[Tribune] Losses should be avoided when setting up your social room

The purpose of a social room? Monitor what’s happening online using the big screen and numerous dashboards. Lots of KPIs, a good database, reliable technical support and your motivation will be enough to make this operation a success?

The first problem to avoid: relying on only one tool

As a marketing manager, digital manager or communications director, you may not be experienced in monitoring social networks, but you do believe that it is important to monitor online conversations by quoting your brand and your competitors. You may have already heard of social listening or data analysis tools, as there are many in the market (Talkwalker, Brandwatch, Digimind, Meltwater, Netbase etc …) and you think this is probably a good starting point for your social room. .

After a few weeks of benchmarks, numerous meetings, and a few quotes, you’ve got a choice: Choose a service provider that is able to configure the tool for you. This will actually help you define keywords and create requests that will allow the social listening tool to crawl content on the web quoting your brand or your competitors.

After a few weeks of use, one of your colleagues in the community management department took stock of the latest communication campaign. Problem: Its statistics are much higher than announced by your social listening tool. Tell yourself that the tool has an error …

It is really very common that from one platform to another, the declared statistics are not identical. Similarly, some of the data inside your own brand will be difficult to access for an external monitoring tool. For example, community managers, administrators of the group’s Facebook or Instagram pages have more data due to external access restrictions in the APIs of these various social networks. The solution exists (token management, page configuration, etc.) but the first observation that can be made is the need to link the company’s internal data (known as ownership data) to external data (obtained through social listening).

However, variations in external data volumes depending on the source can still be noticed. Why? Each social listening tool performs real-time crawling of the web based on various requests made by all its customers. Some tools will perform well on Twitter, others on forums or even online press. The set of all edited searches constitutes the backdata of each tool. Thus, when one wants to find out the history of one or two years, this database is consulted. Totality does not exist on the web. Each tool complements and has its own database based on customer requests.

The amount of content produced online per second is such that claiming completeness when crawling the web is a myth. Therefore, it is easy to understand that no tool can provide identical statistics. Without increasing the number of subscriptions, it is important to give up the necessary freedom to be able to change tools if necessary. The best way to evaluate the relevance of a solution in your sector is to be able to test it before subscribing to any subscription. A software can actually prove to be quite relevant in a given subject and provide much less results in other areas of activity.

Problem 2 to avoid: Avoiding qualitative analysis

Conversation Volume, Voice Share, Tone Percentage, Virality Indicator, Net Promoter Score, Acquired Media Value … all KPIs available. Average, percentage, rate, swear by your management number, and your dashboards are there to meet these expectations.

Yet when it comes to formulating a recommendation, an action plan or a strategy, numbers help a little. And when you are asked about the reasons for increasing your competitor’s voice share, you feel a little helpless. Then you must “ Immerse yourself in the data Be able to extract real insights. The first step in understanding how Internet users express themselves is to read and analyze a semantic sample.

For example, in all of these phrases, sort out the 5 most viral publications, analyze their content and if you see that they all come from Instagram, they use the same lexical field and are published by large audience accounts, so this is probably an impressive one. Marketing campaign. You can now see that your competitors have seen a double-digit increase in their voice share on social networks in the last 3 months, but you also know how.

It remains to be seen how to combat this competitive communication campaign: Do you run an impact campaign too? Copy in another way? If the final answer remains in your hands, you now have all the cards to judge the situation correctly. Without quantitative, qualitative and in-depth human analysis alone, these decisive insights would not have been possible.

3rd Trap to Avoid: Believing in Artificial Intelligence

Imagine: This morning, the social listening tool detected an increase of about 12% in positive conversations. Today there is no specific communication operation, you wonder where this growth could come from. The idea of ​​sharing information in the communication section crosses your mind, but you don’t know how to ignore qualitative analysis. Then you look at the data, and you did well.

Instead of positive conversations, you notice sarcastic tweets, sarcastically quoting your brand. How do tools come to qualify them as positive conversations? These are the limitations of artificial intelligence. Currently, the software is not yet capable of capturing second degree, contextual, embarrassing or humorous feelings. However, humor is the primary vector for spreading information on social networks. So it is not a component to be clearly ignored when analyzing data.

The error rate during automatic tone qualification is between 20% and 50%. Tone errors of about 20% to 50% can lead to a ranking, bias analysis and incorrect strategic decision making within the brand. Automatic qualifications do not verify people, so it is always essential to take a step back from the statistics communicated by the tool.

4th trap to avoid: underestimate the sound

You now take the pinch with artificial intelligence and know that a manual qualitative analysis is required before issuing strategic recommendations directly from the dashboard. Despite the fine-tuning and advanced filtering of the tool to eliminate “spam” type content, you can see that the results always contain irrelevant information. You still think that this noise must be a minority.

Be careful though, because if you take the time to analyze these conversations one by one, you will find that the percentage of irrelevant data is probably not so low. Of course this will depend on the subject being analyzed, but what are the strategic decisions based on 65% Noise Integrated KPI? 65%: This is the ratio of words mentioned by Groupe 361 in the data provided by Social Hearing Tools. And this, despite the artificial intelligence and optimized settings.

All of this data is time consuming to sort manually. However, this task can be assigned to an external service provider. Otherwise, there is always a need to take a step back and be careful when reading dashboards based on data that has not been manually modified.

Trap 5 to avoid: allocating resources for one-shot projects without considering the medium term impact

After a few months of use, you will find that LCD screens will have only aesthetic functionality. Everyone will then become accustomed to their presence in spite of their strategic position. Because even though dashboards are automated, people actually have to put behind this engineering so that the transmitted data can be put to real use. As you can see, setting up a social room without a dedicated team would make little sense.

If the purpose of the social room is to monitor its e-reputation, it is not a purpose in itself. Similarly, dedicating a team without assigning a specific purpose will bring little value. Why would you want to monitor your e-reputation? If a crisis is detected through these channels, do you have the resources to respond? With which lever and with which process? If an opportunity for dashboards is identified, will you be able to seize it? How will marketing departments use the indicators sent to them? A social house? And what’s in it? Before allocating technical resources and human resources to long-term projects, it is necessary to answer the whole problem. Properly expected, the establishment of a social room can truly establish itself as a central and driving component of the marketing, communications and digital departments.

You now have all the cards in your hand to successfully set up your social room You know that it is important to go through different sources of data, qualitatively analyze a sample of the conversation, take a step back from the analysis provided by artificial intelligence and, if possible, manually correct the data and noise of irrelevant publications.

These steps are timely and require the skills of true professionals You can count on the help of an expert who will be able to support you in this type of project. Feel free to contact AACC if you have any questions.

Tribune by Lise Duchatel, Social Licensing Manager, Member of Groupe 361 Agency, AACC.

Leave a Comment