Auditing,
Forecasting, and Market Research
How
They Work Together
1.
Understanding the Roles of Each Component
Auditing involves a
systematic review of an organization’s processes, strategies, and performance
metrics to ensure compliance with established standards and best practices. It
helps identify areas of strength and weakness within marketing functions or
forecasting methods.
Forecasting is the
process of predicting future trends based on historical data and analysis. It
plays a crucial role in planning by providing insights into potential market
conditions, customer behavior, and resource needs.
Market Research
gathers information about consumer preferences, market trends, and competitive
dynamics. This data is essential for making informed decisions regarding
product development, marketing strategies, and overall business direction.
2. Auditing, Forecasting, and Market Research
Relationship
three components are
interconnected in several ways:
Data Integrity: Auditing
ensures that the data used in forecasting is accurate and reliable. By
verifying the quality of data sources through audits, organizations can improve
the accuracy of their forecasts. For instance, if an audit reveals
inconsistencies in sales data collection methods, adjustments can be made to
enhance future forecasting efforts.
Informed Decision-Making:
Market research provides valuable insights that inform both auditing processes
and forecasting models. By understanding consumer behavior and market dynamics
through research, organizations can set realistic goals during audits and
create more accurate forecasts that reflect current market conditions.
Continuous Improvement: The
feedback loop created by these three components fosters continuous improvement.
Audits can identify gaps in forecasting accuracy or market research
methodologies. In turn, improved forecasting techniques can lead to better-informed
market research initiatives that align with organizational goals.
3. Practical Application of
Their Integration
Integrating auditing, forecasting, and market
research allows organizations to:
Develop comprehensive marketing strategies
that are grounded in solid data.
Allocate resources more effectively by
understanding projected demand through accurate forecasts.
Monitor performance against established
benchmarks using insights gained from audits.
For example, a company may
conduct a marketing audit to assess its current strategies’ effectiveness while
simultaneously analyzing forecasted sales figures derived from recent market
research findings. This holistic approach enables decision-makers to adjust
their tactics proactively based on empirical evidence rather than assumptions.
4. Examples of Bringing it Together (excerpted from Springer Book)
Regression and Local Market Area Regression is very valuable in
adjusting predictions using categorical adjustments for various demographics
factors such as geographic region and other characteristics such as gender.
Geographic region can also be a surrogate for income distribution which is
readily available from government published data. For example, there is
available through census data sources that break down income tax collections or
median income by zip code. Such data combined with a company’s sales data is a
good indicator distribution of the socioeconomic characteristics of a
business’ customer base.
The compilation of this
type of socioeconomic dataset requires time, but gaining insight of the
customer base is invaluable. The first step is to set up a dataset from a sales
account with links to customers’ zip code. A summary of sales data by individual
zip code would be very unyielding to interpret. The first step is to group zip
code into broader zip code areas. This is where background research of census
data and a map to develop relevant zip code areas (Table 8.1).
The sales and
sociodemographic research resulted in condensing the data into four geographic
areas. This number was chosen to simplify the presentation of the concept of
using regression to project sales by geographic area. It is most likely that
many more areas would be of value, especially for larger businesses. From a
statistical data perspective, zip codes are categorical data.
Table 8.1 Sales Data by
Zio Code and by Four Geographic Areas
Regression results:
R=.86
R2=.74
Sales = -$30,885 + $47,049 x Zip (Geographic Area number)
Used Excel pivot table to
summarize sales by zip code area
As part of small business forecasting,
it is key to get a picture of the possibilities for selling products or
services in a local market. Looking at local markets will provide information
about the types of individuals who might buy products or services and how
extensive is the company’s geographic reach and what is the competition within
the various market areas.
Create a Customer Profile Next there is a need to determine who
are the people who will buy products or services.
What
age are they?
What
is their income level?
What
is their education level?
What
kind of jobs do they have?
What do they like to do
for entertainment?
It may be too cumbersome and difficult
for a small business to survey for such data. But a small demographically
diverse focus group is a proven way to measure customer opinions. It is set up
in guided or open discussions about new products or current views of the
company to determine reactions that can be expected from a larger population.
The use of focus groups is a market research method that is intended to collect
data through interactive and directed discussions by a researcher. If there are
issues with lagging sales that don’t respond to standard means of marketing,
then arranging for a focus group may be what is needed.
Obtaining Customer Ratings
A Likert scale is a scale
commonly involved in research that employs questionnaires. It is the most
widely used approach to scaling responses in survey research, such that it is
often used interchangeably with rating scale, although there are other types of
rating scales.
The scale is in a
format in which responses are scored along a range. When responding to a Likert
item, respondents specify their level of agreement or disagreement on a
symmetric agree-disagree scale for a series of statements. Thus, the range
captures the intensity of one’s feelings for a given item. The Likert scale has
found widespread use in business and marketing, primarily because of its
simplicity.
A scale can be created as the simple sum or average of
questionnaire responses over the set of individual items. Likert scaling
assumes distances between each choice on the sale are equal. The design of a
set of scale items is such that they are highly correlated but also that
together will capture a full range of customer preferences.
A Likert item is a statement that a customer respondent is
asked to evaluate by giving it a quantitative value based on a level of
agreement/disagreement being the dimension most commonly used.
The format of a typical
five-level Likert item, for example, could be:
·
Strongly
disagree
·
Disagree
·
Neither
agree nor disagree
·
Agree
·
Strongly
agree
It is a bipolar
scaling method, measuring either positive or negative response to a statement.
Sometimes an even-point scale is used, where the middle option of “neither
agree nor disagree” is not available. This is sometimes called a “forced
choice” method, since the neutral option is removed. The neutral option can be
seen as an easy option to take when a customer is unsure, but there is much
discussion if it is a true neutral option or just when the respondent is
confused. It has been discussed that there is not a significant difference in
the use of forced neutral or not.
Likert scales
may be subject to distortion from several causes. Respondents may:
·
Avoid using extreme
response categories (central tendency bias), especially out of a desire to
avoid being perceived as having extremist views.
·
Agree with statements as
presented (acquiescence bias), who may by an eagerness to please.
·
Disagree with sentences as
presented out of a defensive desire to avoid making erroneous statements and/or
avoid negative consequences that respondents may fear will result from their
answers being used against them.
·
Try to portray themselves
in a light that they believe the examiner or society to consider more favorable
than their true beliefs.
The biases listed become
paramount when questioning individuals regarding socially or highly personal
issues. But for the small business manager, the dimensions are primarily three
dimensions of primary concern:
1.
Overall customer
satisfaction with interaction with the business.
2.
Satisfaction of value of
goods or services based on price and other perceived values.
3.
Likelihood of repeat
business and recommend to others.
For the small business, these scales
are a straightforward indicator of customer satisfaction with the business.
They do not deal with more highly charged social and personal assessments. The
issue of bias is not a major concern.
Question Design After the questionnaire is completed,
below are things to keep in mind when formulating individual questions:
·
Make the questions very specific.
Notwithstanding the importance of brevity and simplicity, there are occasions
when it is advisable to lengthen the question by adding clarification. For
example, it is good practice to be specific with time periods.
·
Avoid jargon or shorthand.
It cannot be assumed that respondents will understand words commonly used by
people in the business. Trade jargon, acronyms, and initials should be avoided
unless they are in everyday use.
·
Steer clear of
sophisticated or uncommon words. A question is not a place to score literary
points, so only use words in common parlance. Colloquialisms are acceptable if
they will be understood by everybody.
·
Avoid ambiguous words.
Words such as “usually” or “frequently” have no specific meaning and need
qualifying. Avoid questions with a negative in them. Questions are more
difficult to understand if they are asked in a negative sense. It is better to
say “Do you ever ...?”, as opposed to “Do you never ...?Avoid hypothetical
questions. It is difficult to answer questions on imaginary situations. Answers
may be given but they cannot necessarily be trusted.
· Do not use words which could be
misheard. This is especially important when the question is administered over
the telephone. For example, fifteen and fifty can sound very similar.
· Desensitize questions by using
response bands. Questions which ask about age is best presented as a range of
response bands. This softens the question by indicating that precision isn’t
necessary and only a broad answer is needed.
Forecasting
Economic Potential There
is a treasure trove of economic information for market research that is
contained in business accounting systems. But to unlock that, information
requires careful planning. Auditmetrics experience is that standard accounting
system reports are useful, but the level of detail may not be sufficient. In a
business assessment concerning the lack of growth of a new product, the
immediate response of the business manager is usually “let’s increase the advertising
budget,” a typical marketing research response. A review of product sales data
for the period of time before and after the initiation of the product was
conducted. A random sample was selected to pull records to interview employees
and customers. The results indicated staff were not familiar with the
requirements of the new product that unfortunately led to customers to be
confused. In an audit it was found that there was an unusually high number of
product returns. More training of employees is what helped in increasing sales.
Fortunately, this assessment was done quickly enough to avoid the company
having an image problem.
Routinely Sample Accounts
to Monitor Business Activities Many business managers complain that selecting account
samples takes a lot of time. It is better to sell rather than sample. There is
nothing wrong with pushing for more sales as long as there are not some
unforeseen barriers.
What helps is to
make inroads on the time issue. Regular timely random samples allow the
business manager to deal with small workable subsets of account data
representative of the total book. There is no need to use the gold standard of
3% which would require a fairly substantial random sample. The other time saver
is the Auditmetrics AI assistance software feature. This makes it possible to
rapidly generate a random sample by simply deciding on the desired margin of
error and test different number of strata. The business manager can make more
rapid forecasts using MS Excel in effect putting the business on a monitor.
Account Data and Customer Input In this section Likert customer rating
is linked to customer sales data. It was designed as a preliminary small sample
looking at linking customer ratings on price and satisfaction and whether they
would recommend the business to others. The random sample was derived from a
QuickBooks report that list total sales by customer (Table 8.2):
The variables are:
Zip_Area—That is the customers’ zip code
aggregated by areas based on census data that delineate areas of different
socioeconomic characteristics based on median family income.
Table 8.2 Likert average Likert scale rating and
total sales by zip area
Lprice—Average Likert
customer rating evaluating price and value of the product LSatisfied—Average rating for satisfaction with interaction with
business staff LrRecommend—Average
rating for repeat business and recommend to others
Average Likert rating is used even
though there are scaling issues discussed previously. This assessment is
intended to be a quick look at consumer opinions because time is also of value
to the customer. A more detailed analysis can be done with a focus group
recruiting customers for a more detailed assessment. It usually involves offering
some sort of compensation not necessarily cash but product or service discounts
as an alternative.
The next step in the analysis is to
conduct regression analysis with zip area as the dependent variable and the
Likert ratings as the independent variables with the following results:
Multiple R = 0.46
N =60
The
correlation coefficient is .46. As expected, data variables that measure
people’s attitudes are not as predictable as the prior regressions involving
account data. So as an ongoing prediction model goodness of fit, this provides
a preliminary look. Below is the table that gives each independent variable’s
alpha error. The decision is if the observed coefficients occur by random
chance alone is less than 5%, then one can assume there is a measurable effect
of that scale by zip area. The Likert rating for price does differ among the
different zip areas. Though this is a preliminary snapshot, the observed alpha
error (p-value) is so small that it should require further analysis (Table
8.3).
Zip Area 1 is a geographic area which
is in the lowest end of median family income based on census data. A breakdown
of average sales by that zip area is also at the low end. It may be wise to
have an advertising campaign specifically tailored
Table
8.3 Statistical significance
of Likert ratings
Table
8.4 Average sales by zip area
to those families in Area
1. That is the area that may increase demand if price discounts are offered.
There can be repercussions if sociodemographic data is used, so further
assessment should be conducted. If it happens that this geographic area has an
older population, mostly retirees on fixed incomes, then a senior discount for
all customers is appropriate (Table 8.4).
Total Process Overview The overall process in conducting forecasting and market
research is to:
1.
Start with a random sample
of accounts.
2.
From there use regression
to project revenue and expenses.
3.
Also add to the account
data pertinent variables such as geographic and socio-demographic data.
4.
Set up a mechanism to
obtain customer ratings using Likert scales.
Step 4 should be part of a total
package to obtain customer loyalty. To truly build this loyalty, companies need
to move from transaction interaction with their customers to building company
customer relationships. The first step in building these relationships is
engaging with customers beyond basic one-way dialog. Customers don’t feel
valued when it takes undo time to contact the business they patronize. At the
same time, sending out mass text messages without a prompt response will also
not give customers a satisfying feeling either. Correct proactive outreach can
help organizations maximize productivity, customer satisfaction, and
contributions to the bottom line.
Though much discussion in this book involved quantitative
measurements structured to act as part of a business performance monitoring
process. Is it worth it? The quantitative methods will expand control of day-to-day
operations. Also, when seeking funding for current operations and new business
plans, the quantitative methods discussed follow both AICPA and IRS standards
at a level of statistical sophistication that usually is available only to
large corporations.
5. Conclusion: A Synergistic Approach
In
conclusion, auditing, forecasting, and market research work together
synergistically to enhance organizational effectiveness. By ensuring data
integrity through audits, leveraging insights from market research for informed
decision-making, and utilizing accurate forecasts for strategic planning,
businesses can achieve greater success in their operations.
6. Resources for this Discussion
Auditmetrics Small
Business Power Series books available on Amazon:
Both Kindle and Paper Bound Available
Statistical Audit
Automation
Applying
Artificial Intelligence Techniques
Forecasting
Revenue and Expenses for Small Business
Using
Statistical Analytics
Market Research for Small
Business
Using Statistical
Analytics
7. Value Added in Healthcare
and Public Health
Value added is the
extra value created over and above the original value of something
• For private business it is usually the
products sold to the consumer
• It
is the difference between a product final selling price and the direct and
indirect
expenses incurred in providing that product
In healthcare and
public health the challenge is how to measure value added
•Research into organizations that have
achieved better health outcomes while often
lowering costs suggests a strategic framework
for value-based public health and
healthcare
implementation
•Focusing on health outcomes aligns how
patients experience their health with links to
the investment
incurred
•This
is the basis of cost effectiveness and cost benefit analysis of public health
and
healthcare programs
Also available on Amazon:
HealthLink Wellness: Science for the
Individual