Most publishers, editors, or sales managers ask researchers to conduct surveys to get answers to different questions.
It is natural for research departments or institutes (in many newspapers, at least in France, there is no research department and publishers hire external providers) to seek answers through qualitative or quantitative surveys of readers, subscribers, buyers, or residents — to collect undoubtedly relevant and useful information in a few days.
But there is another way, which we should use more and more. It is the statistical analysis.
Newspapers are sitting on a gold mine. Every day, they produce data: copy sales, subscriptions, classified ads, articles, complaints — by point of sale, by city, and through their Web site analytics. All this data, which measure and describe finely activity and performances, can be transformed into statistics — figures that can be analysed and give us so much meaning.
Let’s have an example of analysing the local performances of the newspaper.
The local news is the basis of a local newspaper. Accessing this local information is one of the reader’s main motivations for purchasing the newspaper. We measured it many times in ad hoc research.
But how can we explain why the newspaper sells better in one area than another? Is it because the newspaper deals better, more in depth or more regularly, with the life of the community?
Are there other explanations? Is it because the population of this area presents more affinity with the information? Is this population more educated, wealthy, or simply more attached to the territory? Is the distribution network better adapted, more efficient?
Can we plan how sales will go if we move certain parameters of the newspaper? More articles, more photos, more frequency?
To understand and answer these questions, we can listen to the inhabitants and readers of the areas where the newspaper succeeds, as well as to the ones from the less successful areas.
In other words, do research as usual!
Instead, we researched the data — population data, editors’ data, sales data. All these statistical data were integrated into a multi-dimensional analysis that allowed us to highlight the correlations between the good rate of penetration of the newspaper in an area and its peculiarities.
We identified the main criteria for success, including those dependent on the editor side (the regularity of editing articles about the area, for example), and the population side (the wellness of the people, etc.), and were able to weight each criterion.
From there, a scoring system allowed us to select the areas with the greatest potential for sales development. This was important information for both editors (who now knew where to focus content) and the sales department (which knew where to devote subscription resources).
We also applied this approach to other areas:
- Help predict the volume of sales of special issues and select best points of sale.
- Explain the rate of transfer between copy sales and subscription and help predict the impact of subscription campaigns.
- Highlight all the factors that contribute to the decrease in sales and predict the effect of an increase in price.
What you need to succeed in this approach is:
- The ability to gather reliable and complete data.
- Human resources skilled in modeling and statistics (but you can work with universities or research companies).
- Sufficient time to evaluate different hypotheses, data, and models.
- The ability to work in a pragmatic way that is oriented toward an actionable result.
In an age where Web sites, mobile phones, and tablets focus our research and deliver a plethora of data, the statistics approach for researchers becomes fundamental, by helping us sort out all the data and fine-tune rough figures.