Best of both worlds

Qualitative or quantitative: that is the question. Put 100 researchers in a room to discuss methods of developing new products and they would probably come up with 100 different recommendations. Most agree that the best new product development (npd) makes use of both qualitative and quantitative methodologies, but opinions vary greatly as to emphasis and detail.

Martyn Richards, research director at SMRC, accepts the need for both approaches. “Quantitative helps define the target market next time you do focus groups, while groups give you the vocabulary to approach that market with words they associate with the brand.”

But he expresses concern that while manufacturers of new products often want facts and figures, research deals in concepts where qualitative knowledge may be more crucial.

“Quantitative assumes we’re all different and only by spreading the net wide enough can you get sufficient views to count as a sample. Qualitative assumes there is more similarity than that between people.

“Your eight housewives in Cheam may not be exactly representative of all housewives but they will be enough alike to make the re search worthwhile. With qualitative you can understand how people relate to brands and how those brands work. This aids positioning, and perception of positioning,” he adds.

Simon Avison, managing director of New Solutions, agrees that qualitative work is important, though he stresses it is no good in isolation. His company operates four stages of research.

The first stage aims to build an understanding of consumers who already buy products similar to those under development.

“Typical research at this stage is fairly loose qualitative work, getting groups together and watching them reacting to products. We’re exploring attitudes, so there’s no point doing widespread quantitative research,” he says.

At stage two the researchers segment the potential market and find out how many people hold the same opinions as the steering group. Those surveyed are asked a range of questions, possibly as part of an omnibus study, to determine their attitudes to the product sector – such as whether they buy the cheapest make or prefer a well-known brand.

Avison adds: “As a result, we do a cluster analysis where we get the computer to analyse our findings into groups of people.

“Then we devise a name and define characteristics to fit those groups. For instance, we might research people who buy baby products and come up with five groups, ranging from those who believe nothing is too much trouble for their baby and will pay premium prices, to those who are looking for the most convenient or cheapest products.

“We’ll overlay this with strategic marketing to enable us to focus on the groups that are really important to us and disregard those that aren’t going to fit our market.” (See box.)

Stage three involves in-depth consumer workshops where groups are recruited, often off the street, and new products will be introduced as concepts, pictures and finally in 3-D. Says Avison: “At the end of this process we have a good idea of two or three product concepts which could work.”

The key quantitative aspect of the process now takes place as re-searchers conduct a hall test, showing the product to a large number of people and attempt to gauge how much they would pay for it.

“Depending on what they say we can work out what will sell and how much profit it will make. Quantitative is good here but it is a blunt instrument. It can tell you how many people hold an opinion or which is best of two options, but not why. You need to know why first.”

The Craton Lodge & Knight (CLK) approach follows similar lines, with the research process breaking down into three component parts: exploratory, nearly always qualitative; executional, which can be qualitative or quantitative depending on the needs of the research; and evaluative, which must by definition be quantitative.

Chairman Creenagh Lodge explains: “When it comes to pros and cons everything rests on why you’re doing the research. If your question is ‘have we got a good hypothesis?’ then qualitative will work well, but if the question is ‘will the new product work?’ then quantitative is better.”

If small steering groups are the best method at the beginning of a new product’s development, then the emphasis at the other end of the process is definitely on quantitative procedures.

Much of research companies’ current work aims to improve the quality of information gleaned from these mass surveys. Research International has a system called MicroTest which not only tries to predict how big sales might be but why. MicroTest is used after the basic npd concept has been established.

Research International npd director Julian Bond says: “The principle behind MicroTest is simple; it aims to measure likely trial of a new product and then to measure likely repeat purchase. We start by interviewing 200 or 300 people representative of typical households, usually in their homes. We will show them the concept and establish their level of interest in it. Then we leave it with them to test and come back later to see what they thought of it.

“Alternatively, if the new product packaging is already finalised we might take them to a hall with a simulated store shelf display. Stage two is when we establish how likely they are to repurchase the product and how frequently. Models take into account the numbers of people who, in the real world, won’t be aware of the new product and/or won’t try it out, and from this data we are able to predict volume sales.”

Not all research companies see npd research as a choice between interviewing small or large groups of unknown consumers. John Orsmond, chairman of Advertising Research Marketing (ARM), argues that the way forward is database-led quantitative research. This gives clients a greater and ongoing understanding of what their customers, markets or products are doing.

“Clients often look to research to tell them things that their database should already be telling them,” says Orsmond.

The ARM approach is to interview large numbers of people but to ask them increasingly qualitative-style questions. With this fusion technique, each time a customer is interviewed the researcher asks a couple of extra questions such as readership habits or lifestyle preferences. Orsmond explains: “It’s a process of accretion: as you add more attributes you make an exponential gain which will help build your database and help you ask more intelligent questions later.

“This combination of database and research techniques with technology gives clients rolling knowledge on a continuous basis. Unfortunately a lot of companies are still not pulling their various sources of data together and many rely too heavily on focus group findings to make what often turn out to be poor management decisions.”

Clearly the melting pot of UK research already contains a wide range of attitudes – from Avison’s standpoint that quantitative is best used as a final check, to Orsmond’s conviction that qualitative is mainly a refining tool.

But Phileas Fogg, one of the great marketing success stories of the past decade, was launched without either qualitative or quantitative research but with manufacturer Derwent Valley Food’s conviction that there was a gap in the market for upmarket snacks.

Spotting a quirky niche like this and then getting the product and the marketing right is a great achievement. But was it a lucky guess, or is instinct actually the best research tool we have?