Video content marketing, the reason why you can’t do without it

Interview with Giorgio Marandola, trainer and consultant

Giorgio Marandola
Consultant and trainer

Online videos are a priority for all brands, today and in the coming years. Companies need to become content creators and media broadcasters now more than ever.

Let's explore with the expert Giorgio Marandola what strategies and tools to adopt, and how Artificial Intelligence and DAM platforms are increasing the effectiveness of videos.

Q: Why should companies consider including video content in their Content Marketing strategies?

A: A couple of years ago I attended an event where a Facebook executive said "Video is the new content". I was so impressed by this sentence that I wanted to know more, where the reasoning came from, and data gave me several confirmations.

First of all, it’s not up to us to decide what to include in our Content Marketing strategies, but our choices must be supported by data. CISCO revealed that within the next year, about 85% of mobile data traffic will be used to watch videos.

This shows that we marketers must be ready for a very simple process, that of interpretation and rapid response to trends that come from the consumption of content by users.

In addition, I'm obviously not the only one to have noticed that the available time span to capture the attention of a user on the feed of any social network is about 3 seconds. This figure alone does not impress, unless we compare it to the 5 seconds-span we used to have at the beginning of 2018: the overload of information - and consequently of content - puts us against the wall, and we need to choose impactful, interactive formats.

It's all about being able to strategically plan the design of video content and its production, with an eye at containing costs, or else you risk spending a lot and obtaining little.

Q: What do you think of Google's 3H model, does it convince you?

A: Google's 3H model is not only valid, but it is also essential to set up a first strategic framework for our Content Marketing project.

Every day I observe many content initiatives, especially B2B, which use social media, blogs and so on, but focusing on corporate blogs and text content, the scenario is desolate: I see companies, even in our industry, use content in a random way, just a succession of content that somehow relate to the industry of reference.

In those cases there is no strategic vision, no relationship between the content and the supposed core business of the company.

What I’ve been preaching in recent years is a very simple and tested process: start from the base of the Google 3H pyramid, with the Help content, the kind of content that intercept the needs and demands of the target audience on search engines or social media. That’s where many (alleged) strategies fail.

Content creation at this stage cannot be random or based on rumors, but we need to know what our target is looking for, so we will use Google Trends and the keyword planning tool of Google Ads to produce the kind of content that our future buyers are looking for.

Once we have intercepted our audience through the Help content, we will produce the so-called Hub content, i.e. those aimed at our prospects, the people who have already entered our ecosystem and are sniffing around our product.

Last but not least the Hero content, which is the spearhead, the type of content that can reach and interest the "general public”, or a very large part of our target, like a television commercial.


Q: When you talk about Video Content Strategy you suggest to act asymmetrically. What you mean by this?  

A: In this field, by asymmetry I mean a broader view than the classic bottom-up approach that your digital channels, such as corporate blogs or social networks, usually have. You have to produce content where the large communities in line with your buyer persona are, so you have to leave your own channels and keep them as the final step of a hypothetical funnel.

To make a concrete example: years ago, Puma launched a new football shoe, and to do this it decided to focus on YouTube to engage younger audiences. The first obvious thought at this point would be to produce a commercial and simply have it run as a preroll. Instead, Puma acted asymmetrically with respect to this - even today - little rudimentary strategy.

It involved two youtubers and pro soccer player Mario Balotelli, the former two challenged each other in different challenges and football tricks, posting the outcome on their respective channels, already reaching with this first branded content several hundred thousand people organically. Afterwards, the winner had the opportunity to challenge Balotelli in a competition of penalty kicks taken in a goal held about 10 meters high by a helicopter.

The video of the feat was published on the YouTube channel of the youtuber, of course, and to date it has received over 1 million views.

Puma itself posted a classic 30-second video highlights of it on its channel, but reached much smaller numbers than the total the content reached on the youtubers’ respective channels.

Overall, the campaign produced several million organic views, compared to the few tens of thousands of views of the corporate piece of content.

That’s what I mean by thinking asymmetrically, outside the box, keeping in mind what result is the objective of the campaign.

Q: In your opinion, what are the advantages of applying AI, specifically convolutional neural networks (CNN), to multimedia content?

A: AI is here to fill a gap that hopefully will become wider and wider over the next few years. I say “hopefully” because in a Content Marketing perspective, it is to be expected that each company will become more and more similar to a media company, even if its core business is the production of metal laminates.

In very practical terms, we expect that the content produced - images, videos, text and web pages - will keep increasing, with more and more effort required from those who deal with rationalizing and organizing it into Digital Asset Management systems.

In this, Artificial Intelligence performs a crucial function: it interprets content and helps humans in the process of rationalization. We are talking about automation, which allows us an organized use of content on front end systems, and about scalability, in a context increasingly rich in content - as already mentioned for the production of videos – making processes as straightforward as possible is the key to be sustainable.


Q: Why is it so important to classify and structure content with metadata?

A: Content is a bit like the universe, always expanding, so rationalizing and organizing these assets is crucial to be effective and efficient in brand and product communication.

Metadata helps the searchability and the sharing and of any content, and also serves as a basis in the training processes of Artificial Intelligence. With it, you can teach AI to recognize key visual elements and associate them to a tag or a word. We constantly use metadata for any content we publish online, can you imagine a piece of content without a reference describing it?

I remember years ago, I was working for DHL, and I got to appreciate their intranet and their Digital Asset Management system, a huge space categorized strictly, with metadata properly associated with each asset. The use of photographic and video content was made easier from the beginning.

Q: Does this process involve a training phase?

A: Absolutely, yes it does. Training is the basis of any AI implementation. In the case we are analyzing, the initial period of training will provide the machine with the essential reasoning to catalog content, to assign metadata and make content re-usable as much as possible, internally or externally.

All the uses of AI imply a complex phase of training, but not only: usually, the machine is left some space to perfect its own level of learning while it is working on the tasks for which it has been programmed.

Google, for example, uses an AI trained to automatically and determine which pages are qualitatively fit to be indexed in its search engine and which cannot.

In this case, a fundamental initial training is made and checkpoints are set when the human being will evaluate the work of the AI and, if necessary, will correct it.