A waste of time and money

Something is missing somewhere. Recently I chaired the Research Data Matters conference in London and heard from a long list of forward-thinking companies about what research should be doing. In a nutshell – it must translate into dollars and not only provide interesting insights but also must truly benefit the bottom line.

It does not do this today as it does not integrate with other business intelligence disciplines in an effective way. Much duplicated or unnecessary effort is due to the lack of an integrated approach.

There were whole systems of customer intelligence with enormous levels of sophistication introduced that do not use research or research data at all. The conclusion of the conference clearly pointed to the fact that people don’t need to be asked as many questions as is traditional or perhaps they do but you just don’t know or can’t justify which ones to ask. What it seems is needed is an integrative research approach.

Now, given the readership of this magazine, and as I fear the backlash from this article, I would like to discuss why so many people believe such a wide margin of truth exists and what might be the reason behind this discrepancy. I will also discuss how you might go about finding and realising the savings that an integrative approach may bring.

Track to success
Let’s start at the beginning. There is now a massive availability of non-research data sources from which consumer insight can be gleaned and there is also a relatively untapped world of cross-study analysis within research. For example, I am writing this article on a train and have just bought a cup of coffee. The receipt lists my train and seat number along with what I bought. This was even lovingly printed on paper for me to take with me. Sadly it is probably not expensible, but it does indicate the worth of questions like, ‘when did you last buy a cup of coffee on a train. How big was it?’

The two rivers
Back in 2003 myself and my then colleagues at WPP and SPSS tried to explore this through a research study and whitepaper called Two Rivers. This study surmised that the worlds of market research and business intelligence were like two swift flowing rivers of information that would come together in the future as they answer many of the same questions. However, there were many barriers to realising this synergy, such as the poor quality and immaturity of corporate data warehouse schemes combined with similar problems in the worlds of survey research and additional issues on data consistency and ownership.

Much has changed since 2003. The data warehouse projects are now more complete and technology has moved on significantly. That said, in real terms there has been little progress in terms of including MR and its wide-angle lens of insight into this world. Why? In my opinion rapid consolidation in the industry has been effective for lowering costs to clients but ineffective in delivering the innovation in technology required to realise the true leaps in value that are potentially available. For example I have worked in the MR industry since 1990 and in my first few weeks I learned computer skills that I could still use to make a living in the industry today. I challenge the reader to think of many industries where computer skills that are 18 years old are still pertinent today. At that time I hadn’t seen a mobile phone (other than the yuppies’ bricks in the bar where I worked) and was five years away from even hearing the word ‘internet’.

So from 1990 to 2003, market research data practises changed really very little and even in 2009 they are still much the same. For example most practitioners are likely to use Microsoft Excel for the bulk of spreadsheet analysis and we all find it very easy to share work with others and work collaboratively using it. However, for our specialist MR tools that allow us to analyse our survey data we still have many to choose from.

In one of my last experiences working in a large agency, we conducted an audit of all the platforms we were using for both data collection and analysis and found we were using over 40 different products in each category. With this level of friction in the research process, it is easy to see why the industry has failed to keep pace with the overall business intelligence market in terms of developing an integrative approach. While you could argue that it is a different discipline, it is still the voice of the consumer within a business, so should be simpler to integrate and use.

Business intelligence
A better way to think about this is to think about the flow of information from one channel to the other – specifically from MR to business intelligence or vice versa. At the Data Matters conference Martin Hayward from Dunnhumby made a very important point. He said: “We work from what you did outwards to why you did it.” Most market research works from questions about why you do things and try to predict what you will do. Surely within this there is an optimal model that will help you ask only the most pertinent questions in the most pertinent way and not waste effort on information that is better sourced elsewhere.

Here are some examples of how an integrative approach can be more efficient and help to realise the savings that so many people believe are out there.

Savings from redundant research
Organise different research under the same platform to achieve synergy. By combining various product/concept test research, you are often able to answer new business questions and eliminate funding superfluous research. Synergy is obtained through combining data,
a) from different time periods (trending)
b) across product/brands/concepts/ business units.
 
Better insights from linking different sources of research data
Additional synergy can be captured from across product linkages as well as trending. As a case in point let’s use chocolate. Across the board men associate chocolate with comfort; whereas women associate it with indulgence, which has great implications for how you communicate with them. While an individual study may provide the same information about a specific concept (e.g., a white chocolate with a bitter orange flavour), we would not know that, in general for women, chocolates are tied to indulgence. With respect to trending, only by linking and creating a trend line for appeal, will we know that chocolate has, over the years, consistently lost appeal among men but not among women.

Research redundancy
We can also think in terms of research redundancy. If we asked a question within a category – such as customer satisfaction or new product development – we are building a picture of the consumer that can be looked at across the surveys we conduct. For example how many times have we ever asked a certain question and how has the context or relevance of that question changed over time. From our own experience we have databases with over half a million responses that show no or little change over time. Using this knowledge in a structured and data-centric way can give us the tools we need to manage our research process more effectively and ensure we don’t duplicate efforts in our research or collect information that is perhaps better captured elsewhere for the sake of it. A primary example is the receipt from the coffee I am now drinking.

Data sharing
There are equally opportunities for data sharing. If many of the data points that we collect are static, why should data not be shared in a way that will, while ensuring confidentiality, provide researchers and their clients with a window on what is genuinely different and what is genuinely insightful from a research study. For example if we know that the primary driver of purchase intent is age, regardless of the product being tested, then why do we not analyse what we already have to make a better-targeted research study and avoid duplication of effort?

Linking research data with other sources of data
For this let’s take an example looking at doctors’ prescription patterns for a new drug. By linking satisfaction and effectiveness data to prescription data you can provide insights regarding both optimal quantity of sales calls as well as the quality of messages to a particular doctor. So if you know that cardiologists tend to write more prescriptions when the salesperson is able to demonstrate “knowledge and competence about the disease state and product benefits” whereas the oncologist writes more prescriptions when the salesperson could show that “he/she cared about the physician’s business practice”. These types of insights lead to better understanding of what drives volume and share of prescriptions for different drugs.

Conclusion
It is easy to see how an MR strategy that is not aligned with other business information streams in a seamless way can make you spend money that you don’t need to. My advice to those who do not believe this is to conduct an audit to find out where and by what means each piece of the research puzzle can be best answered, and by what channel, before another research study is begun.

You’ll be surprised by what you find and how much you can save with an integrative research strategy.