How can we make the digital media industry more data-driven?!


Cosmin Nastasa is the co-founder of Data Revolt, a data-driven media agency that has managed to shift focus from the old way of doing business and adopt data driven thinking at all company levels. But where does our market stand in terms of data adoption? We had an extended talk with Cosmin.

How advanced is the local media industry when it comes to working with data? There is a very popular debate now concerning 1st party and 3rd party data. 

Cosmin: Data is used at a rather low level locally, I dare say. There are, in the first place, a few data categories. 

First, the data generated by media campaigns, either performance media or awareness or lead generation campaigns. 

The second category of data is related to the website, data that comes from the ”digital property”. What I would include here is the intersection with other types of channels, with other traffic categories, I would include user & commercial data, everything profitability, products, prices, types of sales and ultimately, technical data. 

These are the main data categories an industry player should consider. 

But besides that, we have a third category of data, related to market regulations. The 1st party-3rd party data debate I would rather define as strategic data, market approach and overall business activity approach. 

We have a fairly recent history in using data, let’s say some business fields are more driven to using data, they are somehow forced to do that, and here I am talking about the e-commerce types of businesses but generally, other areas don’t see the value in working with data yet. 

And this is generally due to the fact that the data-driven business approach has developed exponentially, it has grown rapidly, with a much higher speed compared with the speed at which advertisers and agencies are used to learning. It could be one of the reasons why data in business is so scarcely used. 

Although, at some point, I have observed some reports that were built outside Romania, with answers coming from digital managers. From what I saw there, our local state of business is quite similar with the one outside our borders. We are maybe just a little below the line because we started a bit later or because we didn’t have so many resources but at a global level the situation is similar: most companies declare they would find it useful to use data more frequently and this is one of their main projects for future years. 

Companies try to invest in capacity and human resources to better manage data but the reality is that all implementation and other connected capabilities companies build at the moment are a bit lower compared to their planning. 

Referring to the volumes and types of data you are mentioning, what should a company start with? What are the first steps that should be made? 

What comes in handy should be the first step. Let’s say, the low-hanging fruits. What Google & Facebook offer, or what other media platforms offer in a day-to-day activity or digital activity. 

Every free feature that is now offered by platforms provides a very high, valuable and powerful level of data. This could be a first step. And what is important is that we have the data, we have all platform features in our hands, and those are extremely valuable. 

Another important aspect would be a very accurate data collection. One of the biggest problems I have encountered by now, in almost 99% of cases, is related to the accuracy of data collection. This is paramount. People should try to find ways in which they can correctly collect data and then find a data collection model that fits their business needs. The approach of many is – we have a free tool, let’s use it -. That is not enough. The tool, per se, doesn’t say anything without accurate interpretation. Data has no value without accurate interpretation. This is why a collection model and an interpretation model are needed. And obviously, this is done with people. There are only a few specialists, people who really understand what is important in data collection. 

Another important step in acquiring data-driven capabilities is understanding of reports. Someone who works with data should be able to look into the reports and understand what those reports offer. Otherwise, the whole process is in vain. 

And maybe realise what metrics matter? There are some metrics known to be misleading and others that are important, and oftentimes overlooked. 

Understanding of metrics is indeed a very important step, it starts from the moment you interpret data. This step is closely tied to a basic understanding of the phenomena and also connected to measurement skills. But let’s say we are over the basics, we have placed the metric in a measurement model and moving further, when interpreting the data, one has to understand exactly what it means, what it defines, why the metric is important in the big picture of a certain business. There is a popular joke on the market about the bounce rate that is a very misleading metric and can be defined in many different ways, based on how you customize it and people can trick themselves into believing wrong things because they don’t understand how the metric was defined, they take it for granted and become fooled by it. 

What is the angle from which data should be observed when answers don’t come easy and data volumes are immense, maybe confusing? What tools or tactics can be used to formulate the right answers? 

When data volumes are immense and confusing, you always go back to the basics of your strategy and ask yourself “why”? Data volumes are generally huge and can be confusing. But if you manage to extract what interests you, everything becomes more simple. 

Cosmin Nastasa

If we are referring to the Romanian market, we haven’t reached the “big data” level. We are still at the point where data is extremely manageable and easy to understand and the tools that are now available provide a great deal of help in terms of filtering information. I dare say that what is absolutely paramount is the level of involvement. I believe data volumes should not be managed by a single entity. There should be more than one entity involved in reading and understanding data. 

And also, tactics are less important than tools, right now, from my point of view. There is a wide list of tools, starting with Google Analytics, DataStudio, reporting possibilities, the market is packed with tools that can aggregate data and make them easier to understand. 

One of our models is based on breaking data categories into the different entities of business, and then everything looks more clear: commercial data goes to the commercial department or, for instance, campaign data, goes into a separate report. But I wouldn’t see this as a problem right now, data volumes or the confusion that is generated, as long as there is a correct data collection model, as long as there is a plan.

The tools available right now are powerful enough to help us understand many aspects. 

On the other hand, we can get into personalization, into customization, where businesses need reports tailored on their specific needs. At that moment, the situation becomes a little more complicated, because besides tools and accurate measurement, we add people that can help the process, and here I refer to the Development and Martech departments that manage to work with raw data, collected at a base level, and turn it into reports that are much easier to understand or interpret. 

I am asking you about the tactics because I remembered that case from WebMixx, where you were mentioning the “missing bullet points theory”, and I was wondering how is it that someone gets to see such things, how does this particular way of seeing and interpreting data appear? 

Concerning that particular case, I would say it is a question of creativity in using data, collecting and understanding it. That case was rather about an interesting use of data and the real question came from somewhere else, not from the people who performed data analysis, but from those who were wondering how to approach the entire campaign or everything related to that certain business. That is where it all started, the hypothesis was born there and it was first defined there. Ultimately, the hypothesis became a theory that was verified inside data sets and we had a case of creative use of data. 

If we take a step further and talk about the importance of specialization and specialists, data becomes valuable when people ask some questions, pose problems and with the help of data they get to solve those problems. 

How does the new ecosystem look with data and ad-tech? 

To me, it looks a bit more complicated. We have data, we have adtech, martech, data marketing, data analytics, so many new specializations, there is an amazing diversity and diversification in the field. Although apparently things look more simple, they become increasingly complicated and here I am not referring to the details, but I am talking about the need to understand different fragments in the digital ecosystem. People need to understand and master data. The development of this industry offers us the opportunity to create another area of specialization. If we think of the internet and the digital field and generally, what gravitates around it, it is fascinating. If I come to think about how I started and what we have now, the difference is immense. 16-17 years ago, digital was represented by a 1 person department. And all developments in digital have generated progress in legal matters, we have a niche dedicated to online there, same with financial and so on. 

Where is the local industry positioned in terms of automation? 

Unfortunately, we are a bit behind here as well, although we have people and knowledge that could lead us towards automation. Romania is well known for its developers, for programmers, we have what we need to accelerate automation but we have probably not reached that need to get to the next level and keep things under control & simplify, in order to reach automation. 

We also have cases where automation is used, e-commerce, for instance, but we are still pretty far from it. Start-ups will probably help, together with all new ideas on the market that will be quickly adopted as they appear. I believe there is potential, there are ideas, but many times automation is dictated by what companies need, big companies, but they haven’t had the first say in accelerating automation, rather, the first step was made by startups that came up with ideas. We will certainly reach the point where big companies, in an aggregated manner, will be able to push automation. 

What would be the client’s role besides the media or data agency? 

It is more and more clear that we are not facing a type of client that is focused on a single  business area anymore. With what we do, it is clear that the client is increasingly involved, educated, he or she sits on the agency’s side and finds unity with the agency through the value that comes from data. Through this invisible thread, decisions are being made together, we don’t receive a task for something we have to develop or make, but we settle common objectives with the client. We have consultative sessions, we discuss, we brainstorm, because we are always facing the data and we get to make decisions together, we see what could work better, if not best. The client should encourage this experimentation path, encourage openness and investigation of new angles, to create this playground that generates the best solutions. 

On the other hand, if the client’s role is rather to generate some fixed ideas that should be passed on, we have lost all the fun in having some data to discuss on, data to interpret or observe and generate other types of ideas, together with the client. 

I even have some examples in this case. Clients who have started with the premise of selling something, and who have changed their minds after a data-driven brainstorming session that made them realise their strategy or objectives are not worth the effort because things can be done differently, the strategy can be changed, they can sell something else and what they had in mind initially cannot be sold. But we found a better way together. 

A permanent dynamic, here is where the client should be, it should be a partnership. Because we can discuss some indicators that allow us to be two pairs of eyes that aim towards the same objective. 

What do you think are the necessary qualities or skills for a professional to become an expert or a thought leader in the digital  media business nowadays? 

If one really aims to become an expert, I think they should not only think about media or data, they should understand that no matter their specialization, let’s say performance specialist or data analyst or a developer in martech, it is important to understand everything that happens around a certain type of business.

Cosmin Nastasa

It is a little weird, because nowadays we have specialists on all sorts of tiny niches. They master that niche, but they fail to understand the big picture. Why do I say it is weird, because the logic should be that one first has to understand marketing or digital and only after that first step is completed can they deep dive into a certain piece of this entire puzzle. 

The types of people we want, or the kind we support and nurture are rather the digital consultants who understand what happens on channels, people who understand media and maybe do not master everything, but understand how media works, have insights about the role of each channel, can understand data with all its technicalities and can see the reasons why some things work while others don’t, although they appear to be working. These types of people also know what communication means, how messages are delivered, they know how a website works. 

It is really hard to become a specialist nowadays, up to a certain click. Up to the first click the user has performed. No one can be a specialist up to the first click a user made. We need a specialist that can remain one up to the last click and maybe even farther. A specialist up to the first click can be one who works in a hyper-specialized industry, or someone who relies on repetitive tasks, but the profile of a true specialist is tied to understanding of the entire journey and he or she can define and come up with questions related to channels, can understand the effects of a certain type of action, commercial intersections, technical intersections, understand the need to perform a series of strategic actions and business actions on a market. That is kind of it. One cannot embrace a niche in this field and be content with it. That would be some sort of lack of synchronicity with the field. 

Where do you think the digital media industry is headed at and where do you see room for improvement? 

The digital media industry is certainly headed towards automation, it is apparently simpler but it is actually finer and more complicated. This misleading simplification comes from the fact that the algorithms are doing their job or should perform the job us humans used to perform. And the reason why I am saying that things are about to get more complicated, is due to the fact that we should be the ones who direct, who know how to direct all the efforts towards a certain area, we should be the gatekeepers or the ones who know how to fit together all the pieces of a puzzle. On each puzzle piece there will be an algorithm that will do its job but one has to know what are the exact pieces and how they can be matched. We are only facing apparent simplification but the reality is that other creative horizons are foreseeable in terms of business, testing, and other ties between different concepts. 

Things will change, the cookieless world will most certainly bring some changes, we will always need “more and better” and people will become increasingly specialized, the level of expertise will be higher and we will know more, people will put a stop to amateurship.

Cosmin Nastasa

What I would add, thanks to these algorithms that help us in automation, we will need to be very mindful and thoughtful about putting the robots at work and never let them have entire control. The robots are only a tool and it totally depends on us if what this tool does is serve our objectives. It is ultimately not about the tool, it is about us. 

About the author


Copywriter and content writer for almost 10 years. Brand storyteller that turns brands into both visual and narrative experiences. Event executive for Re:Think:Analytics, turning ideas into live broadcasts, for the moment.

Add Comment


Copywriter and content writer for almost 10 years. Brand storyteller that turns brands into both visual and narrative experiences. Event executive for Re:Think:Analytics, turning ideas into live broadcasts, for the moment.

Stay connected

We can only promise that this is a place where you will find the information you were looking for but didn’t quite know you did. Let’s re:think the future with data.