Photo by Gavin Whitner
Social listening programs offer many benefits beyond reputation monitoring – it opens up a world of unstructured, online conversations that give brands new avenues for marketing in a way that is personal and effective. It also provides valuable research data that can be used when building a marketing strategy, developing new products and services, or breaking into a new market.
One concern that causes hesitation in setting up a program is the initial setup itself – if a search is not set up effectively, it will create a lot of noise and unnecessary results, which can be frustrating and overwhelming. Fortunately, there are several ways in which a search can be set up to maximize impact and quality of the results.
Step 1: Start SMART
A successful listening program starts with strategic goals. Before you begin, make sure your goal for listening is SMART – specific, measurable, actionable, relevant, and time specific.
As an example, a listening program set out to find out what people like about ice cream won’t work. It’s too broad and will generate such a high volume of content that it will be overwhelming and not at all useful.
A better option might be, “Do people prefer ice cream in a cup, plain cone, or waffle cone?”
Social listening should answer a question that falls in line with your overall strategy and provide information relevant to your needs.
Step 2: Research
Brainstorm for what you DON’T want. Make a list of terms you know will clog up your search. For example, if you’re running a listening program to get consumer insight for a restaurant named Coopers’ Hawk Winery, it’s easy to come up with some initial exclude terms, such as birds, aviation, coop, eggs, vineyards, etc.
What else is out there? Run a Google search on the terms you anticipate using. Take a look at the first few pages of results to see what other terms may be included that aren’t relevant. Another place to look is at the very bottom of the first page of results, under “related searches”. Make a note of these terms to use as exclude phrases when you set up your search.
Another thing you may find during this step is abbreviations or nicknames you didn’t know existed. For example, Cypress Cove, a local waterpark, discovered that many guests referred to it as “the Cove.” This was added for additional result capture.
Step 3: Set up the query
Become BFF’s with Boolean. Boolean queries used to make me nervous – it was like learning a new language. Spending some time learning about the capabilities that are available can make life a lot easier. There are many great options to make searches not only more specific but more relevant when it comes to results. A great Boolean guide can be found here.
Exclude sites & types of content. Say you’re conducting a query to learn more about why people shop online for a specific product. Results may likely produce news articles and blogs talking about this very topic. While they can be useful in gaining background research content, it’s not what you’re looking for. With most listening platforms you can easily exclude news and blog sites – highly recommended, especially for “hot topic” research queries.
Step 4: Initial Run
Now the hard work is done and it’s time to run the initial search. You’re close to the finish line, but not quite there yet.
After the initial run, you’ll want to see what the results actually turn up. You may find that there are still irrelevant results; while they cannot be fully eliminated, taking some time to review your initial search results will let you do one last pass through to refine before it’s ready for analysis.
Step 5: Quick review & removal
Outside of looking at every result, use the basic platform analytics to find clues to determine if there are still outliers that need to be excluded.
Word clouds/themes. Since word clouds produce common words/phrases found in the results, this is a good place to find irrelevant results, this is a good place to start. Many platforms also generate clouds by hashtag, which can prove useful as well. Make note of items found here for exclusion.
Influencer analytics. Similar to excluding news sites, there may be an influencer identified, whether it’s a blogger or news aggregator that publishes a lot of content on a topic but isn’t relevant to the question you’re answering. Make a note of their name, social handle, and/or website to include in the exclude terms.
Now, take the information and go back to the query setup to add these sites, terms, and phrases to the exclude items section.
Step 6: Revise and retest, rinse & repeat
Once the last step is complete, it’s time to do a final run. At this point, the results will likely be as clean as they can be. Test to be sure there aren’t any new irrelevant results showing up, and then run the program.
It’s good practice to review results on a regular basis to make sure there are no other new outliers/irrelevant results popping up. While they can’t be avoided all together, looking for irrelevant outliers can make life easier.
The first few times out can be frustrating, especially if there are common phrases, acronyms, or terms clouding your search, or if you start too broadly. However, using the above as a guide can limit the noise volume and start your program off on a focused and efficient path to getting the information you’re looking for.
Author: Marianne Hynd, SMS
Marianne Hynd is the Director of Operations at the Social Media Research Association, a global trade organization dedicated to forming a community of researchers who aim to define & promote best practices and share ideas to help enhance the effectiveness and value of conducting research using social media.