Whitevector: Analysing online discussions – getting started
The latest report from Whitevector looks at the processes involved in analysing online discussions.
A successful online discussion analysis project starts with a clear understanding of how the final, produced information will be used in marketing decision making. Even though just monitoring interesting online discussions is valuable in itself, significantly greater value can be achieved if the data delivered is used to improve marketing communications, campaign planning, or digital marketing. The course of the analysis process should be gone through carefully bearing the final objectives of the project clearly in mind. This article explains general guidelines for the implementation process and then, taking a typical but purely hypothetical case using a well-known global brand such as McDonald’s, outlines the various steps involved.
The process consists of three essential phases. First, one must be aware of the average volume level of discussions around the chosen topic, as there is obviously no point starting a project only to notice that there is not enough discussion data to analyse. Secondly, the research questions the analysis is trying to address must be clearly defined. Thirdly, organizational and administrative issues must be considered so that the information produced will be produced in a format and to a timetable of greatest use to the client.
The first step is quite simple as it can be done at any time, and without incurring any costs. In fact, any brand that is being discussed online should be interested in what people are saying about them. Ignoring or undervaluing the importance of online discussions is not a wise idea as these discussions inevitably involve a significant number of people, who consider the information covered by the discussions valuable. The volume evaluation can be done by simply browsing through the discussions to get a broad overview of how many there are, and what they are like.
It is recommended that the research questions should be decided using a simple framework created from experience of the sort of issues typically raised in online discussions. Whitevector uses five different categories to create this framework: 1) online reputation management, 2) consumer research, 3) campaign measurement, 4) brand and product management, and 5) online planning. The chosen research questions usually belong to one or two of these categories; the brand itself and its current situation in the market usually give a quite clear starting point for the analysis. The research questions are the questions that the brand owner wishes to have answered. The questions could concern, for example, the client’s brand (and competitors’) online visibility, the various products’ popularity within their own categories, discussion audiences for industry discussion, etc. (See possible research questions below in the McDonald’s example.)
After discussing and defining the research questions, the team involved in the analysis should think of how the analysis will be used. Here, it is important that the brand team and the researchers at the media agency work closely together so that the agency conducts quantitative and qualitative analysis based on a clear knowledge of the client’s business and marketing strategies and objectives. Thus, the agency makes its recommendations based around how the online discussion analysis will be used throughout the organisation.
How might McDonald’s get started with online discussion analysis?
As a hypothetical example, we have taken a famous brand, in this case McDonald’s, and considered its needs from an analysis of online discussions. The steps are very similar regardless of the brand, but the selection of the right categories and the research questions depend on the nature of the brand.
First, a glance at the actual volume data concerning discussions around McDonald’s shows that there are approximately 3000 discussions a month in UK based discussion forums and blogs. Thus, we have a satisfactory volume of data to analyse. Furthermore, the results presented by Whitevector’s Chat Reports seem relevant and consist of interesting discussions. There is no immediate need to analyse the discussions further at this point but if one has no idea of what the discussions might include, it is good practice to use Chat Reports to read a couple of threads to gain a basic understanding of how the discussions usually progress.
Next, the brand team together with the media agency explores possible research directions: what are we trying to find out? It is common that when online discussion analysis is done for the first time, a lot of attention is focused on the kind of content the included within the discussions. However, to make the analysis worthwhile it is important to be as precise as possible in defining research questions and objectives. For McDonald’s, possible research questions could, for example, be:
- Consumer buying analysis
-
- What is said by customers’ about their experiences in McDonald’s restaurants?
- How strongly do customers recommend McDonald’s versus (for instance) Burger King?
- Service feature analysis
- How strongly do customer service and restaurant cleanliness feature in discussions?
- To what extent do customers associate restaurant cleanliness and speed of service with McDonald’s versus (for instance) Burger King?
- Does the issue of the healthiness of fast-food vary by target group?
- Brand analysis
- Has the volume of discussions around McDonald’s changed over time?
- Which target groups are discussing McDonald’s?
- Product portfolio analysis
- What are discussion volumes across McDonald’s product portfolio?
- Are there differences in how McDonald’s products are valued across target groups?
- Online planning
- What are the most influential forums and blogs for the fast-food restaurant category?
- How large an audience does McDonald’s engage in online discussions?
The third step, which is deciding on specific actions resulting from the analysis, is of course going to vary by individual case and will be a matter for the media agency and the advertiser. However, the link between the research questions and the resulting actions should be borne in mind throughout what is an iterative process, to make the use of ongoing online discussion analysis as efficient and relevant as possible.