Millions of items of consumer-generated content are uploaded every day, so how can marketers maximise this resource and learn more about these contributors?

Consumer-generated media (CGM) is a massive, largely unregulated dataset boasting few rules and regulations. Democratic participation, ease of entry and freedom of content are advantageous attributes yet call for vigilance when imposing a measurement structure. It is critical that the data underlying insights based on CGM is of the highest quality. So how do you go about ensuring the data you get from your provider fits this bill? We believe there are a few key questions that will help you separate the wheat from the chaff.

What is included in the data to be measured?
Some research takes a limited view of the vast CGM landscape, focusing exclusively on blogs. Additional sources of CGM include message boards and forums; public discussions (Usenet newsgroups); discussions and forums on large email portals like Yahoo, AOL and MSN; online opinion and review sites and services and online feedback and complaint sites. This is an evolving and growing list of CGM sources, each enabling consumers to express themselves in a unique format -- often with varying motivations with valuable implications for clients. These include opinions, advice, consumer-to-consumer discussions, reviews, shared personal experiences, photos, images, videos, podcasts, webcasts and other emerging forms of this new social media.

Where is the content coming from and how is it harvested?
Some research is undertaken using data that has been supplied via a third-party aggregator or provider and, therefore, is limited to the formats and breadth of data they provide. For organisations that harvest their own data, technologies exist that spider or crawl websites to collect information. Sometimes this technology is unable to capture the proper information from a site or a collection of sites. Thus, changes then need to be made to the harvesting technology to capture the information, but not all companies go to this painstaking and manual level of detail.

How is the data cleaned and prepared?
How does your data provider clean and prepare the data for analysis? What rules are applied to systematically reduce irrelevant conversations (noise) and ensure relevance? For example, if you are interested in CGM insights on the telecommunications provider 'Orange', how do they ensure references to orange as a fruit or colour are excluded? Often, within forum posts, content from a previous post is included in subsequent posts and needs to be removed if the research is to provide accurate quantitative information. There are many other examples of cleaning CGM data.
 
How is the data organised or segmented?
Is the remaining content relevant to the business questions being asked? What are the base, volume and discussion sources being included for classification? How is the data being segmented so it contains the most pertinent consumer discussions around your specific area of interest? By looking at a segment of the online conversations relevant to a research objective, perspective is introduced to provide norms and benchmarks.

How is the data being analysed and are actionable insights delivered?
How is the information actually being analysed? Is it purely done by automated technology or by human analysis, or both? If technologies and software provide the information, how does this technology manage to measure constructs like the sentiment of a conversation accurately? Can technologies help you determine what the important topics are that lead volume or drive a particular sentiment? CGM calls for multiple layers of analysis. Software and technology deliverables provide flexibility, but often are limited on research depth and actionability. Research teams and more manual processes provide depth and accuracy, but are less cost effective. Actionable insights are often found when these two are tightly linked.

How actionable are the insights and how should they be implemented?
Is there a consulting service so that information and data can be transformed into insight? CGM is a vast and ever changing form of media and many organisations need to rely on the expertise and experience of a well seasoned research team. While data can be informational, consumer generated media insights are powerful building blocks that can be used to transform and prepare an organisation for the changing digital landscape.

Brad Little is director, industry solutions, Nielsen Online.