Acting on Analytics: The Official Mineful Blog

Why Customer Satisfaction Should be Trended

A typical customer satisfaction survey generally captures data with regards to overall satisfaction with a product or service. To assess performance and to understand the specific areas that the company should invest in, performance on various other parameters

customer satisfaction trendfor the company and competition is essential too.

Not only does a customer satisfaction survey detail the company’s performance vis-à-vis competition, an analysis of the performance on parameters can give out an action plan as well. This is an analysis based on the importance and performance rating on various parameters and attributes.

While the benefits of a customer satisfaction survey may be understood and appreciated, many people feel that conducting a survey once is quite enough to understand where the company stands and the areas to focus on. However there are various reasons why a customer service survey should be trended. Here are some of them.

  • a. The Satisfaction Index – A customer satisfaction survey provides a concise one number index that can be compared across competition and over time. This means that a single chart that details the satisfaction index over the years along with the performance of competition can give you a picture of how your fare among your customers.
  • b. Drill-Down – The same index can be evaluated among different customer groups – loyal versus floaters, new versus old, open to competition versus closed, young versus old and other parameters that may be relevant in the industry being studied.
  • c. Stay in touch with your clients – The market place is a dynamic place. And unless you keep an eye on the manner in which your customers are moving, you are likely to lose them before you even realize it. Trending the customer satisfaction survey on a regular basis helps in keeping an eye on the movements and commitment levels of your customers.
  • d. Understand the ever-changing customer needs – The importance of conducting full-fledged customer satisfaction survey on a regular basis cannot be stressed enough. In the ever-changing market that we work in, it is not only the satisfaction levels that change. The needs and desires of the customers and the expectations also change. What this means is that the attribute landscape that is created at one point in time changes rapidly. Some of the attributes that may have been identified as growth areas or areas that can increase customer satisfaction may soon become hygiene factors as more and more competitors adopt the attribute and start offering it.

There are some customer satisfaction indices that are based on purchase intent and others that take various factors such as willingness to recommend, intention to continue using, advantage over competition. While some use a black box weighting method, there are others that are relatively simple calculations. What is important is that you use one system and track it over time in the same manner so that you can understand the trend and take decisions that are meaningful. This is easily done with the Mineful dashboards since they allow you to see dashboards online at the click of a button.



Preventing Survey Bias – Part 3

Bias caused by the Respondent
Irrespective of what you do, there are times where the respondent will bring in a bias based on his or her own personality, reasons or past experiences.

  • Social appropriateness – This is a bias that mainly creeps in face-to-face interviews where respondents give answers that they feel are socially acceptable. For example, stating that they tend to get angry with their children frequently is not something that any mother is likely to agree. Unless you have a rebel of a respondent stating that she does not like making friends is not a response anyone is likely to give. This phenomenon is sometimes also called the Best light Phenomenon and can be avoided by not putting respondents in a self effacing situation. Projective techniques can be used in such situations.
  • Acquiescence bias – Also called the friendliness effect or the yea saying effect, this is a systemic bias that can occur if the interviewer or the respondent is over friendly and starts to agree with whatever that is presented.
  • Halo effect – A bias that is caused by the respondent where a respondent carries an overall positive or negative impression about an element and colors all responses with regards to that positively or negatively. In consumer surveys, this often occurs in responses towards the brand used most often where all positive attributes are associated strongly with their own brand.


Preventing Survey Bias – Part 2

Response Bias – Bias due to Measurement Errors
The way in which the question is asked, the wording, the flow, the environment in which it is asked, other people present and the interviewer contribute to measurement errors.

  • Leading questions – Questions that are leading give away the desired response in some manner. If you state what other people are saying about a new product and follow it up with a question with regards to the respondent’s opinion, chances are that the respondents will tend to agree with what you have just mentioned. Such a bias can also be caused verbal and facial expressions in a telephonic interview or a personal interview. Biased questions should not be confused with counter biasing, a technique that is used to increase the willingness of a respondent to answer honestly.
  • Unbalanced scales – When you give options to the respondent to choose from in a close ended question, deciding on these is critical. If a satisfaction scale provides options such as – extremely satisfied, satisfied and dissatisfied, it automatically leads the respondent to choose the satisfied option more than the dissatisfied option. A bias can also creep in if the scale is not valid and therefore does not measure what it is supposed to measure due to the wording of the question.
  • Implicit alternatives – When options are provided to the respondent to look at and then respond, leaving out some options can cause bias. For example, if specific brands are left out of the options for an aided awareness question, these brands are likely to get lower awareness scores, something that may not have been the case had the specific alternative been included in the list.
  • Order bias – Also called position bias or sequential bias, this is caused when respondents prefer to choose specific options based on the position that they are in on the list. Mail interviews are most prone to this bias because personal interviews and internet surveys tend to resolve this issue adequately by the method of rotation.

The issue of leading question, unbalanced scales and implicit alternatives can be tackled by ensuring that you pilot the questionnaire with a set of respondents and fine tune the wording. Trying to keep these potential errors in mind and understanding the various ways in which a question can be interpreted also helps in creating questionnaires that are not biased. It is also essential that you use the funnel approach in deciding the flow of the questionnaire and start with general questions before getting down to the specific ones.



Part 1 of 3: Preventing the 3 Biggest Types of Survey Bias: Selection Bias
One of the first things that any market research professional needs to understand is the fact that survey bias is an inevitable part of almost all primary data. Different types of survey biases can creep into the data based on the number of people that are questioned, the specific profile of people who are questioned, the manner in which the questions are worded and the flow and structure of the questionnaire. In some cases, survey bias is quantifiable and in others it is not. While the sampling error of a particular sample can be calculated to estimate the range within which the actual responses will fall, the errors caused by other kinds of biases are less easy to resolve.

Selection Bias – Errors and Bias due to Unrepresentative Sample
The selection of a sample plays an important role in survey research. When the method of choosing a sample is faulty, the individual respondents of the sample may not collectively define the population accurately. A bias that results from underrepresentation, overrepresentation or improper selection is called a selection bias.

  • Under or over representation bias – This bias is caused when a specific kind of sub profile in the target respondent is over represented or underrepresented. For example, if the study aims at studying women in a developing country, using an Internet survey may result in a biased sample. As the Internet penetration in developing countries is not too high, it might result in a sample that is in favor of high income groups or highly educated women who have adapted to the internet faster than some of the others.
  • Non response bias – A non response bias occurs in specific parts of the survey when respondents refuse to answer specific questions. At other times, specific profiles of people may refrain from responding to a question resulting in a biased statistic. A mail survey is typically prone to the non response bias.
  • Voluntary response bias – Surveys that adopt the voluntary method of data gathering are prone to voluntary response bias. Typically, those who have strong opinions about the topic are likely to make the effort to respond to a question, thereby causing the data to appear polar in nature.

Selection bias can be avoided by ensuring that the sample that is selected is random. Random sampling can ensure that the sample is chosen purely by chance and every eligible respondent has an equal chance of being selected for the survey.



The news about the proposed Islamic center about two blocks away from Ground Zero has led to a lot of fervor in the country. While the Mayor of New York Michael Bloomberg is in favor of the proposal and backs Imam Feisal Abdul Rauf, there are others who oppose the approval as a another feather in the cap of the jehadis. The specific proposal is to create a 13 storey cultural center for Islam along with a mosque a few feet away from where the Twin Towers stood a few years back. The project is to include an auditorium, a swimming pool, art exhibition space, restaurants, bookstores and also a mosque.

Republican Sarah Palin states that it would be a ‘stab in the heart of the families of the innocent victims of those horrific attacks’ and many other Republicans voice her opinion as the New York Mayor stands his ground in the belief that ‘Muslin religious freedom must be respected’.

Those opposing the mosque have tried to invoke the Landmarks preservation Commission but the vote goes towards the architectural development of the disused coat factory. There are also some who believe that the construction of the Cordoba House Cultural Centre would be an act of good inter-religious relations.

While some like US House of Representatives speaker Newt Gingrich call it an ‘act of triumphalism’, others feel that this could be the tipping point in the Muslim-West relations. Polls show that most feel that the building of a mosque is not an issue but the fact that the location is so close to Ground Zero seems tasteless. Some are even questioning the money trail that leads to various countries like Saudi Arabia, UAE, Qatar and Iran.

Even as protesters accumulate at Ground Zero to voice their concerns about this event, various polls across the country are being conducted to understand the opinion and emotion of the general public.

According to the Quinnipiac University Poll conducted among 1183 New Yorkers in the last week of June, 52% did not want a mosque to be built as against 31% who were in favor of the decision. The Quinnipiac University Polling Institute Director Maurice Carroll stated that while New Yorkers are considered to be relatively tolerant of other religions, they were opposed to this decision by large. Further analysis of the data of this survey showed that Manhattan had the least number of people opposing the proposal with only 36 percent rejecting the idea as against Staten Island where a whopping 73% rejected the idea of a mosque two blocks away from Ground Zero. Among the other classifications the black voters rejected the proposal to the least extent (45 %) while 56% of the white voters rejected it as did 60% of the Hispanics. Rejection was high among Jews and Catholics (66 percent) as against 46% among Protestants. In this poll, 22% of the New Yorkers also felt that ‘Islam encourages violence against non-Muslims’.

Another CNN Poll that surveyed 1009 adults across the country summarized that 68% of Americans were opposed to the approval of the proposed mosque and cultural center near Ground Zero. The poll also reveals that 43% of Democrats support the mosque while only 17% of the Republicans are in favor of it. Support among the independents lies in between at 24%.

The Rasmussen Reports national telephone survey among 1000 adults dealt with more than just the proposed mosque issue. While only 20 percent of Americans were in favor of the proposal but it was also stated that about 22 percent were following the story ‘very closely’ and another 29 percent were following it ‘closely’. Some argue that the mosque will honor the 9/11 victims since it promotes peace among Islam and the West. However, this confidence is not felt by more than 30 percent of those who participated in the survey. The survey also showed that those above the age of 40 were more inclined to oppose the mosque than the younger ones. While the Democrats are split half way on the issue, the Republicans clearly (76 percent) oppose building the mosque near what was the site for the Twin Towers. Not many (only 13 percent) feel that allowing this will improve America’s relations with the Muslim world.

There are differences across these polls with regards to the proportion of people opposing the Islam Cultural Center that probably emanate from the manner in which the question is asked or how the proposed building is being referred to – an Islamic Culture Center, a mosque, a mosque near Ground Zero, a mosque in the Manhattan area or even a mosque near the Twin Tower site. However, the overall results seem to indicate that a majority (irrespective of whether we look at Americans, New Yorkers, Republicans, Jews or young and old) does oppose the proposal being agreed to by the Mayor of New York City.



Our Bigger and Faster New Home

This past Sunday, we moved to an all-round better datacenter. With this move, we introduced more, faster, and newer servers to help you use Mineful more efficiently.

The move was an ordeal. It took 2 hours longer than expected, but we got home around midnight happy that Mineful is now running smoothly in its new home. You can see from the photos below how much fun Jose and I had.

newhome2 newhome3
newhome4 newhome5


Better Healthcare Using DataSituation
The Ruth M. Rothstein CORE Center provides all the services that HIV/AIDS and other infectious disease patients need under one roof including medical care, support services, prevention programs, and confidential testing of STDs and HIV. The CORE Center has more than 20 project managers, researchers, and patient counselors who are required to interact with patients based on data regarding visit logs, demographics, medications, and feedback.

Challenge
Part of the CORE Center’s data was housed in a Data Warehouse system with limited access and reporting capabilities. Other data needed to increase patient engagement in care came from survey tools, patient visit log system, and spreadsheets. To understand which patients needed the most urgent interaction, employees would issue a request, sort through a spreadsheet, and use gut instinct rules to select “critical” patients. It took weeks before any patient counselor could follow up with patients that needed the most help.

Solution
The CORE Center invested in 3 powerful components of Mineful: data collection, integration, and dashboard reporting solution. Data is updated continuously using Mineful’s API giving access to near real-time information to counselors, researchers, partners, and grant funding agencies. Researchers are able to dig deeper and use Mineful’s robust analysis engine to easily sort through medical, survey, and behavioral data and establish best practices and rules for patient interaction in order to decrease patient attrition and improve engagement in care.



Average Attendance Jumps 40% in Two Years
As a Chicago analytics startup, we keep track of all our sports teams and their stats. We love stats of all kinds: sports, business, politics, and finance.

Sports fans that truly care about a game know stat leaders, game stats, and all kinds of stat anomalies.

Since we started Mineful, there hasn’t been a stat indicating that any team in town might have a chance to reach past the semifinals. Except for last year, the Chicago Blackhawks were ranked 1st and in many sport betting sites favorites to win, but they didn’t.

Yesterday, it finally happened. The Chicago Blackhawks are chi-town’s champs. Yes, we’ll admit we are not really avid hockey fans, but it sure feels good to root for the home team. Are we a disgrace for not cheering for them in 2005 when the seats were populated by flies? Maybe, but we are not alone. Average attendance has jumped more than 40% in two years. Three years ago, the Blackhawks were ranked next to last in attendance. In 2009 and 2010, they were ranked number 1.


This weekend, everyone is a Blackhawks fan. Who has been a fan since 2005? Who cares, lets celebrate.



The Arizona Immigration Law has been in the news ever since there was news that Governor Jan Brewer would out her signatures to the document and make it a law. The Senate Bill 1070 or the SB 1070 is a law that allows police enforcement personnel to detain and check on individuals in the state with regards to their legal immigrant status. The police have therefore been given the right to ask for proper documentation to be produced during lawful contact if there is reasonable suspicion.

Uproar
The law has given rise to a large amount of uproar among the people and the media is replete with criticism of the law. Those angry and fuming state that the law is only likely to give relentless power in the hands of the police, prejudiced members of which shall misuse the power and harass Hispanics and those of Latino origin. It is being presented as a law that will increase ethnic profiling, giving rise to racial discrimination, community distrust, higher levels of crime and litigations; something that is likely to have national political repercussions too.

Fine for Some
Some, however have been through the details of the law and state that the law has been worded well taking into account the various fears that have been mentioned by some Hispanics. Governor Jan Brewer states that she was aware of the kind of uproar that the law might bring about but she felt that she could not allow illegal immigrants to continue coming into the state since the Federal Government at Washington had been slow in reacting to the needs of the state.

Survey Says
Despite the fact that the issue seems to have created displeasure among many, it is surprising that the poll results show differently.

In the Gallup poll held in April 2010, more than three fourths of Americans had heard about the law. Among these slightly more than half (51%) are in favor of the law that has been passed. Three fourths of the Republicans were in favor as against the 34 percent of Democrats who agreed with the law. Half the Independents were also in favor of the SB 1070. The Gallup poll results are based on a telephonic survey that was conducted among 1013 adult Americans, 18 years and above.

The national survey conducted by Pew Research Center for the People and the Press among 994 adult Americans also had a similar story to tell. At an overall level, almost 60 percent approved of the law and almost three fourths thought that it was justified to ask people to produce documents related to their legal immigrant or resident status. Two thirds were also felt that it was okay to detain someone who could not produce legal documents verifying the status.

A higher proportion of Republicans (82%) were in favor of the law at an overall level as compared to Independents (64%) and Democrats (45%). However, 65 percent of the Democrats also felt that asking people to produce legal documents to verify status was not out of line. The older Americans (65+ years) were more in favor of the law (82%) than the fiery younger ones (18 – 29 years), only 45 percent of whom approved of the law. Here again the younger group also approved of the need to produce legal documents when asked (61%).

Yet another poll conducted by McClatchy-Ipsos among 1016 adult Americans showed that 61 percent and 64 percent of registered voters favored the Arizona Immigration Law.
The latest Rasmussen Report based on telephonic interviews of prospective voters in Arizona showed that 56 percent approved of the manner in which Governor Brewer was conducting her duties. This is an increase from the 40 percent recorded a fortnight earlier.

With all the poll results showing positive numbers in favor of the Arizona Immigration Law, it seems like a case of too much noise being made by some people who are unduly feeling sensitive about the recent law that has been passed. It is merely a case of the dissatisfied making a higher ruckus of things as compared those who are satisfied or in favor.



In conducting a survey, how large a sample do you need to feel confident about the accuracy of the results? Are 100 responses enough, or do you need 1,000? Whatever number you decide on, what response rate can you expect? In other words, how many people do you have to survey to get the number of responses you want? Let’s take a look at these questions.

How Confident Do You Want to Feel?
The size of your sample depends on how confident you want to be about your results. They key question is: “Are these results truly representative of a larger population?” If you would like to have a confidence level of 95%, you would need to have a margin of error of 5% or less. This means that there would only be a 5% chance of your results differing from the results you would get if you surveyed the entire population. (The margin of error is also called the confidence interval.)

There is a simple equation you can use to determine margin of error:
The margin of error equals 1 divided by the square root of the sample size. Suppose you had a sample size of 400. Your equation would be 1 divided by 20 (the square root of 400), which equals 5%. So to have a confidence level of 95%, you would need to have a sample size of 400.

Obviously, the higher the confidence level you want, the larger your sample needs to be. A confidence level of 95% is widely considered to be acceptable.

Of course this assumes that you have a truly random sample. As experienced pollsters know, it is easy for sampling bias to sneak into a survey without anyone noticing.

How Many to Ask?
So once you have decided on a confidence level, you can easily determine how large a sample you need. But how many people do you have to survey to get the desired number of responses? That is a trickier question.

Response rates vary greatly from one population to another. One of Mineful’s clients, a medical association, received a 25% response rate to a member satisfaction survey. Another client had a similar response to a survey conducted in connection with a conference. But such high rates are the exception rather than the rule. For a survey conducted through cold email invitations, a response rate of 1% is considered good.

Response rates are affected by a number of variables:

  • How strongly people feel about the survey topic.
  • How much loyalty people feel toward the organization conducting the survey.
  • How easy it is too complete the survey.
  • How confident people feel that someone will pay attention to their responses.

Of course these factors can be difficult to measure, and experience is the best way to predict response rate.

Fortunately, there is an online sample size calculator that can make this whole process much easier (http://www.raosoft.com/samplesize.html). These are the variables you will need to plug in to use this calculator:

  • Acceptable margin of error
  • Acceptable confidence level
  • Population size (for populations smaller than 20,000
  • Response distribution

The calculator explains what each of these variables means.



« Previous posts | Back to top