Quantitative and Qualitative Variables: A Guide for SMEs

A coffee shop owner in Puebla usually has more data than they think. Every purchase leaves a trail. The time of consumption, the ticket, the chosen beverage, whether they paid with a coupon, if they returned that week, or if they preferred a branch or delivery. The problem is not the lack of information. The problem is that this information arrives mixed up and ends up stored in reports that no one uses to sell more.

This scenario is repeated in thousands of physical businesses. In Mexico, 68% of retail SMEs in regions like Nuevo León and Jalisco report difficulties in segmenting customers by mixed variables, which reduces the ROI of their campaigns by 42%, and furthermore, the use of AI agents in CRM grew by 35% in Monterrey since 2025, allowing hybrid analyses that raise the average ticket by 18% in businesses like car washes and bakeries, according to the data cited in Matemóvil.

The difference between a business that sends promotions to "everyone" and one that sells better is not in having more customers. It lies in understanding what type of data is in front of them and what decision they can make with it. That is where quantitative and qualitative variables come in.

When a business learns to separate what it can measure from what it can classify, it stops seeing a database and starts seeing opportunities. A customer with a high ticket but low frequency calls for one action. A customer who always buys a frappé and visits on Fridays calls for another. A customer who stopped coming days ago calls for a third.

For an SME, that is not desktop statistics. It is commercial operation. It is deciding what promotion to launch, whom to send it to, and how to measure if it worked. This logic is the basis for data-driven decision-making for physical businesses.

Table of Contents

• Introduction: From Data to Profitable Decisions

  • The data that actually changes decisions

  • Accumulating data is not what is profitable

• What Quantitative and Qualitative Variables Are

  • The difference that actually matters in business

  • Practical table for classifying data

  • Two quick signals to avoid confusion

• The Measurement Scales You Must Know

  • Nominal and ordinal

  • Interval and ratio

  • How to use the scale without getting complicated

• Practical Examples for SMEs in Mexico

  • Coffee shop in Puebla

  • Car wash in State of Mexico

  • Gas station in Yucatán

  • What works best in day-to-day operations

• How to Transform Variables for Advanced Analysis

  • When it is convenient to convert categories into numbers

  • Two useful forms of encoding

  • What you should look out for

• Common Errors When Analyzing Customer Data

  • Errors that cost money

  • The practical fix

• Conclusion: Convert Your Data into Sustainable Growth

  • Three steps to start today

Introduction: From Data to Profitable Decisions

A neighborhood coffee shop rarely fails due to lack of effort. It fails when it makes decisions using intuition where it could already be using evidence. If sales are lower on a Tuesday, a general promo is launched. If visits drop, a message is sent to the entire database. If the cost of supplies rises, they try to compensate by raising prices without understanding who can actually handle that adjustment and who cannot.

That pattern wears you down. It also confuses you. The owner sees sales, tickets, and repeat customers, but does not identify why some return and others disappear. Nor do they see which group responds best to a reward, which branch attracts a certain profile, or which product drives recurring consumption.

Rule of thumb: if a business does not distinguish between numerical data and categorical data, it ends up running massive campaigns for problems that are actually very specific.

Quantitative and qualitative variables help bring order to that chaos. They are not college theory. They are a simple way to separate two key questions: How much? and What kind?. With these two questions answered well, an SME can design promotions with more commercial logic.

The data that actually changes decisions

In a coffee shop, "average ticket" is a quantitative variable. "Preferred beverage" is a qualitative one. "Number of visits in the month" is also quantitative. "Branch where they purchase" is qualitative. In practice, this difference decides whether it is better to compare amounts, count recurrence, or segment by tastes.

When this data is mixed up without order, the business misinterprets its customers. A customer may seem of little value because they come a few times, but if each visit leaves high consumption, they deserve a different strategy than someone who comes often and buys the bare minimum.

Accumulating data is not what is profitable

Accumulating records does not generate profit on its own. What is profitable is turning every record into an action. Therefore, daily operations improve when the business identifies three things:

• What variable describes money: ticket, amount, frequency, points.

• What variable describes preferences: product category, branch, type of reward.

• What combination helps sell more: for example, cold beverage customers with a high ticket or breakfast customers with low recurrence.

An SME that masters this logic stops "sending promotions" and starts managing demand, loyalty, and margin.

What Quantitative and Qualitative Variables Are

Quantitative and qualitative variables are two ways of reading business information. The first measures quantities. The second classifies characteristics. If this division is well understood, the owner stops seeing cells in Excel and starts detecting purchasing patterns.

The difference that actually matters in business

A quantitative variable responds with a number that can be counted or measured. In a coffee shop, it would be the average ticket, accumulated points, purchase frequency, or the number of drinks per visit. In a car wash, the number of services per month. In a gas station, the liters or accumulated visits.

A qualitative variable responds with a category. It does not say how much. It says what kind of thing it is. In a physical SME, that could be the branch visited, the favorite product, the preferred contact channel, or the loyalty level when using a scale like low, medium, and high.

The common mistake is believing that only numerical data is useful. That is not true. Qualitative data usually explains the purchasing context. Knowing that a customer spent more helps. Knowing in which category they spend, what type of reward they prefer, or at which branch they consume helps much more to take action.

A business does not need to become an expert in statistics. It needs to classify each piece of data well before using it.

Practical table for classifying data

Two quick signals to avoid confusion

If the data allows arithmetic operations with commercial meaning, it is almost always quantitative. If the data only serves to group or identify profiles, it is almost always qualitative.

It works to think like this:

• Quantitative: "This customer came 6 times."

• Qualitative: "This customer prefers cold drinks."

• Quantitative: "Their ticket is above average."

• Qualitative: "They buy more at the downtown branch."

When the business makes this separation from the beginning, reports stop being ambiguous. It also improves the sales team's analysis, because every campaign starts from a clearer segmentation.

The Measurement Scales You Must Know

Not all business variables are read the same way. Two data points can be qualitative and still require different treatments. The same goes for quantitative ones. The measurement scale defines what comparison makes sense and what analysis should be used.

Nominal and ordinal

The nominal scale groups categories without order. "Puebla Centro Branch," "Cholula Branch," or "payment method" fall here. No category is greater than another. They are just different. This is useful for comparing distribution, share, or preference.

The ordinal scale also classifies, but it does have a hierarchy. "New customer," "frequent customer," and "loyal customer" is a good example. Also "low, medium, high" in loyalty level. Here order does matter, even if the distance between levels is not exact.

For daily operations, that difference changes a lot:

• Nominal is used to see where the customer buys or what product they choose.

• Ordinal is used to prioritize actions. A customer with high loyalty is not treated the same as one at a low level.

Interval and ratio

The interval and ratio scales belong to the quantitative world. Both use numbers, but they do not mean the same thing.

In interval, there is a difference between values, but zero does not always imply actual absence. In ratio, an absolute zero does exist, and that makes commercial comparisons much more useful. Sales, visits, points, and purchase amount fall under ratio. If a customer has zero visits in the month, that does mean an absence of activity.

Operational criterion: if the data allows you to say "double" or "half" with real meaning, you are usually dealing with a ratio variable.

This distinction seems fine, but it prevents misinterpretations. A coffee shop owner needs to know which reports allow comparing purchase intensity and which only describe order or category.

How to use the scale without getting complicated

The most practical way to use this in an SME is to ask three questions:

1. Is it a category or a number?
   If it is a category, you are in nominal or ordinal.

2. If it is a category, does it have a hierarchy?
   If the answer is yes, it is ordinal.

3. If it is a number, does zero mean actual absence?
   If yes, it is ratio.

With this logic, the business avoids choosing the wrong charts, useless crossovers, and poor conclusions. The correct data must not only exist. It must be on the right scale so that the analysis is actually useful.

Practical Examples for SMEs in Mexico

Theory is of little use if it does not land on the counter, cash register, and repeat purchases. In Mexican physical retail, analyzing quantitative variables like average ticket and purchase frequency can raise ROI by 25-40%, and segmenting customers with an average ticket over $500 MXN can increase recurrence by 32%. Furthermore, 68% of retail chains fail in retention by not quantifying metrics like LTV, which averages $2,800 MXN per customer per year, according to the shared reference on applied quantitative analysis.

Coffee shop in Puebla

A coffee shop in Puebla usually has two strong moments. The office morning and the afternoon sweet craving. If the business registers preferred beverage as a qualitative variable and average ticket as a quantitative variable, it can build very useful segments.

For example, one group buys cold drinks and almost always adds sweet bread. Another buys black coffee alone and is just passing through. The first group should not receive the same promotion as the second. A combo works better for the first. For the second, a recurring visit dynamic is best.

A simple way to bring this down to operations is to review sales reports to detect consumption patterns by segment. The goal is not to look at numbers out of curiosity. It is to decide which offer increases frequency and which offer just gives away margin.

Car wash in State of Mexico

In a car wash, variables are usually more obvious but less exploited. Accumulated points is quantitative. Type of service is qualitative. The mix between both allows detecting who is close to a reward and in which service it is convenient to activate it.

A customer who accumulates points quickly and always buys a premium wash does not need a generic discount. They benefit from a reward aligned with their pattern, such as a benefit in detailing or waxing. In contrast, someone who buys only the basic service and lets a long time pass between visits requires a different intervention.

If the reward does not respect the purchasing habit, the customer perceives it as irrelevant, even if the discount is good.

Gas station in Yucatán

In a gas station, a very useful variable is days since the last fuel fill-up. It is quantitative. It allows detecting the risk of churn without waiting for the customer to disappear completely. If it is also crossed with a qualitative variable like preferred reward method, the message becomes more accurate.

Not everyone reacts the same way. One driver may respond better to points. Another to a direct coupon. Another to an incentive per visit. The same business, with the same customer base, gets different results depending on whether it mixes both types of variables well.

What works best in day-to-day operations

In food, services, and high-frequency consumption SMEs, these combinations usually provide clarity:

• Average ticket + favorite category to design combos.

• Purchase frequency + branch visited to detect drops by location.

• Days since last purchase + reward type for reactivation campaigns.

• Accumulated points + loyalty level to define tiered benefits.

What does not work is treating the entire base the same. In a coffee shop, that translates to wasted promotions. In a gas station, to poorly allocated points. In a car wash, to loyal customers who do not feel recognized.

How to Transform Variables for Advanced Analysis

There comes a point where classifying well is no longer enough. If the business wants to predict churn, attribute sales to campaigns, or find less obvious patterns, it needs to convert certain categories into formats that analysis can process better.

The good news is that this is not about making the whole business technical. It is about translating qualitative variables into a useful structure. According to the reference on encoding and analyzing qualitative variables, encoding qualitative variables into dummy variables for regression analysis can raise sales attribution accuracy by up to 76%. Furthermore, ordinal variables like loyalty level show a Spearman correlation of 0.62 with retention, and acting on them with automated campaigns can increase purchase frequency by 19%.

When it is convenient to convert categories into numbers

A category like "branch" cannot be averaged. But it can be encoded to answer more useful questions. For example, which branch is associated most with a certain type of customer or which one has a better response to a campaign.

The same goes for "loyalty level." Low, medium, and high are ordered categories. If they are converted with criteria into a simple scale, the business can better prioritize follow-ups, benefits, and reactivation.

This principle also appears outside of CRM. In search and filtering tasks, understanding structured logic helps to build cleaner rules. That is why it is worth reviewing the foundamentals of boolean sourcing, because they show how to transform loose criteria into clear and operable conditions. This mindset is very useful when segmenting customers.

Two useful forms of encoding

There are two practical paths that an SME can understand without getting tangled up.

Dummy variables for nominal categories

If a business has several branches, it is not convenient to assign them arbitrary numbers like 1, 2, or 3 and pretend that means a hierarchy. The correct way is to create separate columns like "Puebla Centro Branch = 1 or 0" and "Cholula Branch = 1 or 0."

This allows measuring participation by category without inventing an order that does not exist.

Ordered labels for ordinal variables

When there is a hierarchy, such as in loyalty level, a simple scale can be assigned. Low = 1, medium = 2, high = 3. Here the number does not represent money. It represents order.

This helps the business identify more consistent behavior patterns and build customer segmentation based on habits and commercial value.

Converting a category into a number does not change the customer's reality. It only makes a signal that was previously scattered analyzable.

What you should look out for

Not every conversion improves the analysis. If a category is encoded poorly, the business ends up creating false patterns. The rule is simple:

• If there is no real order, use separate variables.

• If there is order, use a short and clear scale.

• If the team does not understand what the number means, the model will not help in operations.

Good encoding does not just serve to make a report look prettier. It serves to decide better whom to activate, with what offer, and at what time.

Common Errors When Analyzing Customer Data

Many businesses do not fail due to lack of data. They fail because they misinterpret what they already have. In Mexico, 55% of marketing managers in SMEs make errors by treating qualitative variables as numerical ones, which biases analyses and can underestimate recurrence by 30%. In this same context, discrete quantitative variables like the number of visits have proven to be more effective than continuous ones in predicting retention, according to the included reference on frequent errors in the treatment of variables.

Errors that cost money

The first mistake is averaging where there are no useful averages. "Branch average," "payment method average," or "zip code average" do not contribute anything. They are categories, not amounts or frequencies.

The second mistake is treating an identifier as if it were a metric. The customer number or the ticket ID are not useful for measuring behavior. They only identify records. If the team includes them in analyses as if they were business variables, it pollutes the result.

The third is choosing visualizations that confuse. A line chart to show reward types or payment methods usually suggests a continuity that does not exist. A comparison by category works better there.

The practical fix

Three adjustments clean up a large part of the problem:

• Separate identifiers from actual variables. Customer ID is not behavior.

• Do not force categories into arbitrary numbers. If there is no hierarchy, do not invent one.

• Choose the variable that actually predicts action. In retention, the number of visits usually says more than a misinterpreted continuous data point.

An incorrect analysis does not just give a bad reading. It also drives the wrong campaigns, unnecessary discounts, and poorly focused follow-ups.

A coffee shop owner in Puebla does not need a more complex dashboard. They need to avoid decisions based on irrelevant averages, poorly encoded categories, and reports that look technical but do not help sell.

Conclusion: Convert Your Data into Sustainable Growth

Quantitative and qualitative variables are not an isolated academic topic. They are the foundation for turning daily consumption into profitable decisions. When an SME understands which data measures value and which describes preferences, it starts to segment better, launch more precise campaigns, and use its reports with commercial intent.

The difference between growing with order or living putting out fires usually starts there. Not with a grand strategy. With something simpler. Knowing if a data point answers "how much" or answers "what kind." After that, everything improves: the offer, the reward, the recurrence, and the customer reading.

For a coffee shop, a car wash, or a gas station, the starting plan can be very concrete.

Three steps to start today

1. Audit the current database
   Separate numerical data from categorical data. Ticket, visits, and points on one side. Branch, favorite product, and reward type on the other.

2. Create a first mixed segment
   Combine a quantitative variable and a qualitative one. For example, high-ticket customers who prefer cold drinks, or low-frequency customers at a certain branch.

3. Launch an action and measure the response
   Send a specific campaign, not a general one. Then review the repurchase, ticket, and recurrence of the targeted segment.

This order makes information actionable. It also avoids one of the most common problems in physical SMEs: having data without turning it into utility.

If a physical business in Mexico wants to move from loose data to campaigns that actually generate recurrence, Swirvle helps centralize customers, segment by purchasing habits, automate loyalty actions, and measure which campaigns actually drive sales. For coffee shops, car washes, gas stations, and chains with several branches, this difference translates into clearer operations and more sustainable growth.