Written by Pål Børresen
In order to make educated decisions in a world that is becoming increasingly complex, with more and more different sources of information to deal with, you have to use methods to make the picture clearer.
Furthermore, we have far more data available than in the past, and the challenge is to put this data together in a way that helps us get something understandable and sensible out of it – thereby learning from it.
This is where modelling comes in. In 2020 we have tools, based on advanced machine learning technology, which allow us to make data modelling both cheaper and more accurate.
Within climate research, we have a number of different types of climate models, from extremely simple energy balance models to the most advanced Earth system models. Common to all climate models is the use of computer programs to calculate different climate parameters such as temperature, pressure, wind and humidity, in order to map all the possible factors that may have a significant impact on the climate over time.
Within marketing and sales, there is also a wide range of factors that interact with each other. What affects developments in sales, or what is it that causes a brand to develop in a positive way?
Only a few years ago, you could see effect with the naked eye. If two graphs went up, there was probably a connection. Many marketers knew exactly what to do to achieve the desired effect.
This is no longer the case. The number of data sources available has exploded, and it has become more challenging than ever to understand how marketing and different media channels impact sales.
What percentage of sales is actually due to the advertising campaign?
And, not least: what should the breakdown of your media investments be in order to maximise sales and make the most of your advertising budget?
In order to find an answer to this in 2020, we need to use the econometric method, also known as modelling. By using econometric models, we can, with great precision and at a low cost, provide answers to the extent to which different internal and external factors impact sales – and, importantly, what effect they would have had if the budget had been allocated slightly differently.
In order to be able to say anything about what affects sales, and in particular the impact of market investment, we need to look at various factors such as the type of media channel, distribution rate and price changes, or external conditions such as competitor activities, seasonal variations or the weather, for that matter.
Imagine you're running an advertising campaign for ice cream during a period of warm weather. Has the advertising increased ice cream sales, or is the increase due to the weather?
With sales modelling, we collect data relating to all of these factors and put it into a data model – just as climate scientists do. Instead of trying to explain what affects temperature, we try to explain what affects sales. In this way, we can put together a model to provide us the answer to this.
So, in 2020, why should we be using modelling? It’s clear to us that data is worth its weight in gold. The only problem is that it’s difficult to refine the raw material that is the data – we can't make gold out of it.
However, using sales modelling, you will be able to refine and create value from data that is already available to your business, or at least should be available.
Based on historical data, you build models that help explain the degree to which different factors contribute to sales, including the value of the brand in isolation. This gives you, among other things, a basis from which to optimise your advertising investments in the future and to calculate the profitability of the measures you are planning.
Data modelling is one of the most effective measures to streamline your marketing activities and thereby your ROI. So make sure you take care of your data, in particular data that shows advertising impact per channel and per week, as this is important input for your models.
The leading businesses learn what encourages short-term and long-term traffic and sales among different target audiences. The quickest way to obtain these insights is through modelling. By performing modelling, you will have much greater confidence in what you are getting in return from your advertising budget, and how to spend that money in the best way possible.
In fact, you could go so far as to say that if you are managing a large marketing budget and not carrying out modelling, then you’re virtually being professionally negligent.
In 2020, sales modelling – using market-leading marketing intelligence tools – is still the only way to understand the correlation between advertising and sales as a basis to optimise your advertising investments.