important to remember that the means of engagement will inevitably change over time, so engagement-related metrics should be as channel-agnostic as possible. It’s easy to fall into a trap of using the metrics that are freely available in social media channels, such as Facebook likes, Twitter followers, YouTube views, or even Klout scores to measure engagement. But, there’s a good chance that these metrics will undergo significant revisions (or might not exist) in several years. Because a target audience can interact with a brand in almost limitless ways through a wide-number of channels, you want to find metrics that are common to as many of these channels as possible. If you can do that, you’ll be able to benchmark engagement over time and can avoid painful social media measurement program revisions every 6 months.

Some people contend that engagement should be considered an outcome metric. This is what social media departments and agencies are expected to do: produce compelling social media content that draws audiences to social media channels and builds relationships. Others, and I consider myself among them, consider engagement to be a precursor to changes in awareness, attitudes, and opinions, thus it is an output. It is talking at people, in the same way as TV news talks at people. Or as Avinash Kaushik put it, “engagement is not a metric, it is an excuse.” So, Seth, is engagement an output, an outcome, or something else?

I’m squarely within the camp that says engagement is an output and never an outcome. Ultimately, it doesn’t matter if an audience is very engaged with branded content if the engagement doesn’t directly or indirectly lead to increased revenue, cost-savings, donations, grant support, or anything that keeps the organization running.

If engagement is an output, how important is it to measure engagement?

I do think that measuring engagement is incredibly important, though. Like you, I tend to think of engagement as something that a target audience does with content that a brand or a third party produces (sharing it on social networks, commenting on it, producing video replies, etc.), which is an important intermediary step between content creating and meeting campaign goals. Measuring engagement is an essential part of campaign planning because, if you create content and find that it leads to an increase in sales but never measure how customers engaged with the content, it might be difficult to reproduce or improve upon that campaign again. You won’t know what content was shared, where it was shared, or by whom. So, although engagement is not an outcome, it’s essential to measure it so you can optimize your content and social media strategy the next time around.

This is probably a controversial perspective, but I do believe there are situations when measuring engagement-related metrics in place of outcomes is excusable. For some organizations, it is difficult, time consuming, and expensive to connect business outcomes such as sales to social media campaign efforts. For these organizations, it might only be possible to measure the business value of a social media on an irregular basis—maybe once a year or so. If these organizations can, at least on occasion, demonstrate a clear correlation between engagement-related metrics and outcome metrics, I think it is acceptable to use the engagement metrics as a stand-in for outcome metrics.

For instance, if a company uses panel-based research to find that one Facebook Like is associated with $10 in revenue during campaigns A and B, but cannot obtain budget to use the panel for the smaller (but similar) campaign C, I would argue that measuring Facebook Likes as an outcome proxy might be acceptable for campaign C. I would never call a Facebook Like an outcome but, if you can demonstrate that it predicts a real outcome, I think that it can act as a temporary stand-in.

I expect some of your readers will take exception to the idea that outcomes can be measured in place of outputs. But, I just don’t think that’s feasible in situations where proper ROI-type measurement might cost more than one third of a campaign’s budget. As long as you are completely transparent about your research methodology and do not try to pass engagement as value to stakeholders, I think that this is a perfectly reasonable alternative (but only if you’ve shown that engagement metrics correlate with outcome metrics in the past).

What are the best ways for PR professionals to keep up on web analytics?

You’ve already mentioned Avinash Kaushik, whose books should be required reading for web analytics beginners. These books are for a non-technical audience and do a great job explaining the importance of creative, customized- approaches in measuring online behavior.

Another great online resource is the Web Analytics Demystified site, which is a web analytics consultancy run by Eric Peterson, John Lovett, and Adam Greco. They have an informative blog as well as a few popular e-books. Web Analytics Demystified also hosts regular free events called Web Analytics Wednesdays, where web analytics professionals get together for drinks and networking at cities across the world.

Once PR professionals have a bit of knowledge and experience with web analytics under their belt, it’s worth following the blogs for the major web analytics vendors (e.g., Google and Adobe’s Omniture) to keep track of the latest technologies. Many of the posts will be for a technical audience, but you’ll see all of the big announcements for advancements that are relevant to PR (e.g., Radian6 integrations, sentiment tracking, etc.)

The eMetrics conferences are another great place to learn about web analytics and meet some of the leading practitioners in the area. The speakers and attendees are a good mix of vendors, consultants, marketers and client-side analysts. Much like the vendor blogs, this is a really good place to learn about the direction where web analytics and digital measurement is headed.

The best resource that PR professionals have is their organization’s own web or digital marketing team. These are usually the people who select the web analytics vendors, tag all of your organizations digital properties so they can be tracked, evaluate online ad effectiveness, and manage search engine marketing. They have a lot of digital measurement knowledge and, from my experience, are always happy to provide data and support to PR professionals trying to integrate web analytics into their toolkit. If you happen to work at a large organization, there’s a good chance that these people also get to beta test new offerings from web analytics vendors—including the occasional social media analysis tool. If you haven’t already, it’s definitely worth talking to them about how web analytics might fit into your PR measurement program. In one conversation, you’ll probably learn that your organization is already collecting 5-10 PR-related metrics that you didn’t know about.

Thanks for talking with us today. Please stop by for future installments of Five Minute Conversations, or read part interviews at

Share this:

Heidy Modarelli handles Growth & Marketing for IPR. She has previously written for Entrepreneur, TechCrunch, The Next Web, and VentureBeat.
Follow on Twitter

2 thoughts on “Five Minutes with… Seth Duncan

  1. Hi Fraser:

    You make a good point that there shouldn’t be a purely dichotomous output vs. outcome measurement taxonomy. I wasn’t familiar with the “outtake” definition that you shared here, but I think “engagement” fits nicely within that definition. I was looking for a concept like outtakes when I described engagement as an intermediary step between content creation and and meeting a campaign goal.

    Thanks for your input.

  2. Engagement is neither an output nor an outcome measure. It is audience action between the receipt of a communication activity output and the communication effect outcome (changes in … ultimately audience behaviour).

    The audience action is called an outtake. The Dictionary of PR Measurement and Evaluation defines it thus:
    Outtake – m. measurement of what audiences have understood and/or heeded and/or responded to a communication product’s call to seek further information from PR messages prior to measuring an outcome; audience reaction to the receipt of a communication product, including favorability of the product, recall and retention of the message embedded in the product, and whether the audience heeded or responded to a call for information
    or action within the message; s. sometimes used as an outcome
    serving as a dependent variable in research.

    Seth’s examples of engagement exactly fit the definition of an outtake measure. He said: “People might re-Tweet a story about an organization in mainstream media. I would consider that engagement because someone has behaved in a way that demonstrates interest in the organization and wants to share information about the organization with his or her online social network. In other cases, someone might engage directly with the organization by either Tweeting at them, visiting their website, or physically walking into a store or office and talking to a company representative. People can engage with organizations in limitless ways.”

    Engagement activities are outtake measures.


Leave a Reply

Your email address will not be published. Required fields are marked *