Saturday, July 25, 2009
E-Tailing at a Children's Toy Company
A regional toy company is considering the Internet as a selling channel and commissioned a background research report to decide if this retailing channel warrants further investigation. I used the Internet to do background research on toy retailing on the Internet. McDaniel and Gates (2008, p 86) recommend listing the distinctive words that might identify articles as a first step for an Internet search. I chose the following words: buy educational toys online. For my first search I did not try to enclose phrases or clichés in quotes because I wanted to see a wider range of results to get other ideas for searching. This starting gambit hit a rich vein of useful material. In addition to educational toy company Web sites, I found this blog that reviews educational toy companies: Suite 101.
The companies reviewed in the blog such as The Discovery Channel provided useful information. Moreover, I followed the McDaniel and Gates guidance (p. 87) to “vary your approach with what you learn,” and picked up some words and other ideas in the review of these companies. E-tailing was a new word and eToys was a high profile failure that merited investigation. I decided to also improve the quality of my source material (see McDaniel and Gates, 2008, p. 87) and started searching with my new terms and information in EBSCOHOST. The following is my report to the toy retailer, call them KidzBiz.
The Market Reality for Online Educational Toy Sales
There is no mistake that revenue growth for online educational toy sales is front-running its brick and mortar relation by a significant margin. Online sales now form a significant share of sales for educational toy companies with the vision to participate in this channel. On example is Fat Brain (see Davis, 2008, p. 1), which has most of its sales online. It also takes advantage of the infrastructure services provided by Amazon to reduce the risk and cost of selling online by sharing Amazon server capacity, technical services and expertise.
Carson (2009, p 1) reports that Fat Brain is now on the Internet Retailer’s Hot 100 List. Furthermore, Inc. Magazine (2008, p 1) reports the astonishing growth in Fat Brain sales of 428% from 2002-2008. It also reports that the company is ranked number 40 in the Top 100 Consumer Product Companies. Not bad for a small family owned business that started in a garage selling educational toys through local stores.
What about established players? According to Internet Retailer (2006, p. 1), Toys-R-Us online sales jumped 20% with increasing strength in both number of orders and average order size. Although they do not specialize in only educational toys, they do sell them and experienced a 33% growth in sales of toys for toddlers. Hughes (p 32) reports that Disney has experienced phenomenal success in selling educational toys by “redefining babies solely as learners whose potential to learn can be released by consuming these products.” The combined message is a powerful indicator that online sales of educational toys for toddlers yield rich returns.
Are there other examples? Consider Leapfrog, an educational toy vendor who sells in brick and mortar retail stores, through its own online store and through online channels like Yahoo Shopping. In its quarterly filing with the SEC (see SEC, 2009, p. 15), Leapfrog reports that it plans to increase its online presence. This is in spite of, or perhaps because of the beating retailers are getting from the economic downturn (see SEC, 2009, p. 14).
The Discovery Channel (2008, p. 12) reported in its 10-Q news conference review that it closed all brick and mortar stores on May 17, 2007 and now will sell its educational products solely through catalog and their online store. They further inform us (p. 3) that they are aggressively investing in their online properties. Web traffic to their Web sites almost tripled from March 2007 to March 2008 from 13 million unique visitors per month to 33 million. Please note that the report does not break out visits to their online store from visits to their informational sites.
What about the high profile failures
The most high profile failure for an online toy company was eToys. Sliwa (2001, p 1) reports that eToys faced strong Internet competition from Amazon. As noted above, Amazon is now willing to partner with toy sellers. Moreover, Gomolski (2001, p. 72) goes into more depth about why eToys withered before Amazon. She reports that the eToys business model was based on competitive pricing but they had neglected to build awareness about their pricing. In addition, they failed to use the Internet to build customer relations of any kind. This resulted not only in inadequate brand awareness but also a failure to connect with children.
Moreover, eToys did not have the critical mass to reach profitability quickly and at the same time did not have cash flow from conventional business operations to sustain itself. The result is that eToys was not able to initially compete online with Amazon and could not sustain itself until it could.
How Can We Trust the Data
Can we trust the information used in this report? The sources are well respected. Inc. magazine, InternetRetailer, ComputerWorld, and InfoWorld are news magazines of note. They follow standard journalistic practice, which is designed to ensure reliable reporting. The Contemporary Issues in Early Childhood is a professional journal. Articles are peer reviewed to again ensure information reliability.
McDaniels and Gates is an academic textbook. It has been reviewed by the editorial board of John Wiley, one of the most trusted publishers. Finally, audited 10-Q statements are filed with the federal government. Heavy penalties under the administrative law are assessed for publishing false information on these statements.
Recommendation
KidzBiz has a lucrative opportunity to expand its sales and establish a presence in a new and growing sales and distribution system, the Internet. That system is placing relentless pressure on traditional brick and mortar operations. Today, KidzBiz is solely dependent on that old tired soldier. The next step should be to review Kidz Biz internal customer, product and sales data.
Johnson (2009, p. 3) says that, “secondary data can be greatly enhanced when merged with internally-generated data.” Kuchinskas (2003, p. 2) reports that in 2000 Dell experienced declining growth in the educational market because of tightening education budgets. They responded with a database-marketing program to the education sector. Database marketing on the Internet should enhance KidzBiz sales.
References
Carson, Mark (January 8, 2009). Fat Brain Toys Named To Internet Retailer’s 2009 Hot 100 List. Fat Brain Press Release. Retrieved on May 26, 2009 from http://www.fatbraintoys.com/about_us/press.cfm?pr_id=62
Davis, Don (September 2008). Advantage Amazon. Internet Retailer. Retrieved on May 26, 2009 from http://www.internetretailer.com/article.asp?id=27583
Discovery Holding (May 8, 2008). Discovery Holding Company First Quarter Earnings Release. Retrieved on May 27, 2009 from http://www.prnewswire.com/cgi-bin/stories.pl?ACCT=104&STORY=/www/story/05-08-2008/0004809260&EDATE
Gomolski, Barb (02/05/2001). Going global: Some lessons from eToys and Yahoo that might help you. InfoWorld. Retrieved on May 29, 2009 from EBSCOHOST.
Hughes, P (March 2005). Baby, It's You: international capital discovers the under threes. Comtemporary issues in early childhood. Contemporary Issues in Early Childhood. Retrieved on May 26, 2009 from
EBSCOHOST and http://www.wwwords.co.uk/pdf/viewpdf.asp?j=ciec&vol=6&issue=1&year=2005&article=4_Hughes_CIEC_6_1_web&id=74.6.25.186.
Inc. Magazine (2008). Company Profile: Fat Brain Toys. Retrieved on May 27, 2009 from http://www.inc.com/inc5000/2008/company-profile.html?id=200808330
Internet Retailer (June 21, 2006). Toysrus.com swings into the black in first quarter. Retrieved on May 26, 2009 from http://www.internetretailer.com/dailyNews.asp?id=18996
Johnson, E. (2009). Using Secondary Data & Databases. Retrieved on May 30, 2009 from WVU www.imc.wvu.edu.
Kuchinskas, Susan (Sep 2003). Data-based Dell. Adweek Magazines' Technology Marketing. Rertieved on June 5, 2009 from WVU IMC 611 week 3 readings.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
SEC (March 31, 2009). Form 10-Q for Leapfrog Enterprises. Retrieved on May 27, 2009 from http://yahoo.brand.edgar-online.com/displayfilinginfo.aspx?FilingID=6580220-812-158026&type=sect&TabIndex=2&companyid=108193&ppu=%252fdefault.aspx%253fcik%253d1138951
Sliwa, Carol (1/8/2001). Facing Tough Rivals, eToys Nears Oblivion. Computerworld. Retrieved on May 29, 2009 from EBSCOHOST.
Wednesday, July 22, 2009
Using ColdFusion to add XML data to a SQL database
As an example, consider the following. An xml file has a field named ICN, among others. We will read the file and update a SQL database. First, we get a file handle.
< cffile variable="gmrXML" file=" “wits.xml" action="read">
Then a ColdFusion XML Document Object datatype, using our file handle and the ColdFusion xmlParse() function:
< cfset myxml=" xmlParse(gmrXML)">
When we use cfoutput to see what we have, all is well:
<cfoutput>
<cfloop to="#arrayLen(myXML.incidentList.Incident)#" from="1" index="i">
#myXml.incidentlist.incident[i].ICN#
…
The output on our Web page shows the ICN value we expect:
200458431
Good enough. We now add it to our database with the following code:
<cfquery datasource="SMC" name="loadSMC">
Insert into aIncident(ICN)
values(#myXml.incidentlist.incident[i].ICN#
< /cfquery>
and when we look, the value in the ICN column in the database is not 200458431 but instead is
<?xml version="1.0"?>
This disappointing result is because in the cfquery above, we were treating the ICN reference as a value. cfouput helped in this deception because we could treat it as a value with this sophisticated function. A look at cfdump shows what happened.
Here is a dump of the reference:
It is not a value but a structure. To get to the value we want, we need to add .xmlText to the end of the reference we used so that the query now looks like:
<cfquery datasource="SMC" name="loadSMC">
Insert into aIncident(ICN)
values(#myXml.incidentlist.incident[i].ICN.xmlText#
</cfquery>
This works as expected.
Monday, July 20, 2009
ColdFusion Program to Read XML File
With XML, a structural model of the data in a file can be encoded along with the data. The structure can be as simple as name and phone or much more complex, with fields like name and phone embedded in other fields like employee. A style sheet, an XSL, can be used to transform the tagged data in an XML file into a Web page. Likewise, a schema file, an XSD, can communicate database information in database operations.
ColdFusion provides a subset of functions that enable a programmer to operate on XML files. A typical operation would be to read an XML file, work on it and write it, perhaps to a database. A datatype in ColdFusion has been created for XML and is known as the XML document object. By doing this, Adobe extends the reach of its already existing ColdFusion structure functions to encompass XML data as well.
Structures consist of objects, properties, and objects embedded in other objects. Name and phone embedded in employee is an example. The dot operator is used to delimit what object in a structure you want to access. The general syntax is <object>.<object>, <object>.<property>. These can be combined to get for instance <object>.<child object>.<property>, so to access my phone number, we would use the dot operator: <employee>.<phone>
The following ColdFusion code uses some of the XML functions to read an XML file. It starts by validating the XML data file to ensure that it is consistent with its schema.
<cfset myResults= XMLValidate ("wits.xml", "wits.xsd")>
<cfoutput> Is Valid? #myResults.status#!
</cfoutput>
Next, a file handle is obtained.
<cffile variable="gmrXML" file=" “wits.xml" action="read">
Then we get a ColdFusion XML Document Object datatype using our file handle and the xmlParse() function:
<cfset myxml=" xmlParse(gmrXML)">
Now we are ready to loop through the data, and display different field values. This test file consists of a series of incidents embedded in an incident list. In addition to its own properties, an incident will have an embedded list, or object, of one or more incident types. In addition to this, it will have, an embedded facilities list, of zero or more facilities involved in the incident. These embedded lists correspond to embedded objects or to child tables in a database.
<cfoutput>
<cfloop index="i" from="1" to="#arrayLen(myXML.incidentList.Incident)#">
Incident<br>
#myXml.incidentlist.incident[i].ICN#<br>
#myXml.incidentlist.incident[i].Subject#<br>
#myXml.incidentlist.incident[i].Summary#<br>
#myXml.incidentlist.incident[i].IncidentDate#<br>
Event Type<br>
<!---Here is an inner loop to get all Event Types for the Incident--->
<cfloop index="j" from="1" to=
"#arrayLen(myXML.incidentList.Incident[i]. EventTypeList.EventType)#">
#myXml.incidentlist.incident[i].EventTypeList.EventType[j]#<br>
</cfloop>
<br>Facilities<br>
<cfif StructKeyExists(myXml.incidentlist.incident[i].FacilityList,"Facility")>
<!---Here is an inner loop to get all Facilities for the Incident--->
<cfloop index="j" from="1" to
"#arrayLen(myXML.incidentList.Incident[i]. FacilityList.Facility)#">
#myXml.incidentlist.incident[i].FacilityList.Facility[j].FacilityType#<br>
#myXml.incidentlist.incident[i].FacilityList.Facility[j].Indicator#<br>
</cfloop>
<cfelse>
No facility for this one
</cfif>
</cfloop>
</cfoutput>
The ColdFusion arraylen is used in the “To” parameter of our loop to return the number of Incidents in this incidentlist. There is one incidentlist per file. Our dot operator starts with the file handle, then refers to the incidentlist we know to be in file (because we validated), and the particular incident is referenced with our index variable [i]:
myXml.incidentlist.incident[i].ICN
The incident number property, ICN, is accessed from the record just read from the file. For EventType, it is very similar. Since we know we will have at least one Event Type in every incident (one or more), the following code is sufficient:
myXml.incidentlist.incident[i].EventTypeList.EventType[j]
On the other hand, we may not have a facility (zero or more) and will get a null pointer exception if we try to dereference a facility in an empty facilitylist. We need to use the ColdFusion StructKeyExists function to test if a facility object is embedded in the facilitylist for the current record.
<cfif StructKeyExists(myXml.incidentlist.incident[i].FacilityList,"Facility")>
If so, then we loop through the facilities involved in the incident, else we indicate no facility.
Sunday, July 19, 2009
The Marketing Research Report: Now and in 1964
Furthermore, research is the basis for exploring new opportunities in the environment. Research helps segment the environment, and match segments with the characteristics or the product or action. Reporting research findings is an essential part of making it actionable. Johnson (2009, p. 1) advises us “the research report, the final step in the research process, requires thoughtful preparation and presentation.”
What are the key components to a research report?
With one interesting exception, the key components of a research report have not changed much since Boyd and Westphal published their Marketing Research text in 1964. Some forty years on, McDaniel and Gates (2008, p. 468-9) still have the following elements:
- Title Page: The name of project.
- Table of contents: A list of major sections.
- Executive summary: Key findings and recommendations.
- Background: Strengths, weaknesses, opportunities and threats.
- Methodology: although Boyd, et al organize this with three major subsections, the research design, collections methods, and sampling.
- Findings: A summary of results for every question in the survey.
- Appendices
Boyd and Westphal have these same sections but also include a major section on Limitations. They note (1964, p. 575) that “A good report sells the results of the study, but it should not oversell. Every project has limitations.”

McDaniel and Gates (p. 470) tell us that the report interprets findings to arrive at conclusions using induction and then deduces recommendations from the conclusions. Induction starts (pp 470-2) by examining one-dimensional tables and then it cross tabulates to see how characteristics both independently and in tandem affect the dependent variable. A recommendation explains how a differential advantage can be obtained (p. 472). Finally, contrary to popular belief, the report does use pictures and graphs.
How would you incorporate those components into an effective research presentation?
Rubin and Babbie (2005, p. 661) admonish us to know our audience. They emphasize that the report writer must distinguish between professional colleagues and business readers. It is critical to not make assumptions about the existing knowledge of business readers. They go on to say that with business readers it is best to keep terminology simple and clear.
Furthermore, summaries and visuals appeal to business readers as does expressing the implications for their area of operations. Neal (1998, p. 23) recommends that we take time to explain complex analysis and data to our business audience. Doing so will build confidence in the research.
Boyd and Westfall (1964, p. 570) say that to make the report effective, start at the beginning. The report writer must keep the study objectives in mind when writing the report. The writer should be selective of what is included in the report, making sure it is related to the objectives. McDaniel and Gates (2008, p. 468) likewise suggest keeping the report strictly oriented to the objectives. They say, “The genesis of the report and the researcher’s thinking are the objectives provided by the client….”
McDaniel and Gates further tell us to be storytellers (p. 468). With mountains of information, the challenge to the report writer is how to package it into a coherent message. Story telling helps. They finally inform us that Microsoft PowerPoint is not only used for the oral presentation but often for the written report (p. 470). This has been my experience at work. Vendors, including prestigious firms like Booz Allen, submit final reports as PowerPoints. It is fitting with our over burdening workloads.
How does the written report differ from the presentation in terms of its function and format?
Boyd and Westphal (1964, p. 579) say the oral presentation demands “greater use of dramatics”, in other words more use of visual aids. They also suggest that transparencies, what we now call PowerPoint slides, make greater use of “headline style” writing. McDaniel and Gates (2008, p. 473) note that the presentation is an assembly of stakeholders who need to get “reacquainted with the research objectives and methodology.” They say that a copy of the full report along with the visual presentation should be handed to the participants.
The presentation must succinctly express the following (p. 475) as part of the persuasion process. First, interpret what the data mean and the impact on the organization. Next, we have learned something and does this new knowledge reveal new opportunities? Finally, what could be done better? This last point goes to the Boyd and Westphal inclusion of a Limitations section in the report itself.
McDaniel and Gates (2008, p. 476) offer the possibility of using Web technology to publish the presentation. We have done that at work. It is easy to publish a PowerPoint as a self-contained Web page or as an attachment to a blog. The blog allows for community feedback and commentary. We deal with an organization of lawyers, so we still need the face-to-face and the cross-examination in person, even if we post on a Web page
References
Boyd, Harper and Ralph Westfall (1964). Marketing Research Text and Cases. Richard Irwin.
Johnson, E. (2009). Communicating Research Results & Managing Marketing Research. Retrieved on July 11, 2009 from www.imc.wvu.edu.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
Neal, William D (Spring 1998). The Marketing Research Methodologist. Marketing Research.
Rubin, Allen and Earl Babbie (2005). Research Methods for Social Work. Thompson/Brooks/Cole.
Saturday, July 4, 2009
The HERI College Senior Survey
McDaniel and Gates (2008, p. 298) note that dichotomous questions are often subject to measurement error because they offer only black and white choices when many times shades of gray are needed. This is not the case with the HERI questionnaire. Question 6 is the only set of dichotomous choices and they are truly yes/no.
The first three pages of the HERI instrument are mostly scaled response questions. They not only allow seniors to express an opinion about a subject but also calibrate the intensity of that feeling (p. 299). A potential problem with scaled response, remembering category options, is avoided in this questionnaire by its careful design.
The fourth page introduces multiple choice questions. McDaniel and Gates (p. 299) warn that the choices may not cover all alternatives but this may be mitigated by offering an ‘other’ choice. HERI does not do this. They should to improve the questionnaire over time. In addition, Question 27 may have positioning bias. The two positive answers about colleges are positioned where they are most easily seen.
The next question reveals another common problem in the questionnaire. McDaniel and Gates (p. 301) say that words should have the same meaning to all respondents and that, additionally, words used in questions should have a precise meaning. In Question 28 both flaming liberals and draconian conservatives could indicate themselves as moderate. This same issue occurs in Question 13 about drinking frequently or occasionally. What I consider occasional, someone else may consider frequent. Finally, in Question 5, what does the word regularly mean in “regularly communicated with professors?” Specific guidance about how many for frequent, occasional and regular is needed.
Sensitive or embarrassing questions are handled by HERI in a robotic manner. In Question 13 students are directly asked about mental depression, the need for professional counseling, and alcoholic drinking habits.
Question 19 asks about marijuana, same sex marriage, denial of services to undocumented immigrants, and affirmative action. Many may be sensitive to these issues and McDaniel and Gates (p. 327) recommend using one of two techniques in such cases: 1.) Counterbiasing; or 2.) 3rd person voice. HERI does not. This is not an anonymous survey.
McDaniel and Gates (p. 301) also admonish us not to ask questions the respondents could not answer correctly. In Question 14, HERI asks students to guess what quintile they are in for various aptitudes or domain knowledges. Ambrose and Anstey (2007, p. 28) state that an important topic category in a survey is measuring the knowledge of the population. By this they mean discovering the levels of understanding. However, in the HERI survey, the students would have no frame of reference for knowing the answers to the various parts of Question 14 and so McDaniel and Gates apply with full force.
McDaniel and Gates (p. 302) also recommend that time periods be kept recent. They highlight a question as poorly worded because its time period is a year. In Question 9 of the HERI instrument, the time period of a year is used in twenty parts to that question. As McDaniel and Gates ask, unless the students kept an accurate diary of each of the twenty activities, how would they know? A better approach is to ask how much the respondent has done in the past two weeks and then if that is more or less than average (p. 302).
All in all, I found the HERI questionnaire to be well done but a few questions could be fine tuned to get more complete and more accurate results.
References
Ambrose, David and John Anstey (March 2007). Better Suvey Design is Stick for an Answer. ABA Bank Marketing. Retieved from http://www.imc.wvu.edu/ on July 4, 2009.
Christian, L and D DIllman (Spring 2004). THE INFLUENCE OF GRAPHICAL AND SYMBOLIC LANGUAGE MANIPULATIONS ON RESPONSES. Retieved from http://www.imc.wvu.edu/ on July 4, 2009.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
Saturday, June 27, 2009
Assessing the Impact of a Sexual Harassment Lawsuit at a Metropolitan Restuarant
Exploratory Research
Harris and Smith (2002, p. C1) define a crisis as any newsworthy event that has happened or will happen in the near future that has the potential for adverse publicity. The sexual harassment event associated with the Restaurant X satisfies their definition of a crisis. In his section on public relations, Duncan (2005, p. 563) says that this type of incident should be treated as a crisis because of the extreme impact it poses to Restaurant X’s “reputation and brand relationships.”
Lichtenstein, et al (2004, p. 29) reinforce Duncan’s premise by noting that corporate social responsibility strengthens brand equity and say its impact goes beyond sales. This means that a consumer perception of Restaurant X as an environment that condones sexual harassment will have a negative effect not only on sales. A reduction of brand equity will also result and with it there will be a consequent loss of owner equity. It is vitally important for Restaurant X to verify through marketing research the attitudes its publics harbor about the incident.
What type of survey research should be initiated
Management should already understand that negative publicity in general affects customer retention. They will want to know if their specific sexual harassment publicity is reducing their customer retention and sales. The goal of this marketing research should be to “shed light on [this specific] association or relationship,” which is how McDaniel and Gates (2008, p. 49) characterize a descriptive research design.
The basic method of research should be survey. McDaniel and Gates (p. 50) inform us that experiments are typically used only for causal research designs. Observation would be an inefficient approach to determine opinions and attitudes on the restaurant’s sexual harassment case because there is no direct interaction with the customer. Surveys are “an orderly and structured approach” to understand “opinions and attitudes.”
This research should be quantitative. Qualitative research is “not necessarily representative of the population of interest,” (see McDaniel and Gates, 2008, p. 110). Surveys are a good tool to find out what factors influence consumers (p. 142); in this case, did the publicity influence customer retention and sales. Of the various types of surveys, I recommend Computer-Assisted Telephone Interviews (CATI).
Why CATI?
In this situation there is a tradeoff between turnaround speed, measurement error, and cost. CATI is the optimal strategy, with Internet Panels a close second. Johnson (2009, p. 2) notes that non-response bias is a significant source of measurement error. He also informs us that both telephone and mail surveys are subject to increasing non-response rates.
Thee (2007, p. 1) reports another problem with telephone surveys for the restaurant: because of the substantial use of cell phones, telephone sampling frames are no longer geographically based. On the other hand, the restaurant public is geographically bound. Thee goes on to say that households with only a cell phone numbered 16% in 2007 (p. 1) and this could grow to 25% by the end of 2008.
CATI can help here. McDaniel and Gates (p. 152) inform us that the computer can customize each questionnaire according to how the respondent answers previous questions. We could ask if they are currently in the metro area. If not and they have no intention of visiting the metro area, then we could close the interview.
McDaniel and Gates (p. 365) also report on a method for determining the extent of nonresponse bias. They cite two studies that suggest nonrespondents are not a clearly distinct subgroup from the general population. To find out for certain in this case, only a very small follow-up survey using nonrespondents as the sampling frame need be conducted.
Johnson (2009, p. 1) informs us that restaurants make frequent use of self-administered surveys. The reason they are not the best solution in this case, is that customers already lost to Restaurant X because of the lawsuit won’t be patronizing the restaurant anymore. In addition, a self-administered survey approach will not have the same quick turnaround as online techniques (see Johnson, 2009, p. 2).
Interviews such as door-to-door and executive would have little traction with the immediate needs of the restaurant. Mall intercept interviews would be too expensive, and in addition, McDaniel and Gates (2008, p. 172) say to avoid such an approach when the incidence rate is low because it will be too expensive. The number of patrons of Restaurant X is low compared to the total population in the metropolitan area.
Internet Panels have been used in other crisis management situations. Duncan (2005, p. 564) reports that McDonald’s setup a panel to protect its reputation from the fallout over a fraud perpetrated in its promotional games. McDaniel and Gates (p. 156) also discuss the advantages of panels. Johnson (p. 2) informs us that “companies like Survey Sampling” provide consulting expertise to “construct a sample to match the target population” and then rent a panel from their large database. Still, as Johnson (p. 2) observes “The internet population is not representative of the population as a whole,” and “Internet users are concerned with privacy and, thus, may refuse the survey.”
What topics should the survey cover?
First up is demographics such as gender, age, and race. The next topic would investigate level of awareness. It would be used to discover the news sources that informed the respondent about the lawsuit, as well as their knowledge about the particulars of the case, and knowledge about Restaurant X policies and training to prevent sexual harassment.
The third topic would be to determine their opinion about how Restaurant X handled the sexual harassment accusation. Duncan (2005, p. 563) says that for such a crisis it is a good idea to survey opinion about the company’s level of honesty, helpfulness, speed of response, clarity, and consistency of message.
The fourth topic would be to find out from patrons about their experience with the restaurant. Have they ever witnessed sexually harassing behavior at the restaurant? Has the staff been professional and courteous? The fifth topic would be to establish the factors that influence Restaurant X patron decisions to frequent a restaurant. An important aspect here is how did the sexual harassment lawsuit affect their opinion about Restaurant X and their willingness to be a customer? The sixth topic would be open-ended questions to find out their opinions about what further steps Restaurant X should take.
The survey population
The population is the existing, lost and potential customers of Restaurant X in the metropolitan area. With no further information from the restuarant advertising or news stories, I will assume that the entire metropolitan area should be viewed as potential customers. To focus only on existing customers would not discover the opinions of lost customers. Additionally it would exclude the only source of growth, new customers.
The sampling units and sampling frame
The sampling unit would be households because people tend to go to restaurants as a family or as a couple and not alone. The sampling frame should not be the telephone directory for the metropolitan area. McDaniel and Gates (pp 333-4) list the problems with the telephone directory including problems with unlisted numbers and their prevalence among non-whites and younger people. Instead random-digit dialing from phone exchanges in the metropolitan area will form our sampling frame.
Simple random sampling will be used to create our random-digit dialing frame. Johnson (2009, p. 3) notes that probability sampling, like simple random, gives the researcher a representative sample of the population. Specifically for simple random sampling, a computer can randomly select the sample and this is consistent with the CATI approach I recommend.
The sample size
The main difference between qualitative and quantitative research is sample size (see McDaniel and Gates, 2009, p.108). Managers are more “comfortable with marketing research based on large samples and high levels of statistical significance” (p. 109). Another factor that will affect our sample size is the required confidence level. Wilson and Ogden (2004, p. 56) say that
"Survey research requires at least a 95% confidence level and a margin of error of less than 5% to be actionable.”Do we need to know the exact population size for the metropolitan area? Using a sample size calculator it quickly becomes apparent that after a certain point increasing population size by an order of magnitude does not increase sample size requirements by more than a single person or two. Creative Research Systems (2009, p. 1) provides a Sample Size Online Calculator that I used to calculate the following sample sizes:
Given the gravity of sexual harassment allegations and the damage to brand equity, owner equity and ongoing sales, I recommend a sample size of 384.
References
Creative Research Systems (2009). Sample Size Calculator. Retrieved on June 17, 2009 from http://www.surveysystem.com/sscalc.htm
Duncan, Tom (2005). Principles of Advertising and IMC. New York: McGraw-Hill Irwin.
Harris and Smith Public Affairs (July 3, 2002). Public Relations Handbook. Retrieved on June 10, 2009 from http://www.awphd.org/PRhandbook/PRhandbook.pdf.
Johnson, E. (2009). Quantitative Research: Surveys & Sampling. Retrieved on June 11, 2009 from http://www.imc.wvu.edu/.
Lichtenstein, Donald R., Drumwright, Minette E. & Bridgette M. Braig (October 2004). The Effect of Corporate Social Responsibility on Customer Donations to Corporate-Supported Nonprofits. Retrieved on June 9, 2009 from http://www.imc.wvu.edu/.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
Thee, Megan (December 7, 2007). Cellphones Challenge Poll Sampling. NY Times. Retrieved on June 10, 2009 from www.imc.wvu.edu.
Wilson, Laurie and Ogden, Joseph (2004). Strategic Communications Planning, 4th Ed. Kendall/Hunt Publishing Company.
Saturday, June 20, 2009
Measuring Attention and Packaging Design
Independent and dependent variables, treatment groups, the design and control for extraneous factors.
Percy (2008, p. 256) notes that mothers are key decision makers for family health and eating habits. Teenager awareness and acceptance, and mother acceptance will jointly determine the purchase decision. Our task is to test the effectiveness of two alternative package designs in getting teenager attention. The dependent variable is teenager attention. The independent variable is package design.
Wilson and Ogden (2004, p. 36) define an influential as an opinion leader, and observe “that people follow the advice of someone they trust and has knowledge on the subject.” We want teenagers who eat cereal but who are also influentials. They will be our subjects. This is a new product and the most important group to reach in this introductory phase is the influentials.
McDaniel and Gates (p. 114) also recommend “Influentials.” They say these people are two to five years ahead of the curve and have significant influence among their peers. They say, “Recruiting influentials is particularly useful when a marketer is trying to determine to launch a new product….” Brooks (2006, pp 31-2) agrees and further says, “A growing body of consumer research suggests the 60-second conversation is dethroning the 30-second ad.”
We will have two experimental groups and one control group. Experimental group one will view package design one. Physiological Measurements will be made to monitor their response to the package design. These measurements can show interest and arousal levels (see McDaniel and Gates, 2008, p 192). The treatment is to show them the design. Likewise for experimental group two but for package design two instead. The control group won't be shown either proposed design.
There are problems with measuring attention because it may change during the testing process and also the subjects know they are being tested for it (see Boyd and Westfall, 1964, pp 105, 111). A control group can help compensate for this. The control group will share the exact same experience as the two treatment groups, except they will not be exposed to a package design.
They will, however, be read the same statement about the new product and package designs and that they will be measured. McDaniel and Gates (p. 219) note that the control group is not subject to treatment, in this case exposure to one of the package designs. Their participation can be used to dampen extraneous factors, like the anticipation of knowing they will be measured.
Experimental design and conduct of the experiment.
The steps are
- Three groups of influentials will be randomly selected
- Physiological Measurements will be made of each group
- Each group will be read each the same statement about the new product
- For the control group, the post measurement will be taken
- For the two treatment groups, show each their respective packaging designs only and take measurements
The research design is a True Experiment Design (see McDaniel and Gates, 2008, p 219), the Before-and-After with control design (p. 219). The test subjects should be assigned to each group randomly to control extraneous effects. The experiment should be held in a lab for high internal validity (p. 211).
References
Boyd, Harper and Ralph Westfall (1964). Marketing Research Text and Cases. Richard Irwin.
Brooks, Steve (November 2006). How to Build Buzz. Restaurant Business. Retrieved on June 19, 2009 from EBSCOHOST.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
Percy, Larry (2008). Strategic Integrated Marketing Communications. Butterworth-Heinemann.
Wilson, Laurie and Ogden, Joseph (2004). Strategic Communications Planning, 4th Ed. Kendall/Hunt Publishing Company.
Sunday, June 14, 2009
A New Logo for Wendy's?
Wendy’s is an international fast food chain that ranks third in hamburger joints behind McDonald’s and Burger King (Hoover’s, 2009, p 1). It has over 6,500 locations in the United States and more than 20 countries worldwide. Management is considering changing the Wendy logo and brand character to keep it current. There is a concern that changing the well-known logo could have adverse repercussions. To study consumer opinion a series of personal interviews will be conducted.

According to Keller (pp 156-159), logos and brand characters are an easily recognizable way to identify a company and product. They are central to advertising campaigns and packaging. He says they can “create perceptions of the brand as fund and interesting.” However, (p. 158) he also notes they must be changed from time to time. Betty Crocker spent over $1M to update its tired brand character (pp 164-5). This $1M price tag is also the case for updating a logo (p. 157).
Sampling Method
The background information shows the importance of correctly updating a brand character, and that an appropriate budget must be allocated to this important task. Having an accurate cross-section that is representative of Wendy’s customers outweighs cost. We do not want to draw incorrect conclusions by using a sampling method known to be inaccurate.
The life-cycle of a sampling plan is to define the population, choose a data-collection method, identify a sampling frame, select a sampling method, determine sample size, develop operational procedures, execute the plan (see McDaniel and Gates, 2008, p330). This specific assignment is to select and justify a sampling method. Our first decision is between probability samples and non-probability samples. McDaniel and Gates (p. 334) note that probability samples are better for obtaining a representative cross section of our population. Given the importance of the logo or brand character, we need to have the best representation we can get. The trade-off is cost but we should have a budget commensurate with the task.
Next, we must decide among the several probability sampling methods. I like the stratified sample because it has a smaller sampling error than the other methods (see McDaniel and Gates, 2008, p. 341). It is more efficient because it eliminates one of more sources of variation through the stratification process. The researcher is making the “sample be more representative by making sure that important dimensions of the population are represented in their true population proportion.”
Wendy's should stratify on country. According to Hawkins, et al (2007, p 40) cultural factors that like language, demographics, values and nonverbal communications can impact marketing communications. A cultural faux pas with a brand character or logo can make Wendy’s appear like confused Martians. Using country to stratify, I can also avoid the main drawback on stratified sampling – we usually don’t know in advance the proportions of the strata (see McDaniel and Gates, 2008, p. 342). However, Wendy's consolidated financial statements do have that information.
Finally, should the method be proportional allocation or disproportional allocation? In this case, the method should be proportional because it is easy to calculate the relative number of locations in each country to the total number of locations (p 342). Within each country (strata), do simple random sampling with a sample size proportional to the relative percentage for the country.
To collect the data, the store intercept survey type can be used (see McDaniel and Gates, 2008, p. 150). As a style of mall-intercept survey, it would share many of the same advantages and disadvantages. The interviewer can show proposed Wendy’s logos, explain and probe (p. 171). On the other hand, the respondent may be distracted, or in a hurry. Further, there is a greater opportunity for interviewers to select the people to interview in a non-probabilistic manner. Strict operational procedures should be defined.
References
Hawkins, Del, David Mothersbaugh and Roger Best (2007). Consumer Behavior. McGraw-Hill/Irwin.
Hoovers (2009). Hoover's Profile: Wendy's International, Inc. Retrieved on June 7, 2009 from http://www.answers.com/topic/wendy-s-international
Johnson, E. (2009). Quantitative Research: Surveys & Sampling. Retrieved on June 7, 2009 from http://www.imc.wvu.edu/.
Keller, K (2008), Strategic Brand Management. Pearson/Prentice-Hall.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
Sunday, June 7, 2009
Ruger at a Crossroads
Ruger is the apex predator in the revolver market in the United States but has little presence overseas. Moreover, its domestic market share is under fire from semi-automatic handguns. It has an opportunity to expand its sales internationally through revolver sales to Europe. In Europe, countries such as Finland, Norway, Sweden, Switzerland, Serbia and the Czech Republic have gun control laws similar to the United States (see Dodds, 2007, pp 1-2 or MacInnis, 2007, pp 1-2 or Wikipedia, 2009, pp 1-5). As an example, the U.S. has 90 guns per 100 citizens and Finland has 56. In Germany, a valid firearms license is required for each firearm that is owned. Nevertheless, even Germany represents a sales opportunity.
In its 10-K Report, Sturm Ruger & Co. states (see SEC, 2008, p 4) 94% of its firearms sales are to the domestic market. Their units shipped reported in the 10-K (p. 23) are
I have used the Google Trends service (www.google.com/trends) to analyze search engine requests for the word “Ruger.” What I found was fascinating. There is a significant overseas search interest in Ruger. I contend that search interest will translate into purchase intent. This is a largely untapped market for Ruger. I then compared Ruger with Glock. Here are the results.
In-depth Interviews
Glock, Beretta and Walther are powerful international competitors to Sturm Ruger & Co. They do not have an appealing lineup of revolvers but instead have focused their attention on the semi-automatic pistol. I need to use in-depth interviews to find out the attitudes, beliefs and feelings of European consumers about revolvers versus semi-automatic handguns. Additionally, I need to define the parameters of Ruger’s relationship with the various European gun retailers, who may be competitors with each other.
Given the prevailing preference for European semi-automatics in Europe, I believe an in-depth interview is more appropriate than a focus group to find out attitudes towards revolvers. McDaniel and Gates (p. 127) say that such interviews allow the researcher to devote more time, to more deeply delve into an area, and “ to reveal feelings and motivations that underlie statements.” We need to know if handgun consumers in Europe would consider revolvers and why. The “if” will let us know if such a campaign is feasible, and the “why” will help set marketing strategic direction.
McDaniel and Gates go on to say that in-depth interviews allow research to explore “casual and tangential remarks,” and obtain insight into the main issue. Also, new directions of inquiry can be quickly and dynamically charted. The traditional discussion guide (p.117) of the focus group limits such flexibility.
Furthermore, one segment we will be interviewing is gun retailing. Here the members are competitors to each other and McDaniel and Gates (p. 128) note that in such a circumstance interviews are a better approach. We will need to discuss inventory, advertising and distribution strategies with them that they may not want to reveal in front of competitors.
I believe qualitative research is the best first step. According to Johnson (2009, p. 1), “It [qualitative research] obtains in-depth information on motivations, feelings and behaviors.” Wilkerson (2003, p. 26) says that qualitative research such as interviews allows for “broad and open-ended questions in hopes that the author would be able to capture any significant information” not in our preconceived notion of the issue. Johnson (2009, p. 3) goes on to say that in-depth interviews can reveal detailed aspects of consumer behavior. This makes qulaitative research the best first step in a campaign to introduce modern Ruger revolvers to Europe.
References
Dodds, Paisley (April 17, 2007). U.S. Gun Laws Draw Heat After Massacre. AP. Retrieved on June 7, 2009 from http://www.sfgate.com/cgi-bin/article.cgi?f=/n/a/2007/04/17/international/i143553D58.DTL
Johnson, E. (2009). Focus Groups & Other Qualitative Methodologies. Retrieved on June 7, 2009 from http://www.imc.wvu.edu/.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
MacInnis, Laurin (August 28, 2007). U.S. most armed country with 90 guns per 100 people. Reuters. Retrieved on June 7, 2009 from http://www.reuters.com/article/topNews/idUSL2834893820070828
SEC (2008). Sturm, Ruger Annual 10-K Report. Retrieved on June 7, from http://www.sec.gov/Archives/edgar/data/95029/000119380509000482/e605095_10k-sturmruger.txt
Wikipedia (2009). Gun Politics. Retrieved on June 7, 2009 from http://en.wikipedia.org/wiki/Gun_politics
Wilkerson, Kristen Courtney (2003). Cyber-Campaigning for Congress: A Cultural Analysis of House Candidate Web Sites. UT Austin Dissertation.
Saturday, May 30, 2009
A Site for Statrats
A B2B Report: The Census Bureau ICT
Summary
The Census Bureau has a survey report on its Web site known as the Information and Communication Technology Survey (ICT). Its purpose is to make “available data on e-business infrastructure investment by nonfarm businesses.” It is a supplement to the Annual Capital Expenditures Survey. Microsoft makes significant sales to businesses and these are aggregated in the ICT supplement. Here is a press report on the ICT:
Press Release
and here is the actual report:
Full Report
It lets me see sales data and trends in the information technology B2B sector of the economy with separate segments for computer and communications. The most current data, published in February 2009, is for 2007. No one accused the Census Bureau of being speedy. The report is a series of hyperlinks that let me dig as deep as I want, and it even provides me with Excel worksheets of data so I can graph and chart according to my specific interests.
I hope Microsoft is making use of this survey Web report! Dell too.
Marketing Research Utility
I worry for Microsoft. It is in a Gambler’s Ruin competition with Open Source software, which strikes at the heart of Microsoft – its value add to the economy. Open Source offers for free the software that Microsoft sells and that is the basis for its enormous revenue stream. Gambler’s Ruin is a rivalry between two parties, one with a large, well-established position and the other with a small position but able to compete effectively on individual transactions. Of course, Microsoft is the house. Prudence would dictate they worry about the growth of Open Source.
They should worry too, according to the results of the Census 2007 ICT survey. This table from the bureau gives the raw data 2007 ICT Table 1a.
I have graphed it to show the growing problem.
I derived these from the Census data. As a side note, Linux (Open Source) demolished Sun Microsystems in business server rooms, and Sun is now defunct. Microsoft is fighting hard to maintain its share in the server room, and almost succeeding.
Microsoft has something else to worry over: the growth in software sales is declining year over year. I wonder what is going on in 2008 and today. Kuchinskas (2003, p. 2) reports that in 2000 Dell experienced declining growth in the educational market because of tightening education budgets. They responded with a database-marketing program to the education sector. The Census data above suggests that worse comes. The entire market may be tightening budgets.
Like Dell, Microsoft does monitor market trends using secondary sources like Census, but they also glean information from data mining. In this sense they are the exception to Krol’s observation (2006, p. 1) that technology companies were late to use data to drive sales. One of the selling points we highlighted in MS SQL Server was its data mining capability and how we used it at Microsoft. We could even link to document collections from places like Census.
Unfortunately, my free Census secondary data leaves me short. 2007 data was published in February 2009. Additionally, secondary data may be insufficient. It may not have the granularity I need for a certain conclusion. To get a better sense on my impending demise, I would need primary market research. However, it is worth noting that secondary data, like Census data, may alert the "researcher to potential problems and/or difficulties" (see McDaniel and Gates, 2008, p. 73).
A B2C Report: the Census Bureau E-Stats
Summary
The Census Web site has useful business to consumer information such as the E-Stats Report, (see Ebusiness 614). This report is the result of a commitment by the Census Bureau to measure the electronic economy. It has sales data on e-commerce activities by sector. The sectors are: manufacturing and merchant wholesaling ecommerce sales (considered to B2B) and retail and select service ecommerce (B2C) sales. The data of this report is available at 2006 Final Report.
The report gives more detailed breakdowns in its appendices that are located at Census 2006 Tables. One of interest is the select services (see Census 2006 Table 4). It shows the revenue vs E-revenue in select industry groups (identified by a NAICS code). The Federal Reserve Board uses this report to track the diffusion of the electronic economy (see Dinlersoz and Murillo, 2005).
Marketing Research Utility
Businesses that deal directly with the customer can use this report to track trends by NAICS industry group to see how fast ecommerce to consumers is growing in its industry. For example, one of the industry groups currently in select services is Arts, Entertainment, and Recreation Services and e-commerce revenue is growing 21.9% per year (latest eStats data is for 2006). A company like Domino’s Pizza can use such a trend to partner with a company in the growth sector to expand pizza delivery to the consumer.
TiVo (2009, p 2) claims to have the most advanced capacity in the television industry to track behavior. Delany and Steel (2007, p. 1) note that Microsoft engages in behavioral targeting, and according to Jopling (2006, p 3-5), Microsoft has taken a commanding lead in IPTV technology. Microsoft is always looking for new industry segments to exploit, but the question for marketing research is which ones. The Census E-Stats report has useful data to tell us which NAICS codes have strong electronic diffusion. Also, as IPTV grows, Microsoft may eventually want to partner with TiVo and its advanced technology, to provide service to the Arts, Entertainment, and Recreation industry.
References
Delany, Kevin and Emily Steel (October 2007). Firm Mines Offline Data To Target Online Ads. Wall Street Journal.
Dinlersoz, EM and Hernández-Murillo, R (January-February 2005). The Diffusion of Electronic Business in the United States. FEDERAL RESERVE BANK OF ST. LOUIS REVIEW. Retrieved on May 30, 2009 from EBSCOHOST.
Jopling, Elroy (4 January 2006). Microsoft's Global Consumer Play Begins to Unfold. Gartner, id number G00136841
Krol, Carol (June 12, 2006). Connecting the data dots. Pro Quest.
Kuchinskas, Susan (Sep 2003). Data-based Dell. Adweek Magazines' Technology Marketing.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
Tivo (November 2008). Q3 2009 TIVO INC Earnings Conference Call - Final. Fair Disclosure Wire (Quarterly Earnings Reports). Retrieved on May 30, 2009 from EBSCOHOST.
Sunday, May 24, 2009
Cheskin Research

Expertise and Methods
Cheskin is what McDaniel and Gates (2008, p 8) would call an applied research firm, one that helps companies better understand the market. Their specialty is multicultural markets in the US, which have grown faster than any other US consumer market. Cheskin applies a proven research process to help firms gain the insight they need into this area of explosive growth.
Cheskin also notes that the "general market" is showing increasing signs of reaching a tipping point on diversity - having powerful ethnic characteristics that will soon invalidate "general market" strategies. McDaniel and Gates (p. 9) observe that companies use firms like Cheskin to do programmatic research to understand “market segmentation, opportunity analysis, or consumer attitude and product usage studies” (see Cheskin on Opportunity).
Their clearly defined research process is very similar to the marketing research process described in the McDaniel and Gates text.
- The first step is to envision or frame the objectives of the research (see Envision)
- Next, explore to understand consumer needs, segment characteristics, and trends for both competition and the market (see Explore)
- Third, create or evaluate the concepts born from the exploration. (see Create) This means that they additionally do Selective research (see McDaniel and Gates, 2008, p 9)
- An important next step is to craft a well-told story. (see Inspire) McDaniel and Gates (p. 52) say that “this is a key step” because research “must convince management the results are credible.…”
- Finally comes the solid innovation and differentiation for Cheskin, - for a fee they will also act as consumer advocates after the research phase has completed to assure consistency of management action with market need. They call this the Express service (see Express).
Adequacy of Web site
As you can see by the material at the end of the above links, the Cheskin Web site does a good job describing and promoting their services. Furthermore, their site has blogs and podcasts (see Cheskin Blog) and various articles of interest (for example their article on the ROI of Diversity). I also appreciate that their search function has both a directory as well as free text search (see Search). I can go to a topic area and look at only the subject of interest, avoiding the usual irrelevance of many returned results from free text.
How to Improve Web Site
That said, I don’t think they have the perfect site. It is incomplete. As an example, I searched their directory to find if Cheskin used panels and how they handled panel effects. I clicked on Methods and Techniques and the following was presented:
I used their free test search with no better result. Using Yahoo, I was able to find that Cheskin does conduct panels (see Living Room Panel). Dennis (2001, p 1) reports that research firms may create professional panelists who respond differently than the rest of us, the panel effect.
Another issue with the Web site is that it does not organize around customer profiles, but rather it is organized around Cheskin and its functions. In contrast stands the WVU Web site (see WVU Web) that does have customer profiles as well as functions. Each profile tab has content organized according to the interest and comfort level of the profile. University of Maryland is the same (see UM Web). Among others, Lisa Sanders (2007, p 1) advises Website designers to use the concept of “personas” when creating a site. Personas are ”archetypical characters [who] represent specific consumer segments.”
References
Buchanan, Mark (2002). Nexus: Small Worlds and the Groundbreaking Theory of Networks. Norton.
Dennis, J Michael (2001). Are Internet panels creating professional respondents? Marketing Research.
McDaniels, C and R Gates (2008). Marketing Research Essentials. John Wiley.
Sanders, Lisa (4/9/2007). Major marketers get wise to the power of assigning personas. Advertising Age, 00018899 Vol. 78, Issue 15. Retrieved from EBSCOHOST on May 23, 2009
Sunday, May 17, 2009
Everybody Got a Chevy
The artists, Cadillac Don & J-Money did a series of videos on Chevy models. Here’s an example in a Chevy lowrider:
Peanut Butter and Jelly
The "Everybody Got a Chevy" series of videos features a different model in each. It ties in beautifully with print ads commissioned by Chevy employing the Hip Hop artists. The fact that the videos are not traditional video ads only reinforces the leverage in the print ads, since the music videos are now associated with the cars and they appear less mediated. I think we will see more and more non-traditional work federated with traditional.
Sunday, May 10, 2009
Evaluation of The Black Swan
The book is a rich resource for people who must make decisions in the face of uncertainty and incomplete information. It is written at an intuitive level so it does not require sophisticated training in mathematics or philosophy to understand. On the other hand, the subject matter is thoroughly presented. For practioners, its risk management implications cross-cut many business disciplines. For
The Black Swan by Nassim Taleb
Strengths
It is easy to misuse survey statistics and, according to Taleb, quite often done. He provides high-level guidelines that indicate if a statistical approach is questionable (see Implications section of this post). He also gives many aphorisms for how to think clearly and avoid common decision making mistakes. Here is a short sampling:
- Focus on anti-knowledge, what we know as false instead of what we know. What we know is captive to the unknown, but if we disprove the justification for an effort we can save ourselves time and money (p. xxi).
- The future will be less predictable but we still must predict (p. 203).
- To escape the narrative fallacy, make testable predictions. (p. 72).
- Favor experiments over history and theories (p. 83-4, 120).
- Don’t focus. Uncertainty can come from any direction (p. 133).
- Understand scalable and non-scalable variability
- If we can predict people’s actions then they are automatons. Is this a valid (p. 183)?
- Rank assumptions not by likelihood but by harm they can cause (p. 203).
- Invest in preparedness, not in prediction (p. 208).
- Seize asymmetric opportunities: low cost, high payoff activities even if the payoff has a low probability (p. 205).
- Look to be broadly right rather than precisely wrong (p. 284).
It would be easy to misinterpret this work as advocating a carelessness of futurity (for example, see EO, 2007, p3 in his review of the book on Amazon). Such is not the case. It is the realization that Rationalism is dead, Positivism is dead. What do we do now? Taleb is actually giving guidance on the most responsible approach for considering the future. This should be expressly stated: this is not giving up on our dealings with the future but a start on how to properly deal with it.
Information Sources
His could not be stronger. He has personally worked with Benoit Mandelbrot the leading fractal mathematician in the world. He makes extensive use of the work by Nobel Laureate Daniel Kahneman in the Psychology of Uncertainty. It’s the same for network theory, citing the seminal works of mathematicians Watts and Strogatz. Additionally, he brings his own hard earned experience in the trenches of international finance.
Implications for Consumer Research
Foremost, do question surveys and the validity of their statistical assumptions. Taleb notes that physical quantities such as height and weight are subject to some sort of Gaussian distribution (p. 33). Measures dealing with social matters are not usually Gaussian and if not, we have two choices: 1.) the use of Fractal mathematics if the problem is tractable; 2.) run experiments and analyze evidence if intractable.
How do we decide if the population measure is fractal-able? If the measure follows a power rule, in mathematical terms is self-similar, then fractal mathematics applies. A simple power law is the Pareto 80/20 rule, as discussed in last weeks post. If there is a power rule that applies then the distribution is fractal.
How has consumer research faired by universally applying a Gaussian distribution to social matters? As noted in last week's post, there are mixed results and one of the failures may be instructive. Consider New Coke.
Wilson and Ogden (2004, pp 30-1, 52-3) report that Coca Cola used both statistical survey
The conclusion drawn from the statistical surveys was to change the formula. The result from the focus groups was the opposite; change would be a big mistake. Coca Cola decided to discount the soft evidence from focus groups and go with hard data from proven mathematical statistics.
The measures were not physical attributes but opinions and beliefs. Based on Taleb’s work, we should question the applicability of Gaussian distribution assumptions to such measures. They may be valid, but maybe not. Wilson and Ogden conclude (p. 52) “survey research is becoming less credible as an accurate representation of publics….” Contrary to the hard data predictions, New Coke did not fair well.
A second and more positive implication is the explosive power of Internet marketing. Internet connections are fractal – a popular Web page is linked to by other pages, these pages in turn linked to by their interested parties, and soon you get a fractal like spider web of connections. This is the basis for viral marketing, which is fractal. Internet marketing can give us asymmetric leverage, where the Black Swan that occurs is positive for us. A classic example of the fractal explosiveness in viral marketing is Smirnoff’s Tea Partay (see Johnson, 2009, p. 26).
Hawkins, et al (2007, p. 249) note that viral marketing is an online strategy to generate buzz and word of mouth (WOM). Buzz is an idea contagion that “[creates] an exponential expansion of word of mouth.” It is low cost but they go on to say (p. 248) that such strategies must be used with care so no miscommunications diminish the brand.
Taleb (2007, p. 220) would say idea contagions are fractal as well as exponential. Fractal worlds are winner take all. They follow a scalable power rule, something like 10% of the videos capture 90% of the traffic, and 10% of that top 10% capture 90% of that 90%. In such a world the top 1% owns a lot, 81% in this example. For WOM marketing online, this means you front $600K to produce a video but most times it doesn’t hit, but when it does it can hit big.
Taleb advises us (p. 205) that living today requires a lot more imagination because it is a world dominated by extremes and the unknown. He goes on to say that consumer behaviorists should seize asymmetric opportunities where you make a series of small bets that you will lose for the occasional big payoff. This is the most effective strategy in the world we are rapidly becoming.
References
Cacioppo, John and Richard Petty (1986.) The Elaboration Likelihood Model of Persuasion. Retrieved on April 13, 2009 from the EBSCOHost database.
EO (April 25, 2007).Many important ideas, many flaws that detract from the message. Retrieved on April 14, 2009 from http://www.amazon.com/Black-Swan-Impact-Highly-Improbable/dp/1400063515/ref=sr_1_1?ie=UTF8&s=books&qid=1239724317&sr=8-1
Hawkins, Del, David Mothersbaugh and Roger Best (2007). Consumer Behavior. McGraw-Hill/Irwin.
Johnson, Celia (March 6, 2009). 10 of the Best. BANDT-COM.AU. Retrieved on April 18, 2009 from EBSCOHOST.
Ortega y Gasset, Jose (1994). The Revolt of the Masses. W. W. Norton & Company.
Simon, H.A. (1960). Administrative Behavior: A Study of Decision-Making Processes in Administrative Organization. Macmillan.
Taleb, Nassim Nickolas (2007). The Black Swan. Random House.
Wallace, AFC (1963). Culture and Personality. Random House.
Wilson, L. and Ogden, J. (2004). Strategic Communications Planning For Effective Public Relations and Marketing, 4th Ed. Kendall/Hunt Publishing
Sunday, May 3, 2009
Review of The Black Swan by Nassim Taleb
The Black Swan by Nassim Taleb is a skeptical view of the Rationalism employed in modern behavioral sciences. It is written with eloquence so that complex ideas and mathematics are made intuitively clear. It is particularly relevant for consumer and marketing research, which have continuously experienced embarrassing and costly failures such as New Coke, Life Savers Soda, Colgate Kitchen Entrees, Pond’s toothpaste, Clairol’s ‘Touch of Yogurt’ shampoo, Frito-Lay Lemonade, Pepsi AM, and Heinz’s All Natural Cleaning Vinegar.
Experienced industry professionals in leading companies did these projects. It is not just consumer research either, as American financial models have recently gone bust in a highly visible manner with the rest of the planet watching in total horror. Why the mixed results from research based on the Rationalist models?
Taleb gives a roadmap that not only explains the misuse of mathematics in such predictive attempts but also explains the fallacies of reasoning possible with Rationalism that lead to a false confidence in our undertakings and an understatement of the risk from random but material future events. The Black Swan is his metaphor for a risk from unknown events with consequential effects.
My review starts with Taleb’s recounting the numerous points of failure in Rationalist reasoning such as domain specificity, post hoc rationalization, the narrative fallacy, and silent evidence. It then explains the abuse of mathematics cited by Taleb starting with the circularity of statistics, and the pervasive but often invalid assumption that our distribution of attributes is non-scalable.
Examination
Nassim Taleb gives us a practioner’s guide to the pitfalls in Rationalist reasoning. He starts with the all too human tendency to wrongly translate an absence of proof into proof of absence concerning risk. His delightful example is a thought experiment with a turkey that is well fed and cared for by his human host. Using the inductive methods of Rationalism with day after day supporting proof, the turkey concludes that his human benefactors have his best interests at heart. There is a sudden “revision of belief” (p. 40) on the Wednesday just before Thanksgiving. Of human malice, the turkey had confused absence of proof with proof of absence regarding the risk he faced.
In the first section of his book, Taleb explains the common fallacies of Rationalism. These fallacies include (p. 50) the confirmation error, the narrative fallacy, and the distortion of silent evidence.
Confirmation Error
Taleb observes that the context of the information presented to us influences our thinking about that information (p. 53). The information does not stand on its own merit but that of its presentation context as well. Taleb calls this Domain Specificity, and Hawkins, et al (pp 299-300) call it contextual cues, and explain its impact on consumer behavior.
Another confirmation error is naïve empiricism (p. 55). This is the human inclination to look for support of our vision and to orient research with this positive frame of mind. It only takes past instances that confirm current proposals.
Narrative Fallacy
This is a predilection for simple explanations in place of complex truths. Taleb’s exposition uses a cognition model similar to the Elaboration Likelihood Model used in consumer behavior (see Hawkins, 2007, p 409-10). Taleb describes (p. 81) the cognitive model devised by the eminent psychologist Kahneman. This model organizes cognition into System 1 thinking and System 2 thinking. System 1 is intuitive and quick, relying on heuristic short cuts. It gives easy and obvious narratives but overemphasizes the emotional and the sensational.
System 2 is what we would characterize as central route processing. It is a derived sequence of thought. It is easy to retrace reasoning to rethink our strategy based on feedback. On the other hand, System 1 thinking is prone to narrative fallacies.
Narrative fallacies take several forms. One is Post Hoc Rationalization. This fallacy provides an artificial explanation of an event after the fact rather than establishing causal relationships during the event. Taleb gives the classic example (p. 65) of a group of consumers who each selected a pair of nylons from a set of twelve. A while later they were asked why they made their particular choice. The answers ranged from better color to better texture. The twelve pairs were in fact identical. Hawkins, et al (2007, p 326) report on a similar happening with Disney and Bugs Bunny.
Finally, the more randomness in information the harder it is to remember (p. 69). We therefore seek to summarize random information and impose our own order on it. We fold meanings into convenient dimensions of existing knowledge. This reduces the dimensionality making it less complex and so easier to store and retrieve. This makes the world look less random, and therefore less risky. This is why we tend to underestimate risk, especially risk that does not fit into our existing knowledge dimensions.
Distortion of Silent Evidence
History is a graveyard of Silent Evidence, as Taleb calls it. The simplification biases discussed above reduce complex evidence into convenient summaries. The omissions add to the silent evidence we ignore, which distorts our view of reality. The manifestation of silent evidence is a false sense of stability (p. 117).
The Scandal of Prediction
In the later sections of the book (pp 136-211), Taleb makes an intuitive case for why our predictive models fail. One critical aspect of a system being modeled is its scalability. In the behavioral sciences most ranges are assumed to be non-scalable. In other words, as you leave the mean, not only is the count less, but that it is increasingly less. This is a convenient assumption because it permits the use of statistical mathematics based on the Bell curve (a.k.a. Gaussian distribution) or a derivation of it.
This assumption about a Bell curve for our populations, as Taleb (2007, pp 229-247) argues, is “that great intellectual fraud.” This is not a true attribute of all the populations where behavioral scientists are applying statistical surveys in a wooden and perfunctory manner. He notes that while it is true for physical characteristics such as height and weight, it is usually not true for social measures. Ranges become scalable so the non-scalable assumption of the Bell Curve is invalid. Scalable system behavior leads to non-uniform concentrations rather than smooth distributions, they are thus Fractal. The rich get richer.
Fractal worlds follow a scalable power rule. As a simple illustration, Pareto found that 20% of the Italian population owned 80% of the land (p. 235), and the 20% of that top 20% owned 80% of that 80%. In such a world the top 1% owns a lot, 64% in Pareto’s case. Taleb also uses book sales as an example (p. 264). It does not follow a Bell Curve. It is fractal and follows a complex power rule. Because it is a fractal world it has winner take all, lop-sided distributions.
He uses height and wealth as examples of applying Bell Curve models in each world (non-scalable and scalable). For height, if you pick 100 people randomly, you will derive a meaningful understanding about the height of the population. Adding another person, the 101st, won’t measurably change the average or deviation, even if it is a tall person, say seven feet. This is the standard Gaussian system.
The social measure of wealth is different. If the 101st person you add is Bill Gates, the average and deviation is changed, appreciably. This is an extreme example to make a point, but Taleb discusses his days on Wall Street where non-scalable assumptions were made for scalable systems, which led to misunderstandings of risk and incorrect investment strategies. He shows how scalable systems are described by fractal mathematics.
Traditional mathematical modeling in social sciences, including behavioral sciences, is flawed. The process of employing mathematics starts with the Circularity of Statistics flaw (p. 269). We need data to know if the population is Gaussian or Fractal. But, we need to know if the population is Gaussian or Fractal to know how much data to collect to decide if it is Gaussian or Fractal.
Let’s say we can get by this problem. Then we encounter another problem for Gaussian distributions (p. 251). Gaussian models in pure mathematics are based on the assumption that each event is mutually exclusive. This is true of flipping a coin but not true in most social actions, where there is usually some cumulative advantage effect like learning. In other words, there should be improvement in the probability of a certain outcome over time because of the cumulative advantage effect of learning.
For the other case, if the model turns out to be Fractal rather than Gaussian, we still have problems (p. 272). Fractal mathematics for randomness does not yield precise answers. The Gaussian does and that is why scientists like to make Gaussian assumptions.
The next post will apply Taleb to consumer research.
References
Cacioppo, John and Richard Petty (1986.) The Elaboration Likelihood Model of Persuasion. Retrieved on April 13, 2009 from the EBSCOHost database.
EO (April 25, 2007).Many important ideas, many flaws that detract from the message. Retrieved on April 14, 2009 from http://www.amazon.com/Black-Swan-Impact-Highly-Improbable/dp/1400063515/ref=sr_1_1?ie=UTF8&s=books&qid=1239724317&sr=8-1
Hawkins, Del, David Mothersbaugh and Roger Best (2007). Consumer Behavior. McGraw-Hill/Irwin.
Johnson, Celia (March 6, 2009). 10 of the Best. BANDT-COM.AU. Retrieved on April 18, 2009 from EBSCOHOST.
Ortega y Gasset, Jose (1994). The Revolt of the Masses. W. W. Norton & Company.
Simon, H.A. (1960). Administrative Behavior: A Study of Decision-Making Processes in Administrative Organization. Macmillan.
Taleb, Nassim Nickolas (2007). The Black Swan. Random House.
Wallace, AFC (1963). Culture and Personality. Random House.
Wilson, L. and Ogden, J. (2004). Strategic Communications Planning For Effective Public Relations and Marketing, 4th Ed. Kendall/Hunt Publishing
Saturday, April 25, 2009
Marketing Blunder or Urban Myth?
Many gaffes are humorous and because of this there is also a set of urban myths that have been fabricated. To help avoid these, a literary search of myth busters should be done when reporting on marketing blunders. Here are three that are not urban myths you might find amusing.
Eat Elegancita
Hershey’s Elegancita candy bar is filled with caramel cream, cajeta in Mexican, and covered with chocolate. A popular ad campaign in Mexico was extended to other Latin American countries that spoke Spanish with unfortunate result. Cajeta in Argentina and Uruguay, who speak Castilian Spanish, is a slang term for a women’s vagina.
The singer Thalia tells all that she loves cajeta. (see Hoag, 2005, p 1). Hoag goes on to say “Such linguistic pitfalls, where a common word may be understood in a different way in different countries, happen with some frequency in Spanish.” A contributing factor: there are sixteen different dialects of Spanish (see Spanish Dialects and Varieties) .
This story withstands the scrutiny of an urban myth buster (see About Spanish). This guy does torpedo many other urban myths about marketing gaffes, including allegedly misusing the Spanish language. Nova Myth and Other Urban Myths.
Cultural Confusion at IKEA
Money Magazine (see Bad Translations) reports on IKEA, as does Brand Strategy (2003, p 1). The IKEA bunk bed for children is named Gutvik. This is a homonym in German for Good Fxxx. Muy Bueno! Just the right name for a children’s product. The Urban Myth busters say that to be fair, the German words are Gut Fick, and the phonetics are slightly different. (see TV Tropes) However, they do cite IKEA for another misstep, marketing the Viren Toilet Brush (see IKEA Catalog) in Finland. The legendary runner Lasse Viren hails from Finland.

Hakkaa päälle, Pohjan poika!
I Could Not Love You More If I Loved Pepsodent Less
Dr. Monica Bolesta at the University of Maryland (2008, p 2) discusses how Pepsodent failed to launch its toothpaste in Southeast Asia because it emphasized the whitening aspects.
Pepsodent wrongly assumed the American cultural preference for white teeth was universal. In Southeast Asia, darkened teeth are considered more attractive and the locals even chew betel nuts to stain their teeth. The betel nut itself is considered a sign of love and longevity (see Betel Nut Smile).
The Culitvated Drink for Apes
Pavane is an upscale bitters distilled and marketed by Bacardi. According to both CNN Money (CNN Bad Translations) and Zouhali-Worrall (2008, p 1) in a Fortune Small Business article, product marketing became a cultural blunder in Germany. The German homonym Pavian means baboon.
References
Bolesta , Monica (August 22, 2008). Cross-Cultural Awareness: Avoiding Global Marketplace Missteps. Retrieved on April 3, 2009 from http://www.umuc.edu/departments/podcasts/podcasts_bep/content/shows/sh2008/transcript/Transcript_Diversity_Cultural_Awareness.pdf
Brand Strategy (May 2003). Global village idiot. Source:Brand Strategy; May2003 Issue 171, p39, 1/9p. Retrieved on EBSCOHOST.
Hawkins, Del, David Mothersbaugh and Roger Best (2007). Consumer Behavior. McGraw-Hill/Irwin.
Hoag , Christina (03/08/2005). The Miami Herald. Retrieved on April 3, 2009 from EBSCOHOST.
Zouhali-Worrall, Malika (Jul/Aug2008). Watch Your Language! Fortune Small Business. Retrieved on April 3, 2009 from EBSCOHOST.
