FEATURE1 November 2010
FEATURE1 November 2010
DATA SPECIAL— Mike Page of Cognicient on how companies can integrate data from various sources to generate new insights and fuel innovation.
Over the last few years the breadth of insight sources and the maturing of analysis technologies and techniques have begun to open up new opportunities. We can now find patterns and trends from different data sources and, indeed, find new insights from existing data which would once simply have been lost in the silo’d, survey-by-survey methods of the past. Now, organisations can begin to construct an integrative research function or ‘research blender’ that will allow them to work systematically with their research data as an asset and use it in new ways.
The best ways to develop integrative research resulted from the rise of more cost-effective and easier to use business intelligence and data-mining platforms. These proved their worth in other aspects of data strategy, e.g. fraud prevention or cross-selling and up-selling. This has been combined with a growing desire from all sides to show a return on investment in any data source, particularly one as expensive to the organisation as a piece of quantitative research.
So how can integrated or multi-source data give improved ROI over an approach that puts research data in silos? There are a number of approaches. The first is cross-validation of different techniques or data sources – either using market research methods to validate data mining or vice versa. The second is using the two techniques to conduct follow-up analysis on each others’ results.
The other two approaches are ad hoc data convergence and full data convergence – this is where the true ROI of a convergent research strategy can be realised.
Ad hoc data convergence
Ad hoc convergence is the most typical model of how integrative research projects work today. The current research process remains unchanged, but through a separate initiative of some kind a new combined data store is created that incorporates either multiple different survey sources or a specific survey type such as product or concept testing. This can then be combined into a single database that also incorporates related behavioural data ( such as sales volume or the feedback from social media monitoring ) and a new set of hypotheses for models or new business intelligence deliverables can be developed. These allow a new, more holistic set of insights to be generated.
Full data convergence
The majority of our clients are looking to move to full data convergence. The important change here is that the integration is systematic and an automatic part of the research process. This requires a level of organisational change within businesses, and with the growth of insight as opposed to research-only departments we are seeing a migration that allows these divisions to take direct responsibility for the deployment of all insights that relate to the customer or market.
Ad hoc in practice
We see examples where a number of CPG companies are looking to exploit the value of the vast amounts of data they collect. In most cases this will begin with the hypothesis that if they can bring together their views on, say, branding or innovation, the organisation will have a consistent approach to this type of information and can begin to hook in the other parts of the marketing function to develop ways of measuring this aspect of business performance.
Our experience of these projects has been that they are usually successful but they are not easy to implement. When it comes to bringing together the disparate data sources, the first question is almost always: Who will pay for it? Ad hoc integration is an additional activity conducted within the current organisational framework. We have seen various examples of how this has been made to happen but, so far, those clients that have a specific budget for data integration almost always have a strategic desire to create a dedicated integrative research function.
Full convergence in practice
We’ve seen a financial services company begin to develop an integrative research function, and while the desire to increase the ROI of their research data is the same, they are looking to achieve it by creating a function within the organisation that brings together the different disciplines and data types under a new group within the business, which has responsibility – and budget – for both.
This truly integrative research function is then able to plan the projects and access the budgets seamlessly. This is often an institutionally slower process, but a more satisfying one.
From our observations, a number of companies we have worked with in an ad hoc fashion are now looking for full convergence. This is the only way to avoid the pitfall of the new uses of the existing data having to take budget directly away from the collection of new data.
Finding new value
Often businesses do not think of the total value of their research budget when they think about integration. Looking at bringing together and normalising product and concept tests or branding research across categories and geographies can often seem like a daunting task. It is, certainly in the beginning. Different techniques, different software and different agencies may be involved. The organisation itself might be decentralised in its structure, making it difficult for the central group to create the environment of cooperation required to make the project work. These types of barriers mean that, while ad hoc projects are clearly showing the power of integrative research, there is still a lot of friction involved in making integrative research work in an ad-hoc environment.
The implementation of an integrative research function with a set of new guidelines and process models for best practice will clear the way to ensure that in the future we see a much more dynamic research environment, which will better cope with the emergence and integration requirements of the new insight sources that will undoubtedly be appearing over the next few years. It’s going to be an exciting time to
be in research.