Apr 27, 2013
Social Media Club NYC Recap: Social Media Measurement
Last Thursday, Social Media Club NYC met to discuss the topic of social media measurement. Doh Young Jung, data scientist at Brandwatch, was one of the speakers at the event. The second speaker was Martin Murtland, vice president of platform management at PR Newswire. The moderator for the evening was Howard Greenstein, president and organizer of SMCNYC.
Q: What is your role in your company?
Murtland: I am responsible for developing the roadmap for a lot of the products. Some interesting research is that 56 percent of brands and agencies are equating the value of their social media activities to their business outcomes. So we need to know how to show businesses the value of what they are doing with their social media activities. I am a firm believer that the key to this is for practitioners to talk the language of business, which isn’t necessarily talking about all the metrics you can have but more about trying to understand how you can link to those metrics with what you are trying to achieve from a business perspective.
Jung: I am part of the analytics team. We do a lot of consulting services with clients, and we try to help them understand social media as well as how to use our tools better. In addition, I do a great deal of reporting for clients when they have specific social media questions.
Q: What are we talking about when we say social media measurement?
Murtland: It goes back to what you are trying to achieve from a business perspective. You can look at it like a marketing funnel which you flip over, and then you have to push your leads through the different areas. And you have to work very hard to get them through. Try to think about what you are doing with your campaigns; what metrics are appropriate in each of those general areas, as well as what you are trying to do inside the marketing funnel. For me, it is important to look at where the industry is going and what companies are doing to create these tools to enable users.
Jung: Our goal is to always deliver relevant content in a timely manner. When we talk about relevance it is about understanding our client’s objectives in terms of the data that they want and knowing when they need that data. We always want to make sure that our tool is easy for the practitioner to use and the reporting is easy to understand. Many of our clients come from PR and marketing agencies, and then we also support their clients. In addition, we have some larger financial clients that use social media monitoring for their product offerings.
Q: Why is social media measurement more difficult than just turning on these tools that you offer and letting them do the work?
Murtland: The software providers know part of the puzzle but it also takes work on behalf of the user to understand what issue they are trying to solve. It is important to know what you want to achieve consistently over time. One of the key things from a measurement perspective is to benchmark yourself. Don’t worry so much about what metric you use in the beginning, but try to benchmark what are you doing — otherwise you will not know what’s having an impact and improving. If you are able to do it well then include in your benchmark some of your competitors. You want to try to create reference points to see how well you are doing. From there you can think about what kind of metrics you can cover and what metrics you should be covering from a business perspective. Then look for an overlap between these two groups of metrics, and that should be the metrics you use.
Q: You (Jung) are a data scientist, so what is the science of what you are doing?
Jung: We deal a lot with numbers. We do want to show the different trends going on with social media data. As companies start to collect this type of data and look into it, the more accurate of a vision they can have of relating it back to their business purposes, such as the their marketing or financial results.
Q: Do you consult with companies about the purpose of the stats they are collecting?
Murtland: We do have a team for that. The first question to ask is: What are you trying to achieve from a business perspective? No metric or tool will resolve your business problem, you have to start by identifying the problem and then let everything else drive it.
Jung: Our starting point for every discussion is helping clients ask the right question. For example, if there is a case where a company is starting with zero awareness about whatever they are releasing then we have to do competitive research. So if they are releasing something on the market that already has competitors, we go into competitive data sets and see how they are doing in the market and then we tell the client what the competitor is doing successfully or wrong. This gives them some type of strategy.
Q: Now that we have established a baseline and know what business goal we are trying to achieve with our social, what’s next?
Murtland: The next step is to understand some kind of cause and effect. It is important to log and record the type of activities you have been doing. You want to show that what you are doing is actually driving the change.
Q: Can you have a tool where you are can both send out your social and measure it?
Murtland: We have a product that is an engagement console where you are able to track some of your activities. Likewise we have different tools for more earned media. You are able to log your activities in there.
Jung: We started out as a monitoring tool, so that is our core focus. We have seen more requests for engagement, and this is an area we want to venture into.
Q: Not all the networks (Facebook, LinkedIn, Tumblr, etc.) make their metrics available, so how do you bring all this stuff together into one global picture that you can start to understand as a marketer?
Murtland: Work out what metrics you can measure and what metrics you should be measuring. The other thing to try to validate is where that data is coming from and what it means. I think there is a lot of jargon and ways to express different things, so try to understand it.
Q: How do you view a single metric vs. combo metrics, because the combo metrics seem more accessible?
Jung: It depends on your business goals. Also different types of clients have different things they are more interested in, so there is no one easy formula. PR agencies are more interested in influencer identification, which is trying to identify whether a tweet from a personal account is different than a tweet from a more influential account. They want to find those Twitter handles that have more influence and impact on social media.
Q: How do you determine what is influential for that particular brand?
Murtland: What is important to me is the contextual influence, so what is the person’s domain and whether they are influencers around that. You can also check if they are an influencer by seeing if their followers are active; look for retweets.
Jung: Our tool can collect historical data as far back as two-and-a-half years. We begin by identifying Twitter handles or any sort of users that mention a relevant brand or marketing campaign topic. We then delve into what they are posting about and look for the topic in their conversation.
Q: How much semantic or sentiment analysis are you doing, and how do you decide if it makes any sense?
Jung: We do have built-in universal sentiment engines and they are based on things like swear words. We are able to customize syntax and understand the language better of certain conversations that have been surrounding positive or negative topics. We can manually change the rules, tweak it, and make sentiment more reliable.
Murtland: There are a couple things you want from a sentiment tool. They are: 1) automated sentiment, looking and analyzing large volumes of content and identifying trends inside it; 2) manually being able to override the scores.
Q: What do we need to do next to tie what we are doing (getting inquiries, selling products, etc.) to some sort of a business metric?
Murtland: You need to start by looking at the peaks and troughs, and try to see if there is a correlation between them. You can try to see the causes and effects that are happening and the correlations, then you can begin understanding and seeing what’s working and not working. Do more of what’s working and less of what’s not working. Repeat and then see the effect.
Jung: As a company becomes increasingly savvy about social data, one thing they can do is set a target to reach. For a lot of PR agencies, the target is often key message penetration. They want to see that a message they crafted is actually being delivered through social media to the audience that they want to reach. An increase in key message penetration has resulted in positive/negative business performance.
Q: How do you keep out confounding data? An example of this was when the “Old Spice Guy” first came out and there was a huge spike in sales, but then someone noted that P&G had a major couponing campaign going on.
Jung: Our entire app is based on Boolean, so if we see a peak we are able to delve into it. We can cut it out and see what the marketing volume was about as well as the coupon conversation. Then we look at the relationship there, and if we see both things increasing then that can mean both have worked.
You can watch a video of the event here:
(If you’re unable to view the video on this page, please go to: youtu.be/TXGg6rXLMcs)
Whether you’re a reporter, blogger, author or other content creator, ProfNet can help you with your search for expert sources. You can send a query to tens of thousands of experts and PR agents, search the more than 60,000 profiles on ProfNet Connect, or get timely experts and story ideas by email — all for free! Need help getting started? Email us at firstname.lastname@example.org.