How network listening software works and how to use it.
Network listening software are all different from each other, however, they all function similarly.
Specifically, they revolve around four main phases:
- Setup and setting up searches
- Review and control of data
- Analysis of results and reports
- Review and continuous improvement
As you can see, these are quite understandable steps: before you carry out amarket analysis you have to set up the WHAT you want to research, then you have to review and clean the data, then derive overall results and then create the reports.
Let’s look at them one at a time!
1) Setup and setting up searches.
If you don’t let the software know what it’s supposed to look for, it’s hard for it to do that, right? And to set it up, we have to translate what we need into a language that the software can understand.
This language is called Boolean, because it was devised by one such Boole.
It is not too difficult. Remember when you studied sets in school?
The prof insisted on the seemingly trivial concept that if I add two sets together, I will have the total of their elements; if I intersect them, however, I will have only the common elements between them.
I remember it seemed like such a simple thing that I thought, “But of course! When will I ever need these things?”
And that’s where every circle comes full circle: sets are very much needed for the network listening setup.
It is not my goal now to give you a lecture on Boolean logic, because I could write a book on that alone, but I would like to make you understand the principle.
Read also: The 9 points to consider when choosing the right Web Listening Tools
Boolean logic as a setup principle for web listening software.
He thinks he wants to monitor conversations about Coke.
To understand both who writes it detached and who writes it attached (cocacola), we have to ask the software to SUM the two sets, right?
We want you to take BOTH all mentions with the word detached, BOTH all those with the word detached.
To tell the web listening software this, logical operators are used, usually capitalized keywords that the algorithm recognizes.
The operator for joining usually is the “OR,” which means “either this, or that” precisely because both fit you at the same time. The wording then becomes:
“Coke” OR cocacola
Then what if we took it a step further and wanted to see who drinks Coke while eating pizza?
In this case we would like to INTERSECT the conversations about the drink (both with the word attached and detached) and the conversations about pizza, not add them all up–because otherwise we would get so many more conversations that we don’t need.
The operator in this case is usually AND, because it means “and this, and that.” Our query, which is called a “query” in the jargon, then becomes:
(“Coke” OR cocacola) AND (pizza)
We need the parentheses to tell the software the order in which we want it to do the different operations: in fact, we want it to first find all the words that mention cocacola, and then choose only those that also mention pizza!
Finally, if we wanted to exclude those who talk about Pepsi, we would have to add a negative operator, a nice “NOT” to say avoid that term:
((“Coke” OR cocacola) AND (pizza)) NOT (pepsi)
This is a very simple example to give you an idea of how the software works and how it thinks.
Don’t worry if everything is still not crystal clear! Often you don’t have to do this particular step yourself but are supported by the person who sells the software or the person who does the support.
Any marketing agency that manages the software for you could also take care of it (just make sure the team knows how to use Boolean logic).
However, I wanted to give you a smattering so that you can understand the logic behind it!
2) Review and control of data.
Set up the searches, the software starts to grind and collects all the mentions we asked it for.
Finally you can move on to the next step: reading results and if needed “moderating” them, deleting irrelevant results, changing the sentiment if in your opinion it was not detected correctly by the software, and so on.
This step should not be done only once like the setup, of course!
It is the key to knowing right away what is going on and seizing important sales opportunities.
Indeed, for that very reason, monitoring and moderation can be assigned to the customer service team . In addition to correcting and cleaning data, they will be able to respond directly to users when needed and interact with the community!
Social media managers can also participate in this phase, to see what content is being shared the most and get a pulse on what is happening in their industry online.
Read here to learn how different business figures can apply social listening: Social media listening and monitoring: 7 practical applications in business
3) Analysis of results and report.
The third phase should be done less often than the second, periodically depending on your goals: it can be weekly, monthly or at most quarterly.
What changes from the previous stage? The way you look at the data!
Instead of reading individual conversations, you look at the results in aggregate terms, filtering them in different ways and exploring them from different perspectives.
Find out trends over time, the most talked about topics, comparison between multiple competitors or multiple topics, and so on. In short, find out the information that the different business figures you identified above need to keep track of the situation and make their own decisions!
Usually this phase of analysis and reporting is the task-indeed, the analyst, that is, that figure who collects all the data on everything to present to management or whoever they serve in the company.
If you don’t have a business analyst, a social media manager is fine, but the important thing is that they are familiar with numbers and a little bit of statistics.
The different social listening software, in fact, more or less helps you at this stage, offering graphs and possibilities to filter the data… But in some cases it may also be necessary to export all the results to Excel and work them from there.
4) Review and improvement.
The previous three steps have different deadlines: setup should be done at the beginning, monitoring constantly, and analysis of results at well-defined periods.
But there is more: these steps are embedded in an ongoing process, because the setup is not immutable, nor written in stone like the Tablets of the Law of Moses.
Rather, each time customer care reads the mentions or the analyst collects the results, it should mark potential changes to be made to the setup.
Extra words to track, or words to exclude because they litter the data; new topics that may be useful to keep track of, such as new brands or new trends, and so on.
Mind you: I don’t mean that you have to change the setup every day.
If there is any serious error then yes, otherwise it should be done only after closing one periodic report and before working on the next, so that the graphs and data already collected are consistent and valid.
Remember, however, that working with such software is a trial-and-error process that never ends!
When you start with the first setup it is only the first step, you know almost nothing yet about the buzz in your industry and so you hypothesize, put forth the first words and themes that are most likely important, and then humbly check what comes out.
“To improve is to change. To be perfect means to change often!”, Winston Churchill
Check, find new ideas, improve, and try again! So forward month after month, because the market changes and people change the way they think and express themselves.
He who fossilizes is lost!
A hug.
