What Is the 4P Marketing Mix? Applying It to a Research Model
What Is the 4P Marketing Mix Model
The 4P Marketing Mix is a framework for analysing marketing activities through four groups of variables: Product, Price, Place and Promotion. When you open a topic on purchase intention, satisfaction, loyalty or brand choice, these four groups are often used to describe what a business offers and how customers access the product.
In quantitative research, you should not put 4P into your questionnaire as four broad labels. Each P needs to become a latent variable and then be measured through multiple items. For example, Product may be measured through quality, design and fit with customer needs. Price may relate to the price level, perceived fairness and the value customers feel they receive.
The 4P framework is commonly associated with E. Jerome McCarthy in marketing literature. When you write your thesis, however, check the exact source required by your lecturer instead of adding a publication year that does not appear in your reference list. The 4P framework is also different from the AIDA model. AIDA describes stages in a customer's response process, while 4P describes groups of marketing tools.
Components of the Model
The four components of 4P can be independent variables in a model explaining customer behaviour. You must choose the dependent variable according to the specific research question, such as purchase intention or satisfaction. Do not assume that all four Ps have the same effect in every industry, because their relevance depends on the product, market and survey context.
| Component | What needs to be measured | Examples of items that could be developed |
|---|---|---|
| Product | Product or service characteristics and perceived value | The product meets needs, quality is stable, design is suitable |
| Price | Customer perceptions of price and the value received | The price is reasonable, proportional to quality, and acceptable compared with other options |
| Place | Product accessibility and the convenience of the distribution channel | Easy to find and purchase, convenient location, suitable delivery |
| Promotion | Communication activities and purchase incentives | Information is easy to understand, promotions are suitable, communication content is trustworthy |
Product is not limited to physical quality. For a service or digital platform, Product may include features, user experience, support and the extent to which it solves a problem. Price is also more than the listed number, because respondents may evaluate price based on benefits, switching costs or competitors' prices.
Place needs to be adapted to the context. For a retail store, you may ask about location and ease of purchase. For e-commerce, this variable may focus on the sales channel, delivery speed and convenience when receiving the product. Promotion should remain separate from Product, because effective advertising does not mean that the product itself is high quality.
The Original Model and Common Extensions
The 4P model is most suitable when your topic studies a product or marketing activities from the customer's perspective. When studying services, many sources extend 4P to 7P by adding People, Process and Physical Evidence. You can see the distinction in What Is the 7P Marketing Mix, but do not add these three components simply because the 7P model is popular.
Another extension combines 4P with a model that explains behaviour. For example, 4P can be the group of independent variables, Perceived Value can be a mediator, and Purchase Intention can be the dependent variable. If your topic focuses on technology acceptance, you can refer to the AIDA model to distinguish variables describing communication responses from variables measuring behavioural intention. Combining models must follow the research question and the scales, rather than an attempt to place as many variables as possible in the diagram.
You can also use 4P as a way to group variables before building the model. For example, a study of bottled coffee may retain only Product, Price and Promotion if Place is not part of the research problem. In that case, explain why Place was excluded in both the theoretical background and methodology chapters.
The table below contains technical benchmarks that commonly appear when you convert 4P into a quantitative model. These are criteria for assessing data and scales. They are not thresholds for concluding that a real-world marketing strategy will certainly succeed.
| Check | Reference benchmark | Source |
|---|---|---|
| Cronbach's Alpha of the scale | 0.7 or above; an exploratory scale may be considered from 0.6 | (Nunnally, 1978); (Hair et al., 2010) |
| Corrected Item-Total Correlation | 0.3 or above | (Nunnally and Bernstein, 1994) |
| Factor loading in EFA | 0.5 or above | (Hair et al., 2010) |
| KMO when running EFA | 0.5 or above | (Kaiser, 1974) |
| Total variance explained | 50% or above | (Hair et al., 2010) |
| Outer loading in PLS-SEM | 0.7 or above | (Chin, 1998); (Hair et al., 2022) |
| AVE and Composite Reliability | AVE at 0.5, CR at 0.7 | (Fornell and Larcker, 1981) |
| HTMT between two constructs | Below 0.85, or below 0.90 for closely related constructs | (Henseler et al., 2015) |
Scales Commonly Used for Each Construct
4P is not a single scale that you can copy word for word for every topic. You need to find a scale that fits the industry, product, respondents and language used in the study. In the foundational marketing sources, SERVQUAL by Parasuraman et al. focuses on service quality, so you should not describe all of SERVQUAL as a 4P scale. Check the specific origin of each scale before placing it in Chapter 3.
| Construct | Direction for developing the scale | Suitable reference source |
|---|---|---|
| Product | Quality, features, design and fit with customer needs | (Nguyễn Đình Thọ, 2011) for the scale development and validation process |
| Price | Reasonableness of price, perceived value and affordability | (Nguyễn Đình Thọ, 2011) for the process of adapting the scale |
| Place | Accessibility and convenience of the distribution channel | (Nguyễn Đình Thọ, 2011) for the quantitative survey process |
| Promotion | Clarity, attractiveness and trustworthiness of communication | (Nguyễn Đình Thọ, 2011) for the process of developing items |
| Purchase Intention | Likelihood, plan or desire to purchase | You need to choose the original study for the correct context, rather than attaching a source before checking it |
The rows above are development directions, not confirmed translations. When drafting the questionnaire, write each item as a statement that respondents can rate on a Likert scale. The Likert scale was introduced in Likert's work, so you can cite (Likert, 1932) when explaining the use of this scale.
Use the same form of address, rating anchors and number of points within one scale group. If you have a reverse-coded item, mark it separately in the data file and check the coding again before running Cronbach's Alpha. Translating items from English into another language also requires a contextual review, because a grammatically correct sentence can still be difficult for respondents to understand.
Applying the Model to Your Topic
Suppose your topic is: “The effect of the Marketing Mix on domestic cosmetics purchase intention among students in Ho Chi Minh City.” You could build an initial model with four independent variables, Product, Price, Place and Promotion, and Purchase Intention as the dependent variable.
The conceptual model is:
Product, Price, Place, Promotion → Purchase Intention
The hypotheses need to be written as complete sentences:
- H1: Product has a positive effect on students' Purchase Intention toward domestic cosmetics.
- H2: Price has a positive effect on students' Purchase Intention toward domestic cosmetics.
- H3: Place has a positive effect on students' Purchase Intention toward domestic cosmetics.
- H4: Promotion has a positive effect on students' Purchase Intention toward domestic cosmetics.
This is an illustrative model, not the result of an actual study. When using it for your topic, replace the industry, respondents, location and dependent variable in the hypotheses. If you want to study loyalty, replace Purchase Intention with Loyalty and provide a theoretical basis for that relationship.
Next, create a variable coding table. For example, Product can use the codes PRD1, PRD2 and PRD3. Price can use PRI1, PRI2 and PRI3. Place can use PLC1, PLC2 and PLC3, while Promotion can use PRO1, PRO2 and PRO3. The codes must be consistent across the questionnaire, the .sav or .csv file, and the SmartPLS diagram.
When the data are ready, check missing data, straight-lining responses, reverse coding and cases that do not belong to the target population. Only then should you assess scale reliability and validity. If an item is removed, record the item code, reason and rerun result instead of silently changing the results table.
How to write this in your thesis: “Based on the 4P Marketing Mix framework and the context of [industry/product], the study proposes four factors, [Product], [Price], [Place] and [Promotion], that affect [dependent variable] among [survey respondents] in [location]. Hypotheses H1 to H4 are tested using [SPSS/SmartPLS] on data collected from [number] respondents.”
How to Test the Model in SPSS or SmartPLS
SPSS is suitable when you use linear regression, reliability testing and EFA, and need to read the Reliability Statistics, KMO and Bartlett's Test, Rotated Component Matrix and Coefficients tables. The basic process is to run Cronbach's Alpha for each group, run EFA if required by the research design, calculate representative scores for the factors, and then run regression with Purchase Intention as the dependent variable.
SmartPLS is suitable when 4P is modelled through latent variables and you want to assess the measurement model and structural model together. In SmartPLS 4, import the .csv file, draw the constructs, connect the arrows according to the hypotheses, run PLS-SEM Algorithm, and then run Bootstrapping. When reporting PLS-SEM, you can use 5,000 bootstrap subsamples following the guidance of (Hair et al., 2022).
The table below shows the shape of the output. It is not data from a real study.
illustrative output
| Path Coefficients | Original sample (O) | Sample mean (M) | Standard deviation (STDEV) | T statistics | P values |
|---|---|---|---|---|---|
| Product → Purchase Intention | 0.284 | 0.287 | 0.071 | 4.000 | 0.000 |
| Price → Purchase Intention | 0.156 | 0.159 | 0.068 | 2.294 | 0.022 |
| Place → Purchase Intention | 0.091 | 0.094 | 0.064 | 1.422 | 0.155 |
| Promotion → Purchase Intention | 0.331 | 0.328 | 0.075 | 4.413 | 0.000 |
You must not draw conclusions about your own data from the illustrative table. With the real file, check the sign of the path coefficient, p-value, bootstrap confidence interval and VIF. VIF below 5 is a commonly referenced benchmark in PLS-SEM (Hair et al., 2019). The R² of Purchase Intention shows how much of the variance is explained by the variables in the model, while Q² above 0 is used to assess predictive relevance (Stone, 1974).
In SPSS, go to Analyze > Scale > Reliability Analysis to run Cronbach's Alpha. Select the items for each P separately and open Statistics to select Scale if item deleted. Then go to Analyze > Dimension Reduction > Factor to run EFA and check KMO, Bartlett, Total Variance Explained and Rotated Component Matrix. For regression, go to Analyze > Regression > Linear, place Purchase Intention in the Dependent box and the factor scores in Independent(s).
Common Errors When Using This Model
The first error is treating 4P as four individual items. One P is usually a construct made up of multiple items, so asking only one question about Product or Price makes it difficult to assess scale reliability.
The second error is copying a complete set of items from another industry. A Promotion scale for banking may not fit cosmetics or a food delivery application. You need to preserve the construct meaning, adapt the wording to the context and check it again through a pilot survey.
The third error is adding People, Process and Physical Evidence while still calling the model 4P. If you add these three components, call it 7P and explain the reason for the extension. Similarly, RFM theory and the RFM model are used for segmentation or customer value analysis, not as automatic variations of 4P.
The fourth error is keeping an item only because it increases Cronbach's Alpha. A high Alpha is not enough to prove content validity, convergent validity and discriminant validity. You need to combine theory, Corrected Item-Total Correlation, factor loading or outer loading, and the factor structure.
The final error is writing conclusions that make overly strong causal claims. Cross-sectional survey data can show relationships within the model and the extent to which the findings fit the hypotheses. Write “has an effect according to the model results” within the scope of the study. Do not write that one P always makes customers buy more in every situation.
Frequently asked questions
Is 4P Marketing Mix a theory
4P is commonly used as a framework or practical model for classifying marketing tools. In your thesis, place 4P in the theoretical background and explain why each component may be related to the dependent variable in your topic.
Do I have to use all four Ps
No. You can select the components that fit your research question, but you must explain why a variable was excluded. If the topic focuses on the sales channel, Place may be the central component. If the study examines service quality, 7P or a service quality model may be more suitable.
Should I run 4P in SPSS or SmartPLS
Both can be suitable. Choose SPSS when the model is simple and you need EFA and regression on factor scores. Choose SmartPLS when you want to retain latent variables and assess the measurement model and structural model at the same time. Software does not replace the need to choose suitable scales and hypotheses.
Can I use 4P to study purchase intention
Yes, if you can build an argument for the relationship between each P and purchase intention in a specific context. Clearly identify the respondents, product, location and way of measuring Purchase Intention. Then test each hypothesis instead of drawing a conclusion from personal impressions.
What should I do if Cronbach's Alpha for one P is low
Check Corrected Item-Total Correlation, the item content, coding and missing data. Remove an item only when there is both a statistical and substantive reason, then rerun the scale and record each processing round in the thesis. Do not delete several items simply to reach an attractive number.
Open your outline and data file again, finalise the dependent variable, create codes for the four P groups and check each scale in the correct order before running the model. If you need to run the analysis on your own .sav or .csv file, you can use DoThesis M4 data analysis.