A question of duopoly and alternative possibilities
Silviu Toma is a passionate media and analytics professional. He takes pride in the ”best team ever” he has built over the years. Silviu’s everyday activities revolve around clients such as Orange, ING, Vodafone and Molson Coors, to name a few. Silviu works at Microanalytics, a company that offers analytics services for media and digital products and Allinmedia, a media audit company.
We are presenting a perspective coming from an expert with over 15 years of experience in media and marketing, who has been in the middle of a great transition from offline to digital and is well aware of all consequences of change, transformation and pace shifting in the field. We had an extended dialogue, shared with you below.
What is the general understanding of companies when it comes to the advantages brought by data focused and data driven thinking, especially in the marketing field?
Silviu Toma: I think here, in the marketing field, we are in the middle of a hype with data driven thinking, because the companies that share the greatest power ever, globally, are two giants that are truly seen as tech companies: Google and Facebook.
Globally, 70 to 80% of marketing and media plans are executed through them. So I assume that 70 or 80% of data marketing strategies are also planned through them and invested in them. What remains in terms of planning and investment are other thousands of economic players in the digital field. When I see a presentation coming from a media or marketing agency describing plans on how to use data and the like, I remember the percentages above.
When we want to target a group of people with ads, we are using data delivered by Google and Facebook. Even the results of our campaigns are measured with instruments they have created. There is no independent company listed by the Romanian market to check the truth of what these two giant platforms are selling.
Have a look at the picture below. It is a screenshot of what Facebook sells to a marketer as a list of options to target people interested in food. Truth is it looks amazing when we see so many targeting possibilities and so many segments but there is no authority, no company that can verify if those groups created by Facebook are genuine and if not, for instance, people interested in veganism are mixed up with those interested in fast food. We don’t know if those segments have the exact types of people we want to target.
This is why I’m saying that all this data driven marketing is a hype created to make us believe in the power of data, in this case only applied to the digital advertising business.
Actually I do believe in the power of data but only in relation with gaining the capacity to measure what I received from my investments. I do not believe in the power of data applied to creating groups of people accurately targeted. In this particular case, I believe in the power of good content.
People have been creating content for centuries and will continue to do so. What is different now is that we give free rein to algorithms, which are some sort of robots, to choose what content is delivered to whom, and what kind of content we consume.
What is really amazing is that Google and Facebook are not content creators, these two are not creating content. They are presenting themselves as technology companies that are mysteriously meddling data to give us almost endless possibilities to target people. They are gathering pieces of content from us or from established content businesses and create something similar to a network of content. But if we are looking at where their main revenue is coming, they sell advertising. This tells us that Google and Facebook are in the media business, not in the technology business.
Here is the trick: data driven marketing is the story they sell us to make us forget that actually, content is important. They somehow managed to create this bubble of trust around them, because they have amazing data engineers who can help us target consumers and execute marketing strategies based on data-driven thinking. But I am not buying into this 100%.
What can be improved, considering the autonomous possibilities new technologies offer?
My idea of using data is related to building an independent, objective point of view about what we get from digital marketing. Apparently, people interested in food, for instance, are ideally easy to target, I backed my example with the screenshot above. Taking this into account, literally anyone can open an advertising account and start targeting people interested in food or any other topic or subject of interest. The data provided when we choose the targeting criteria is a so-called “fast-food data”. It is rapidly delivered but we do not know how healthy it is. And it’s not even a bad thing, but one can not stay healthy only with fast food, we need something more to create a balance, to stay healthy, we need veggies too, to use a metaphor.
And to back my idea of fast-food data and lack of standardized measurements in terms of data accuracy, it is a well known fact that when we go to the market to buy a kilo of cherries, the seller measures our cherries with a standardized, independent, approved instrument, the cherries are weighed on a scale, not on the seller’s proprietary and owned platforms with its own standards, unknown by any other seller on that cherry market. Google and Facebook have their own standards related to audience measurement. As an alternative, we have a few standard metrics only related to buying performance. The Media Rating Council together with IAB have done a really great job in defining and verifying these metrics. Fortunately, we can measure an impression or a click with transparency, but this is not enough. We need more.
In the digital market there are some important initiatives to build universal metrics and instruments to measure what we spend money on but we are far from the standards that we have in everyday life. And we should choose wisely how to measure what we are buying and what data to rely on.Silviu Toma
Despite all the hype around this digital world, we are still in the Middle Ages . If we are crossing the river from Google to Facebook, the video viewed by a user on each platform is measured differently. Same with the Middle Ages: if you would have crossed the river from the land of the Lord X to the Lord Y, the system of measurement would have been different. A universal system of measurement is the invention of the State, of centralized power. But the states are still struggling to figure out what is happening.
Don’t you think people need to be more aware in terms of owning and extracting value from first party data? I cannot know for sure how much of our local industry or marketing departments are built to use first party data. Money is predominantly invested in digital campaigns managed by Google and Facebook.
Silviu Toma: This is something only the market can decide and act upon. In my view, where there is a way that makes it possible to gather and process first party data, there is also value.
In order to become more efficient, a company can grow its abilities to extract value from it’s own data sources. But not all companies can do this because there are companies that don’t generate data from digital interactions with their clients.
And that’s ok, not every company is “digital at the core”. Let’s take companies who produce beer as an example. Their product is destined for offline consumption. They need a different framework in order to independently measure what digital media are giving them in exchange to their investments. We need something similar to the TV industry standards, where there is a third party company that measures what TV stations sell. And taking this classic example into account, I think it’s time we had the same kind of measurement system for digital media as well, that would enable us to measure audiences. I don’t necessarily assume that the digital providers are not doing a good job, but it is always safe to check and have a third party in place that is able to measure audiences, results, investments, in an objective manner.
I would really like to highlight this difference between fast food data and healthy data investing, things that are done with higher investments maybe, with better trained people, who are more in tune with objectives, people who have a greater level of knowledge in terms of what they want to invest in or extract value from.
Silviu Toma: I am cautious with that, because the market eventually dictates what is good or bad, healthy or not healthy. We have this tendency to put things this way, like, hey, why don’t you do things like that, because it’s better. I don’t agree with that.
Companies should see for themselves and understand where the value is if they work with data: fast food data or slow food data are not fixed concepts, we can define these concepts based on business needs.
The market decides.
Generally, the debate is directed towards how data is used by the giants and how ways of working with data can be improved.
Silviu Toma: Yes, this discussion has been first and foremost, generated by crises. Although it is very noble and kind to think that people are inclined to improve, in reality we all know that one does not make an effort to improve things if things are going well. If things go bad, then people will try to improve them. It is in our nature to act like that.
One debate was generated by the 2016 crisis, when Facebook was using all audience data, and everybody reacted to Facebook’s way of doing business.
Then, there was another crisis, related to children’s dedicated content that had pornographic comments on YouTube channels. At that time, P&G or Mars, I can’t exactly remember, one of them has ceased their communication on Social Media for a few months, until Google has addressed the problem, updated their algorithms, etc.
Recently, we had the crisis in Australia. Facebook and the Australian Government failed to come to an agreement on a fee that the latter had to pay to the publishers. Facebook has consequently cut off all news from that country and people have witnessed the amount of power Facebook can make use of and ultimately concluded that it is not good such monopoly exists.
Basically, each crisis has revealed to us from multiple angles that at least when it comes to communication and data about people, Google and Facebook have an immense power, we live in a duopoly that was never witnessed before.
And then again, in this area of marketing and data, both Google and Facebook have built their apps in such a way that many people who use the platforms misunderstand the difference between the truth and what they see on that platform.
Let’s take a simple question: How many inhabitants does Romania have? Some people see that information on Facebook and that becomes the truth. At least when it comes to the commercial public.
The interface was created to make marketers feel overwhelmed by the amount of data and have the feeling that these giant companies have all the data, it becomes weirdly clear for them that there is no other source of data and this kind of feature, this kind of gimmick is created by design.
It creates a feeling of saturation and you say that’s it, I don’t need anything else, I’m all good, I can handle it. Which is false.
Because you shouldn’t have the data you spend your money on from the guys you give your money to. You need to be independent when you want to measure how many people are in Romania, in terms of commercial target. These products, by design, have arrived with all the features and interfaces to seem very convincing, they seem to own the truth in terms of data marketing.Silviu Toma
I will say this again, this is not necessarily bad. These media platforms simply follow their interests and they offer us all the reasons to believe in their game. What I am really saying is that we should pursue our own interests in terms of how our money is spent and what we spend our money on.
More than 50% of the content we see online is paid content. How could a balance be created between this media democracy that is welcome and necessary and the genuine, organic information that should still be available to people?
Silviu Toma: Some say that the market should balance itself but the thing is the very market is distorted by the lack of a standard regarding how we measure what we are buying.
Others say that the State should intervene, through a set of laws that already exist in the States and Europe as well, to prevent the media monopoly.
But the weird thing and the thing that does not allow such a simple solution is the following: from the State’s side, the monopoly was often seen as a question of price regulation. So, when two companies control the market, price competition disappears, people simply pay for a more expensive and poor service. In the case of new companies on the market that work with marketing data, there is no price.
We don’t have a unique and verifiable metric to compare prices in the advertising market, across digital and classic media. So we don’t really know how the duopoly is influencing the price. So we don’t know how to act. And this is not Google or Facebook’s blame.
Do you think professional structures, professional people, have the power to balance things in such a way that what appears on the internet, what is communicated through and from a department made up by professionals, to the internet, on the internet, can be made more responsible, or maybe different?
Silviu Toma: At the present time, no. We are in a very awkward situation and I want to highlight that and reveal it as clearly and as simply as I can:
We literally don’t know what content one sees on the internet. To some degree, we are aware of it but for instance I don’t know what you are seeing, because you are exposed to mixed content, created by algorithms. And some of the times, this mix allows for something terrible to be inserted, or something false or horrible. Instead of hardware robots that attack us in Science Fiction movies, we have software robots that attempt an attack on our minds.
We don’t know what the people next to us see on the internet. But we cannot say that what we see in our feeds is private because at the same time the content each of us sees and reads is delivered by algorithms. So it’s not public and not private either.Silviu Toma
We have never faced such a situation in our entire history. If I want to find out what a certain TV channel broadcasts, I press a button and I can see clearly. Same with a piece of paper. You can ask me to switch the channel on ProTv and we will both instantly have the same content in front of our eyes. Or we can read the same book. But we cannot do the same thing on Facebook. Sharing content is just a fragment of an entire story. We don’t have a button that allows us to switch feeds and see what the other one is seeing. I can’t see what you are seeing.
Companies want to deliver a message to large audiences.
They will want to buy media and make their message known. They need to know where their message appears and the answer is they will not know.
What marketers have in their hands is some fancy buttons and checkboxes showing them that their message is delivered to an audience of 20-40 YO, based in Bucharest. But they will not know what kind of content is in front of that audience, where exactly it is that their message appears.
The algorithm on social media platforms is so fragmented and spontaneously created, aggregated, combined, divided, that you never know where exactly your message appears. And for that matter, you cannot really choose and condition.
There are some minimum pressure rules, you can create a whitelist, where you mention the preferred websites you would like your ad to appear on or blacklists for that matter, when you want out of certain websites.
We have too little control over message dissemination. One knows their message will be on Facebook, on YouTube, on a few sites via Google. But they don’t have complete control, they are unaware of the context.Silviu Toma
There is a certain amount of control but if one starts asking some questions and tries to go a bit more in-depth, they realise they don’t have control, indeed.
For instance, you are a beer seller and you would exclude all child content channels. You have excluded them all, you have done everything that is common sense and necessary but at some point you may appear on a channel dedicated to children and this would be a question of a wrongly directed algorithm – a new channel was created, you haven’t blacklisted it, it is possible you appear there as well.
You are never 100% certain that you won’t be there. And you never know what happens in the comments section of a video, there are many unknowns that are hard to control and sometimes they are not ideally solved by algorithms.
I would make everything a question of budgets. If you wouldn’t know what you are paying for, would you be willing to pay for it still?
The majority of answers would be negative. So, when you get money out of your pockets and you want to spend it you can ask yourself: How much of my spending can I control? I cannot have 100% control but how much could I control? To see how much, to draw that line and to aim at it. To use that money and gain as much control over your spendings. To be able to control where you appear, how you appear, what kind of content your message associates with, so that resources are not spent on inappropriate channels.
How does the custom agnostic attribution modeling work?
Silviu Toma: I think it’s good for the market to adopt this tactic. But generally, it is clear that the market offers a set of channels: radio, TV, print, online, cinema, OOH and others. And there are also a bunch of tools, such as Google Analytics that provides you with attribution models, exclusively dedicated to digital. We should develop attribution models for all communication channels. It would help us see what digital offers in terms of efficiency.
It is a method that allows you to have as much control as you can over your budget, because you know where to press and how to press to gain more conversions or the other way around, you know where and how to redirect an overspending, because that investment is obviously inefficient. This is how an agnostic attribution model works.
And it is a completely autonomous process. In digital, it manipulates data from Google and Facebook but does not rely on it solely.
Silviu Toma: It doesn’t rely on them exclusively. If we were to look at Facebook’s attribution algorithm, we would see that 90% of the people who buy something come from Facebook.
This is what I try to highlight. It is a completely autonomous type of measurement and it succeeds in independently extracting real data from all channels, either online or offline.
Silviu Toma: Yes, should be so.
Adidas used it and they discovered that they were overspending in digital. Others could discover that they are underspending. But what is important is for one to have their own measurement scales.
It is important for us to know that alternatives exist, they are independent and helpful for both professionals and final consumers, enabling both to see and think and act in a more balanced way. Because now, even if the year is 2021 and we spend 10 hours in front of our computer screens, not everything happens online, not all advertising campaigns are suitable for the digital medium. There are still channels that represent a good alternative and can bring a fresh perspective and fresh public.
Silviu Toma: Surely, this is the way things will always be. Here is another important point we overlook. For a very long time we have communicated based on an audience we did not know, but we used to associate it with a book we have read, with a certain article, with an editor, with a TV show. Somehow, the content was used to create the public we wanted to reach out to.
In the digital world, with the rise of programmatic, the algorithms look at certain data signals and they create a certain audience pool, depending on what we do online, on what we prefer to see, etc. So the audience as a pool is decided by the robots even if the content is human.Silviu Toma
We cannot see how algorithms work, they are some kind of work of art, they are not transparent for us to see how a robot has profiled “photography lovers”, for instance.
We don’t know for sure, the contemporary digital world lacks this association between content, editorial references, personality, affinity and the messages we deliver. We trust the algorithms, basically we trust the robots. I believe we should be more curious about this and more cautious. I understand why we need robots to do profiling and create an audience from billions of data signals but I don’t understand why this is a secret process. In the age of information we should manage the robots and not let the robots choose for us. So I think it is very important for us to have an independent framework to measure the efficiency of an advertising budget. Data is important for clients and agencies only if it works this way.