Social Media Marketing Industry Report - 2019. Platform use: b2b marketers
Image: Social Media Marketing Industry Report – 2019

The Internet produces content exponentially. By now, even the most die-hard, old-school PR and Marketing people, who still remember the glory days of the 80s and early 90s when large sweatshops were collecting and processing information like busy little bees, know that they need technology to monitor and make sense of it.  The automated technology can cope with the quantity, but can it deliver the qualitative output that only a human intellect can produce? The short answer is – YES. The Natural Language Processing (NLP) technologies made several quantum leaps over the last 7 years, facilitated by major progress on the algorithm, application and computation side.

Let’s take a closer look into the various functional use cases to understand how automated media monitoring technologies ensure the end users enjoy the same benefits that was delivered through human processing.

Early crisis detection and prevention

With a media monitoring service, you can closely follow what is written about your company and products, the volume of the conversations and what is the tone of the discussions.
Usually, the issues are unpredictable! It takes years to create a good reputation but sometimes even a single negative statement, a review on a website or a single comment in social can destroy your brand’s image.

The early crisis detection without compromises calls for relentless 360-degree monitoring. This means – the PR and Marketing need to build a circle defence around the company; to capture, process, analyse and understand every relevant signal in order to flag it as harmless or as a potential crisis trigger. The sheer diversity of communication channels nowadays – websites, blogs, forums, social networks, TV stations, Radio stations, Print outlets is impossible to be covered by human resource. The reason – cost.

To identify key influencers, opinion leaders and multipliers

Brands and various organizations recognised the power of utilizing established social graphs of trust. This is why they look for the so called influencers who are nothing more than entry points into selected networks of trust.

It is an analytical process, where the suitable profiles are singled out of the majority by deduction. The result is a number of profiles most suitable to be the company’s brand ambassadors – leaders whose opinions are valued in a specific industry. Collaborations with the right influencers can increase your brand awareness, strengthen customer relationships and boost your sales.

The PR people look for those few media outlets that have the right audience and reach to be considered as potential targets for their communication.
In this case, the availability of rich and correct metadata on a large scale is key. Only automated technology for harvesting, metadata extraction and data enrichment can cope with a large scale audience profiling.

Human processing would severely limit the ability of the PR and Marketing people to profile user profiles and online sources. Instead of profiling near 99% of the audiences that would potentially be useful, through human processing, they would cover not more than 1-3%.

Hand Loom vs. Modern Loom
Modern industrial looms can weave at 2,000 weft insertions per minute. This performance is impossible on a hand loom.

To track and measure the success of your marketing PR campaigns

You start a campaign and you need to measure its performance against a set of predefined KPIs. In this use case, uninterrupted feedback is key.
Tools like Sensika perform real-time campaign performance feedback continuously and for as long as the user team requires. The measurement process is objective and uninterrupted.
Human campaign tracking, despite very precise, lacks of continuity and coverage – people take rest overnight and during weekends. This leads to data holes and inconsistencies.

To analyse your competitor’s coverage and activities

The competitive analysis is a critical part of the company’s development plan. A company may collect and analyse all available data on all relevant channels, but only when the metrics are benchmarked against the competition, a conclusion about the performance of the business can be drawn.

By monitoring online, social and broadcast presence one can easily appraise the competitors by placing them in strategic groups according to how directly they compete for a share of the customer’s money.

But also, if you know better the strengths and weaknesses of your competitors you can easily define all potential threats and opportunities.

As the competitive intelligence is the key arbitrary factor for an overall business performance appraisal – speed is critical. You don’t need to know how you perform with much delay. You don’t want to change the performance evaluation KPIs, you don’t want volatile analysis. This is why you need a standardized and automated process that doesn’t face bottle necks and delivers fast and relentless.

Automated technology is getting smarter and more powerful with every passing day. There are numerous examples in the past that are telling us that this is the way ahead. Yet, we are in an interim period where the information processing and analysis technologies are just breaking out of infancy and getting into teenage age – they are not always capable of delivering all that we need.

The trick is to know where and when to use automated technology and when and where – human intelligence.

Data interpretation and conclusions

The media monitoring tools automatically analyze large data amounts to create an output of metrics. But eventually, the privilege to interpret the data remains for humans. Only human intelligence can make multidimensional connections and contextualization, and draw correct conclusions in hindsight of past experience and future targets.

Content categorizations and Sentiment scoring – still a big challenge


Your analysis could be in favour of the brand, competitor or government client. But what will be the categorization and scoring of the article? It depends on the clients, on the report purposes, coverage and many other factors.

While machines become better and better in single-dimensional categorization like – this article is about soccer or culture, or business, multidimensional categorization still eludes them.
For example – an article that is about a “Humanitarian crisis causes by a Military conflict”. For now, no algorithm can determine this causality and produce a proper classification.

The sentiment is another field where human and machines can still compete.  Sentiment scoring is the process of determining whether a piece of content is positive, negative or neutral in tonality. Basically, this is one of the major challenges in media monitoring.

Most highly machine-trained systems can reach 80% accuracy on average. It can’t fully understand the complexities of human language and it definitely can’t be categorised simply as positive, negative or neutral.

Usually, we are good at judging sentiment in a given context. But detection of feelings like irony, sarcasm, scepticism, anxiety that depends on a simple thing as the tone of one’s voice is a complex task, even for some humans, but for the machines it is impossible.

Content scoring depends also on the client’s viewpoint. Two different clients may interpret the attitude of the same text differently.


Companies need to consider the proper mix of automated and human-driven media monitoring in order to be able to enjoy the proper collection and preprocessing of data, but also the efficient data interpretation and conclusions

The future direction is clear: machine learning technologies will increase the capability of AI to the point where humans will be only the beneficiaries, but no longer the tools of producing insights.

Would you like to try our solution, it’s FREE.