This stage of market research is completely an art. Being able to interpret better requires a couple of things. You should firstly know what the data is all about. Market research best practices require the researcher to know the data and the metrics well enough to know what they indicate. In addition to that the researcher also needs to have a keen sense of understanding of the niche that he is doing marketing research in. A good understanding of the industry enables you to understand the metrics and then apply them to the situation. Meaningful recommendations can only happen if you understand the industry.

When you are starting out with interpretations it is a good idea to involve other people like colleagues and customers. Discussing your thoughts with them will give you a different perspective about the organization and the data that you have seen. Sometimes just chatting up with customers can also give you insights that you could never have come up with unless you had looked at the data in a different light.

The implementation stages of market research include questionnaire development, data collection, and analysis. At the questionnaire stage avoiding questionnaire bias is important. The manner in which you ask a question can sometimes affect the response that you get. At other times, there are some socially unacceptable answers that you will never get unless you use projective techniques to mask them. All attribute questions should be rotated among respondents so that order bias does not creep in. Unless you are careful about this, the data that you get from the survey will not be useable.

Some of the other aspects that need to be remembered are specific to the kind of research that you are doing. If you are doing a product test then you may want to decide whether you want the same customer to try out the various product variants that you want to test or whether you will prefer to have monadic panel testing. The one that you want to do depends largely on the product category that you are researching and other factors. These are not dictated by the book and have to be decided upon based on experience and practice.

Then there is the aspect of deciding the analysis that you want to do to confirm or negate the hypothesis. The part where you decide the specific analysis is an art but conducting the specific analysis is a science that you can learn from the book. For example, when you want to understand the cause and effect relationship between satisfaction on specific attributes and the overall satisfaction, you need to use regression analysis. Correlation analysis can only tell the simultaneous existence of two variables and does not indicate cause and effect.

While the fundamentals of market research are seeded in statistical analysis, there is a lot more to market research then just statistical analysis. As they say data can be molded to state what you want to see. In addition to that, if the research is conducted among the wrong demographic segment or a biased sample is chosen, the entire results of the market research can take you in a direction that will end in disaster. You will create strategies based on wrong interpretations and none of these will bear fruit.
It is important that market research should be done in the right manner and in order to do that, you have to ensure that you take care of some tasks before you start the market research process. Below are the details of what you should keep in mind at these three stages of market research.

This is the stage before you actually start the process of market research. And this starts even before you start to make the list of questions that you need to ask the customers. The first thing that is important is to understand why you are doing the research. Doing research just to gather some data does not make sense and you are likely to get junk data that you cannot possibly use to your benefit.

The objectives of the research are extremely important. Even if you are not required to justify the research to anyone, make sure that you list what you expect from the research at the end of the process and the answers that you are seeking. Market research best practices also require you to make a list of some hypothesis that you have from your understanding of the business so that you can either prove or negate the hypothesis after analyzing the data.

One of the major mistakes that people do is to forget to make a list of the information areas that will be covered in the research followed by an analysis plan that will be used to answer the questions that the marketing research has set out to answer. The reason why market research best practices requires you to do this is because many times one finishes data collection only to realize that you need a specific variable to be able to answer some questions.

When William Sealy Gosset published the t-test in 1908, he was thinking business (specifically Guiness’), but not necessarily marketing. More than 100 years later, this test has found its way into many fields including marketing. In this post you will find two examples of applications of the t-test in marketing. For more information about the t-test visit this wiki and for details on how to run a t-test using Mineful, go to our t-test analysis page.

Example 1: Pre-Post Marketing Campaign
You have started a new marketing campaign and are seeing the sales numbers come in. The numbers look slightly higher year-to-year and you want to know if they are indeed different or are they slightly higher by pure chance. Since you are taking measurements on the same “process”, but in two different occasions, you want to run a “Paired t-test”. The paired t-test eliminate the effects other confounding factors by matching observations of similar traits. For example, sales in October 2009 vs sales in October 2010. Or in observational studies, the blood pressure of a person before the pill paired with the blood pressure of that same person after the pill. Join that with many other people or months of sales data, and you have a paired sample.

Basically, the t-test will determine whether the average sales numbers are significantly different before and after the new marketing campaign. Go ahead, upload or connect your before and after campaign data and determine if something has changed significantly.

Example 2: Average Satisfaction Between Two Groups
You have survey data and used the Average by Category analysis to see the average overall satisfaction by gender, market, customer segment, or other group. The results show that in Chicago the average satisfaction is 5.6 and in Madison 6.1. But is the .5 difference significantly different? Should I take action and question my service people in Chicago because of this difference? A two-sample t-test allows you to test if this difference occurred by pure chance or if there is something else that’s driving customer satisfaction lower in Chicago.

There are many other uses of the t-test in marketing. Want to share one with the internet? Comment below and tell us how you used the t-test in your marketing research.

There are various studies that mention how the cost of a new customer acquisition is ten times more than retaining one. And therefore companies know that when they have a customer they need to hang on to what they have. Proper customer satisfaction analysis is what can help in arriving at the action that needs to be taken to retain customers.

Just knowing the level of customer satisfaction is not enough. The level of satisfaction does not give diagnostics with regards to what should be done to improve the score and increase the level of happiness of customers. Neither does it help in understanding the various potential areas that can create customer delight.

What this essentially means is that there are various other metrics, other than the overall satisfaction that need to be tracked. A customer satisfaction dashboard is a great way in which various parameters can be tracked on a regular basis. There are also some software options that can provide customer satisfaction alerts in case the satisfaction of a particular area, region, division or center dips below a certain level.

Some of the must have metrics that should be used to create a customer satisfaction dashboard are listed below.

1. Customer Satisfaction Index – Most obviously, this is the most important parameter that should be used to create a customer satisfaction dashboard. The index should take into account all the information that is captured at an overall basis and create a ‘one-number’ index that can be tracked across various players in the industry.

2. Customer Satisfaction Index by Segments – While the overall index can provide an insight into the performance of the company or the brand as a whole, it is necessary that the same parameter should be tracked for various regions, divisions, dealerships or stores to be able to assess their performance. It is only fair that the various segment stakeholders be able to see their performance easily since customer satisfaction index is often an important KPI that is set for managers.

3. Trends in Satisfaction Levels – Not only is it important to look at satisfaction levels at an absolute level, the trend is also important. A time series customer satisfaction dashboard helps in understanding if the performance if dipping or not. This is helpful because at times a dip may not apparent if the company is still performing well. But a trend customer satisfaction analysis can help in taking corrective measures in time.

4. Customer Typology – This is an interesting analysis that you can do to understand the spread of the customers that you have. If you create a grid of customers based on low loyalty, high loyalty, low satisfaction and high satisfaction you can classify your customers into mercenaries (low loyalty, high satisfaction), evangelists (high loyalty, high satisfaction), hostages (high loyalty, low satisfaction) and goners (low loyalty, low satisfaction). This customer satisfaction analysis can also be done on an ongoing basis to understand the changing face of the customer base.

5. Driver Analysis – This analysis can be performed to understand the attributes that are genuinely important. It can be done by regressing the satisfaction scores of individual attributes to the overall satisfaction score. The higher the beta coefficient, the more important the attribute is towards increasing customer satisfaction.

6. Key Attributes – While it may not be able to put all the attributes that are being tracked on to a customer satisfaction dashboard, it may be pertinent to identify the top five attributes and create a dashboard for the same. This will help the managers to ensure that they never lose sight of the parameters that are extremely important for the client.

7. Issue Tracking – This is a relatively qualitative part of the dashboard where extremely disgruntled customers can be brought to the notice of managers by customer satisfaction alerts. These special cases can then be tackled to reduce and cap the damage that has been done. This is an essential part of customer satisfaction management since an unhappy customer can bring about immense bad word of mouth.

8. Changes in Important Attributes – While this is a customer satisfaction analysis that you do not have to do every time, creating such a dashboard and updating it once in a while is helpful. This is because you may not generally see changes in the attributes that are important very frequently in the industry unless you are in the technology niche. But creating a dashboard like this will help in remaining ahead of the crowd.

9. Comment Scroller – A comment scroller is a customer satisfaction analysis that allows a company to understand the positive and negative comments that customers share. Managers can look at this scroller to get insights into the manner their customers think.

10. Competitive Indexing – Customer satisfaction index can never be treated in isolation. It is always interpreted in relation to the manner in which competition performs. A competitive customer satisfaction analysis can be converted into a dashboard where the difference between the scores of the company and the nearest competitor can be seen at all times.

According to AcuPOLL Research Inc, the proportion of product launches that fail is anywhere between 80 percent and 95 percent. And the amount of money that is completely wasted in research and development, creating the product, scaling it in manufacturing units and then marketing the product is something that one can only imagine.

The right kind of market research can ensure that you reduce the risks of failure. While market research was deployed towards the end of the product development stage, there is adequate proof of the fact that if you research the product launch at the ideation stage or even before that, you are likely to reduce the risk considerably.

This means that if you have researched the market and competition carefully, you will know the specific niches that are too competitive. In addition to that an exploratory market research conducted to understand the needs of the consumers can help you start the research and development in the right direction. This means that you can increase the return on investment on the dollars spent on R&D too.

When you use market research at each stage in the product development process you ensure that you are not just creating a product but a customer experience that has its seeds in customer need in the first place. Some of the questions that need to be answered include the specific need that the product solves, whether the product is a ‘must have’ or just another option, who the target audience is, the positioning that is likely to resonate with them and the price at which the target customer will buy.

The product development process starts with an evaluation of the market – the competitiveness, potential, growth opportunities, scope of expansion and the like. Post an analysis of the market, companies should get into an evaluation of their strengths and weaknesses in the areas that they are getting into and what the new launch adds to the company and the business.

The Stage Gate Process is considered to be the 5 step product development process that includes:

• Idea generation
• Idea screening and evaluation
• Business analysis
• Development
• Testing and commercialization

Market research can be performed at each stage to increase the chances of moving in the right direction. So while you can use expert interviews and synectics (interviews and focus groups with creative people) for idea generation, you can do a quick quantitative concept testing to screen out the ideas that do not generate appeal among the target audience.

The business analysis may require you to take on a slightly more complicated research that can help you estimate the volumes and growth based on specific inputs.

While there are various market estimation models that exist, the estimation is as good as the inputs that you provide with regards to marketing, logistics and distribution.

The development stage may require a few rounds of market research as the research and development fine tune the prototype of the product that is being created. This may involve an in-use or at-home product testing or product testing at a venue depending on the specific product in question.

Before commercialization and final product launch, there are aspects other than the product that need to be researched. These include promotional material, positioning statement, advertisement and packaging, if applicable.

With a large amount of completion in every industry, there is hardly anyone who does not keep a close watch on the basic financial parameters of a business. Some of the key parameters that every business measures include volume, revenue, profitability, investments, return on investment, volume growth, revenue growth, conversion rates or trial rations, retention and more. This aspect of the task takes care of the measuring and monitoring of the business parameters. Tracking these parameters over time gives the managers an idea of the progress of the business.

However, despite this level of analysis there is an element that is missing. While these reports help in understand the state of the business as it is, they do not provide any insight into what steps should be taken if the sales are low. Neither can these tell you the specific parameters to be changed so as to ensure better acceptance of the product or service in the marketplace. This aspect of managing these parameters can only be achieved if you know the factors that affect these parameters. Basically, you need to know the key factors that drive purchase to be able to ensure higher sales for your product.

What is Driver Analysis?
Key driver analysis is a statistical tool that uses multiple regression to identify the specific parameters of a product or a service that drive a particular action. Some of the questions that driver analysis answers are:

• What aspects of my restaurant business results in higher customer retention?
• What are the specific product features that lead to purchase?
• What are the specific areas that I can choose to ignore without losing out on my current clientele?
• Is recommendation for my brand occurring due to the product performance or service satisfaction?

The interesting aspect of key driver analysis is that you could use it to understand any of the variables that you need to understand. This means that you could understand what drives loyalty, purchase, repeat purchase, satisfaction or recommendation.

It is important to understand that direct questioning does not provide accurate results all the time. While some respondents are likely to give politically right responses, there are other issues too. Direct responses result in hygiene factors being confused as drivers. This is mainly because these are factors that are never really stating as ‘not important’.

How does Driver Analysis work?
Driver analysis should be used with care since inadequate understanding of the manner in which the statistical tool should be used can result in wrong deductions. It is extremely important to ensure that all the relevant parameters for the product or service are covered in the overall data collection. In addition to that the wording of the parameters is also important.

Driver analysis includes a dependent variable and various independent variables. The dependent variable is the aspect that you need to understand better. So when you want to check out the factors that drive purchase, the response to purchase become the dependent variable. Other variables that you have captured data on like performance on specific attributes can be considered to be independent variables. It is assumed here that these independent variables shall in some way drive the dependent variable. The whole idea is to understand the manner in which these factors drive purchase and the extent to which they do.

The key output of a driver analysis that uses regression models is a score out of a total of 100 indicating relative importance of each of the independent variables that have been tested. A table of importance that immediately allows you to understand the features and parameters that are truly important helps you in deciding where to focus your monies.

Benefits of Driver Analysis
The main benefit that key driver analysis lends for your business is an understanding of the factors that you need to tweak to make things turn. For example, an understanding of the features of a product that drive purchase can help you devise a communication strategy around the feature ensuring better sales. If you are concerned about customer retention for your business, driver analysis to understand the factors that drive retention can help you better service levels.

It is practically impossible to allocate funds towards improving all areas of product features or service in a business. Careful prioritization is essential. Driver analysis helps you prioritize the aspects that you need to concentrate on for short term and long term success.

Customer service professionals will tell us that each individual is different and therefore needs to be treated accordingly. While this is definitely true, retailers and marketing personnel cannot service this individuality when they are marketing mass products. Competition has made it necessary to ensure that your product is not a jack of all trades. Differentiated products for the specific needs of different kinds of people are a necessity today.

Market segmentation is a process that has helps in splitting and grouping customers into various clusters. The customers belonging to a single group are considered to be homogenous and comparable. A market segment is homogenous within and heterogeneous without. If the marketing segmentation has been carried out properly, a market segment is expected to react similarly to a market stimulus.

According to Philip Kotler, an effective marketing segmentation exercise should result in segments that are measurable, relevant, accessible, distinguishable and feasible. These are aspects that ensure that the marketing segmentation has been done on relevant parameters and that further action on these groups is practical.

Marketing segmentation helps marketers understand the otherwise extremely heterogeneous market so that the strategic and tactical marketing decisions can be taken in accordance with customer needs and desires.

Understand the market

At the first level marketing segmentation helps in a better understanding of the market in relation to the industry in question. When each individual is grouped, it is easy to understand the common characteristics of the group rather than assessing each individual customer (something that is quite impossible too) or treating everyone as a single group with similar needs and preferences.

Create product variations effectively

Once you have a certain number of groups that define your market, product variations can be launched depending on the specific needs of the group. So while those who treat a bath as a time to pamper themselves would buy soap with a moisturizer added in, those who feel that a bath is for cleansing may well prefer to have one that fights germs. An insight such as this can help you develop relevant products for specific people.

Perfect your positioning

Not only does marketing segmentation help in creating new products but it also aids in the creating of the appropriate positioning. The words used in the body copy and tag line of the advertisement can be modified based on the expected reaction of the targeted demographic too.

Create customer empathy

With targeted and more relevant products and positioning, the marketer can be sure to create an empathy with the customer; a bond that is difficult to break, thereby encouraging higher customer retention.

Identify need gaps

Conducted for exploratory research, marketing segmentation can allow you to identify need gaps in the market so that the research and development division can be given specific parameters to work with. Sometimes the gap identified can also be used for repositioning of an existing product.

Focus your resources

When the resources are limited, identification of the relevant market segments that are showing growth can be used to ensure that the budgets are allocated in an appropriate manner.

Targeted media selection

An understanding of the customers in each segment can help in better media selection, media scheduling

Using Cluster Analysis for Marketing Segmentation

Cluster analysis is a multivariate statistical tool that is used for marketing segmentation. It helps in identifying groups of people from the sample that have similar responses with respect to any of the parameters that are used for clustering.

Cluster analysis can use any data that has been collected to segment customers into various groups. Some of the broad parameters that can be used are:

• Behavioral parameters like purchase history, amount of product bought, product usage details, brand loyalty, shopping behavior and more.
• Attitudinal parameters that can be used in cluster analysis include (among others) reaction to various belief statements, likes and dislikes measured on a scale and general tendency towards innovation.
• Psychographic variables like lifestyle, values, personality and more can also be obtained using pre tested scales for the same. Customized ones can also be developed depending on the category.
• Demographic data is a key part of the analysis because irrespective of whether it is used in the cluster analysis and segmentation analysis or not, this data linked with each respondent helps in identifying the kind of target audience that you are looking to cater to.

With Mineful all that you need to do is the upload your database with the survey details and the clustering will be done with a few clicks. You can expect to get the specific clusters, cluster sizes, cluster naming and a detailed understanding of each of the clusters in terms of the relevant parameters. Not only will you get an understanding of each cluster but also the sketch of a typical customer in each group.

Mac Users Like It, Of Course
According to our non-scientific survey of 48 respondents, those who consider themselves to be a “Mac” are more likely to purchase the iPad than those respondents who consider themselves a “PC”. Mac users have the highest interest level in an iPad, followed by those who consider themselves neither a Mac or a PC. Mac users are significantly more interested than PC users at nearly twice the interest level.

Elections place a heavy burden on mail carriers. Their bags are stuffed with political advertising telling voters why they should support this or that candidate. Most of this mail is ignored or discarded with barely a glance, mainly because it does not address issues that really matter to people.

It doesn’t have to be this way. With a little effort and a relatively small investment, candidates can conduct their own political research using simple survey tools to discover what people care about and what they are looking for in their elected officials.

Discovering What’s Important to People
Effective political surveys have a lot in common with surveys used for market research or any other effort intended to learn about personal interests and preferences. They should begin with an upbeat invitation, such as “Thank you for taking the time to make a difference in our community,” or “Thank you for letting us know what’s on your mind.”

Like other types of surveys, political surveys should be brief and easy to complete. The early questions should be pleasant and fairly simple. Questions about difficult or controversial topics should be put off until the end of the survey, when people feel more comfortable about expressing their opinions. If these questions appear too early, people may abandon the survey because they find it too challenging.

On multiple choice questions, it’s good to give people an option of “Don’t Know” or something similar. On other types of surveys, the Don’t Know response usually doesn’t provide useful information, but on a political survey it can be highly instructive. If a large number of people don’t know what they think about an issue, a candidate might decide that the issue is not very important to them. Of course there are other ways to determine what people care about most. The simplest is just to ask them to rank a number of issues in terms of their importance. Responses to such questions probably won’t lead candidates to change their positions, but they will help candidates decide which issues to emphasize.

The Online Advantage
In the past, conducting political opinion surveys was too expensive and labor-intensive for most candidates. The Internet has changed that. Online surveys yield results very quickly, often within a day or two after they are posted. Once a survey has been posted online, it costs very little to collect and interpret the results. Online surveys can use graphics, colors, and even video to make it easy and entertaining for people to complete them. And, in the Facebook era, many people will feel more comfortable sharing their opinions online than they would responding to a phone survey or even filling out a paper questionnaire.

Online survey software, such as Mineful’s products, can make it easy for candidates to get the most out of surveys. For example, candidates can use this software to sort responses by location or by political affiliation. They can also see how opinions on one issue might be correlated within opinions on another. Affordable, easy-to-use software tools can make it easier for candidates to deliver the right message to the right people.