2.3 Market Research

Primary and secondary research, hypothesis testing, and the data visualizations that communicate findings.

Part 1 of 5 What Market Research Is

Market research is the planned collection of detailed information about markets, products and customer behavior, so marketing decisions rest on evidence. A business plans its research before it knows the answer. The business designs the research so that the result can show the business's expectation is wrong.

The running example of this unit is a teenager who rebuilds used bikes and sells them at a Saturday market with a signed inspection tag. The numbers below are example figures.

Two kinds of data

	Quantitative	Qualitative
What it is	Numbers that can be measured and analyzed	Descriptive data in words and images
It answers	How many, how much, how often	Why and how
At the stall	Seven of nine bikes sold	"The signature on the tag is what convinced me"

Three checks before a launch

A product is	When	For the inspected bike
Desirable	It creates value by achieving problem-solution fit	Parents cannot tell a safe used bike from a dangerous one; a documented check answers exactly that
Feasible	The business can produce and provide it within its resources, technology, expertise and time	An hour per bike, a checklist and a zip tie fit inside a Saturday
Viable	It has the potential to be profitable in that market	The tag earns about $20 more for that hour

Research does not stop at launch. A business with existing products keeps researching to follow customers' changing needs, spot openings for innovation, and hold or grow market share.

Part 2 of 5 Secondary Sources First

Secondary research gathers information that already exists in government, commercial and academic publications and databases: market size in dollars and in customers, trends, segments, and what drives customer decisions. Businesses use both secondary and primary research to learn about customers, competitors and the market landscape, including the PESTEL factors (political, economic, social, technological, environmental, legal). Secondary research comes first because reading data somebody else paid to collect is far cheaper than collecting your own. Reading sixty local listings in one evening shows the competitive landscape. Untagged commuter bikes are listed near $75, and no listing documents an inspection.

Part 3 of 5 Testing a Hypothesis With Primary Research

A business hypothesis states an assumption about a customer, a product or a market, and the business tests it before committing to a course of action. The teenager's hypothesis is that a bike with a documented inspection tag sells for twenty dollars more, and sells faster, than an identical bike without a tag. The business then picks the primary research method that produces the data the hypothesis needs.

Choosing a primary method

Method	Use it when you need
Surveys	A large volume of quantitative data reflecting a whole population
Focus groups and interviews	Deep qualitative data from a few engaged customers, with follow-up questions
Experiments and observations	What customers do rather than what they say: controlled in an experiment, natural in an observation
A/B testing, a type of experiment	Authentic responses to two viable alternatives in a real setting

A bar chart: nine tagged bikes at ninety-five dollars, seven sold, seventy-eight percent; eight untagged bikes at seventy-five dollars, three sold, thirty-eight percent. — Example numbers. The tagged bikes sold at about double the rate of the untagged bikes, at a price $20 higher.

The teenager offered seventeen comparable bikes over two Saturdays: nine tagged bikes at $95 and eight untagged bikes at $75. Seven of the nine tagged bikes sold, and three of the eight untagged bikes sold. Twelve parent interviews then supplied the reason: most parents pointed at the signature line. Seventeen bikes is too small a sample to be sure of anything. The result supports the hypothesis. The next test should use more bikes.

Part 4 of 5 Keeping the Data Honest

Skewed data has two standard causes: a sample that does not stand for the population, and questions that are not neutral. The first fix is a sample that is large enough and drawn from across the whole population. Everyone interviewed stood at one Saturday-market stall, so the findings say nothing about buyers who shop only online. The second fix is neutral questions. Asking whether the tag makes you feel safer suggests the answer the seller wants. Ask what, if anything, influenced the decision today.

Part 5 of 5 Showing the Findings

A data visualization turns rows of data into a pattern, trend or insight a decision-maker can absorb quickly. Pick the chart that shows the relationship you mean to communicate.

Four chart types

Chart	Shows	Example
Bar chart	Comparisons between individual data points	Share sold, tagged rack against untagged
Stacked bar chart	A total broken into its subcategories	Monthly sales split across three price tiers
Line graph	A trend over time	Customers per month climbing toward August
Pie chart	Part-to-whole relationships	Share of a market held by each seller

The count reports what happened. The interview answers explain why. Real decisions need both.

Worked Examples

Reading an A/B test result

Convert raw A/B test counts into comparable rates and state what the test does and does not prove.

Across two Saturdays Theo racks seventeen comparable commuter bikes. Nine carry a signed inspection tag with a price of ninety five dollars. The other eight are untagged and priced at seventy five. By the close of the second Saturday, seven of the tagged bikes and three of the untagged bikes have sold. Convert the counts into rates and interpret the result against his hypothesis that a documented inspection sells a bike for twenty dollars more, and faster.

Tagged bikes racked: 9
Tagged bikes sold: 7
Untagged bikes racked: 8
Untagged bikes sold: 3
Tagged price: $95
Untagged price: $75

1. Explain why raw counts cannot be compared
Seven and three look comparable, but the two racks did not hold the same number of bikes. Nine and eight are different denominators, so the counts have to become rates before either group can be judged against the other.
2. Compute the sell through rate for the tagged rack
Seven sold out of nine racked is about seventy eight percent.
$\frac{7}{9} \approx 0.78$
3. Compute the sell through rate for the untagged rack
Three sold out of eight racked is about thirty eight percent.
$\frac{3}{8} = 0.375$
4. Compare the two rates
Seventy eight percent against thirty eight percent is roughly double the sell through rate, and the tagged rack achieved it while charging twenty dollars more per bike.
$\frac{0.78}{0.375} \approx 2.1$
5. State the limits of what was measured
Seventeen bikes over two Saturdays at one market is a small sample drawn from one place. The evidence supports the hypothesis and does not settle it, and weather or a holiday crowd could bend a result this size.

Check answer

Answer
78% tagged against 38% untagged. The tagged rack sold at about seventy eight percent against about thirty eight percent untagged, roughly double the rate at a twenty dollar higher price. The hypothesis is supported by a small sample.

Why it matters
An A/B test earns its authority from measuring behavior rather than opinion, and it loses that authority if the two groups are not otherwise comparable. Always convert to rates before comparing, and always name the sample limitation in the same breath as the result.

Revenue from each side of the test

Turn A/B test results into revenue so the marketing decision follows from money rather than from percentages.

Using the same two Saturday test, compute the revenue each rack produced and state which arrangement the business should adopt. Tagged bikes sold at ninety five dollars and untagged bikes at seventy five dollars.

Tagged bikes sold: 7
Tagged price: $95
Untagged bikes sold: 3
Untagged price: $75

1. Compute revenue from the tagged rack
Seven bikes at ninety five dollars each produce six hundred sixty five dollars.
$7 \times $ 95 = $ 665$
2. Compute revenue from the untagged rack
Three bikes at seventy five dollars each produce two hundred twenty five dollars.
$3 \times $ 75 = $ 225$
3. Find the gap
Six hundred sixty five minus two hundred twenty five is four hundred forty dollars, over the same two Saturdays, from one extra hour of inspection work per bike.
$$ 665 - $ 225 = $ 440$
4. Convert the result into a decision
The higher price did not suppress sales, so the tagged arrangement wins on rate and on revenue at the same time. Tag every bike going forward and keep logging sales so the test never really stops.

Check answer

Answer
$665 tagged against $225 untagged. The tagged rack produced six hundred sixty five dollars against two hundred twenty five dollars untagged, a gap of four hundred forty dollars.

Why it matters
Rates tell you what customers did; revenue tells you what it was worth. A research finding is only finished when it has been converted into the decision it implies, which here is a rule about every future rebuild rather than a fact about two Saturdays.

Key Terms

Practice Questions

6 questions. Nothing here is recorded or scored.

Question 12.3.A.1, 2.3.C.2
Two Bucket Detail
Tomas Herrera and his sister Bree run Two Bucket Detail, a two-person mobile crew that details cars in customers' driveways. Their menu lists eight separate services, and bookings stall while callers ask what to pick, so the crew plans to replace the list with one flat 100-dollar bundle. They draft two candidates: Inside Out, centered on interior vacuuming, seat cleaning, and windows, and Showroom, centered on exterior wash, wax, and wheels. Tomas writes down what he suspects: customers booking a family vehicle will choose Inside Out over Showroom. For four weekends, every booking inquiry is quoted one of the two bundles, assigned at random, and the crew logs which version converts more inquiries into bookings. Afterward, Bree calls ten recent customers, five from each group, asking open-ended questions about how they decided, with follow-ups whenever an answer surprises her.
The sentence Tomas writes down is best identified as which of the following?
- A.A finding from primary-source research, because it comes from the crew's own booking records.
- B.A market segment, because it names the customers who book a family vehicle for detailing.
- C.A qualitative observation, because Tomas recorded the claim in words rather than in numbers.
- D.A business hypothesis, because it states a testable assumption about customer preferences.
Check answerHide answer
Answer: D
A.
The tell is timing. A hypothesis goes into research, and a finding comes out. Tomas holds a hunch and zero results.
B.
Borrows an earlier topic's vocabulary. A segment is a group of customers sharing traits, and this sentence is a testable claim about what that group prefers.
C.
Being written in words is not what makes data qualitative, and the sentence is not data at all. It is the untested assumption the research is designed to check; the qualitative material arrives later, in Bree's interview answers.
D.
Correct. A business hypothesis is an assumption about a customer, a product, or a market, written down so the business can check whether it holds before committing to a course of action. Testing it is what market research is for: gathering detailed information about markets, products, and customer behavior to steer a marketing decision. Two Bucket is an established business doing exactly what established businesses use research for: reading changing customer wants, finding an opening to innovate, and holding or growing its share.
Question 22.3.C.1, 2.3.C.7
The four-weekend bundle comparison is best described as which of the following?
- A.A focus group, because two bundle options are put in front of customers for their reactions.
- B.An A/B test, because two bundle versions are randomly assigned and real bookings are measured.
- C.An observation, because the crew records customer decisions inside a natural environment.
- D.A survey, because each caller answers a question about which bundle the menu should list.
Check answerHide answer
Answer: B
A.
Swaps behavior for opinions. A focus group gathers a small group to discuss reactions in words, and these callers discussed nothing; they booked or they did not.
B.
Correct. Primary research runs on surveys, interviews, focus groups, observations, experiments, and A/B tests, and an A/B test is a type of experiment: put two workable versions in front of real customers, split them at random, and measure what each group actually does. Booking or declining is behavior, collected in a real-world setting where no caller knows a test is running.
C.
The tell is intervention. The crew changed what each caller heard, and behavior measured after a deliberate change is an experiment. An observation watches a natural environment without touching it.
D.
A survey asks a set of questions and collects the answers as data. These callers were quoted one bundle rather than questioned about the menu, and the data is what they did next, which makes the method an experiment on behavior.
Question 32.3.A.2.ii, 2.3.C.5
Which of the following best explains why Bree used one-on-one interviews for the follow-up calls?
- A.Interviews collect in-depth qualitative data, and follow-ups surface the why.
- B.Interviews produce the large quantitative sample needed to represent the whole customer population.
- C.Interviews record what customers do inside the natural environment where they usually buy.
- D.Interviews remove bias from the findings, because customers answer in their own words.
Check answerHide answer
Answer: A
A.
Correct. Match the method to the data still missing. The A/B test delivered quantitative data, numbers that answer how many, how much, and how often: it counts bookings per bundle and explains none of them. The calls collect qualitative data, descriptions in words that answer why and how, and interviews and focus groups exist for that: deep answers from a few people, ideally highly engaged customers, with follow-up questions pulling out the full explanation.
B.
Describes a survey, the tool for a large amount of quantitative data reflecting a whole population. Ten calls represent nobody and were never meant to.
C.
Recording behavior in a natural environment is observation, and a phone call is the opposite of it: Bree interrupts the customer's day to ask questions. The calls collect explanations in words after the behavior already happened.
D.
Overstates what a method can do. Skew is reduced by neutral question wording and by a sample large enough and representative enough to mirror the population studied, and an interviewer's follow-up questions can introduce bias as easily as remove it.
Question 42.3.B.2, 2.3.B.3
Which of the following best explains why a business assessing a new market opportunity typically begins with secondary-source research?
- A.Secondary sources already record the buying preferences of this business's own existing customers.
- B.Secondary data is more current and more accurate than the data a business gathers for itself.
- C.Secondary sources already hold market size, trends, segments, and rival profiles at low cost.
- D.Secondary research is the standard way to test a hypothesis about specific customer preferences.
Check answerHide answer
Answer: C
A.
Reverses the division of labor. A business's own customers' preferences live in primary research, because no government database has interviewed them.
B.
Overstates the tradeoff. Published data can be years old and aggregated across markets the business does not serve, and its real advantages are cost and coverage rather than freshness or precision.
C.
Correct. Secondary-source research pulls existing quantitative and qualitative information from external sources: government publications and databases, commercial reports, academic studies. Those sources already measure market size in dollars and in total customers, track trends, map segments, and profile rival businesses, and reading what someone else gathered costs a fraction of fielding your own study, so businesses size up the landscape secondhand and save primary research for questions only their own customers can answer.
D.
Backward on both ends. A hypothesis about customer preferences is usually tested with primary research, because the answer lives with the customers themselves, and secondary sources describe the wider market that others have already measured.
Question 52.3.A.3, 2.3.C.4
Fernline Nursery
Fernline Nursery, a garden-plant retailer, surveyed customers with neutrally worded questions about a proposed home-delivery service. The table shows the results.
Survey result | Customers Customers surveyed | 200 Would use home delivery | 120 Would not use home delivery | 80 Of the 120 interested customers, willing to pay a $10 delivery fee | 30 Which conclusion is supported by the table?
- A.25 percent of the 200 surveyed customers said they would pay the nursery's 10-dollar delivery fee.
- B.A majority would use delivery, and 15 percent of those surveyed would also pay the fee.
- C.The delivery service will be profitable, because most of the surveyed customers want the service.
- D.The 80 customers who declined would use home delivery if the nursery dropped the 10-dollar fee.
Check answerHide answer
Answer: B
A.
Performs a correct division on the wrong base. 30 over 120 is 25 percent, the share of the interested customers. This claim uses the 200 respondents as its base, and 30 out of 200 is 15 percent.
B.
Correct. 120 of 200 is 60 percent, a majority, so the first half holds. 30 of 200 is 15 percent, so the second half holds. A supported conclusion stays inside what the 200 respondents actually said, and both halves do.
C.
Leaps from wanting to profiting. Interest is evidence of desirability, a product fitting a real customer problem. Viability, the potential to earn a profit, needs revenue weighed against delivery costs the table never shows, and feasibility, whether the nursery has the vans, staff, and hours to run routes, is a third test the table also skips.
D.
The survey never asked the 80 decliners about a free version, so any claim about what they would do without the fee is invented. Their reasons could be the fee, or timing, or simply preferring to browse plants in person; the table cannot say.
Question 62.3.D.2, 2.3.D.4
A pottery studio earns revenue from classes, finished pieces, and kiln-firing services. The owner wants one chart showing how total revenue changed across three years and how much each line contributed to each year's total. Which of the following best identifies the visualization to choose?
- A.A pie chart showing what share of one year's revenue each of the three lines contributed.
- B.A line graph plotting total revenue for each of the three years the studio has operated.
- C.A bar chart comparing revenue from the three lines during the studio's most recent year.
- D.A stacked bar chart with one bar per year, each bar divided into segments for the three lines.
Check answerHide answer
Answer: D
A.
Carries the mix half and drops the trend. A pie chart shows a part-to-whole split at one moment, so three years of change would take three separate pies with no clean comparison of totals.
B.
Carries the trend half and drops the mix. A line graph tracks one value over time, like customers per year, and shows nothing inside each point.
C.
A plain bar chart comparing the three lines in one year answers a different question: it shows the most recent mix and drops the three-year story entirely. The owner asked for change across years and composition, and a single year's bars deliver neither.
D.
Correct. Visualizations turn findings into patterns, trends, and insights stakeholders can read at a glance, and the rule is to choose the tool that shows the exact insight you need to communicate. This request carries two layers: totals compared across years, and the mix inside each total. Stacking delivers both: each bar's height is that year's total revenue, and its segments show what classes, pieces, and firing contributed.

In a class? These questions are not recorded.

Take the same questions as a scored quiz and your teacher will see that you have finished this section.

Take the scored quiz →

8 Common Mistakes on 2.3

The wrong moves students actually make on these questions, why each one is wrong, and what to do instead. Part of the practice tier.

See what is included

CED Codes

2.3.A.1 · 2.3.A.2 · 2.3.A.2.i · 2.3.A.2.ii · 2.3.A.3 · 2.3.A.3.i · 2.3.A.3.ii · 2.3.A.3.iii · 2.3.A.4 · 2.3.B.1 · 2.3.B.2 · 2.3.B.3 · 2.3.C.1 · 2.3.C.2 · 2.3.C.3 · 2.3.C.4 · 2.3.C.5 · 2.3.C.6 · 2.3.C.7 · 2.3.C.8 · 2.3.C.9 · 2.3.D.1 · 2.3.D.2 · 2.3.D.3 · 2.3.D.4 · 2.3.D.5 · 2.3.D.6