Brand positioning optimization method based on big data

Abstract

The invention relates to the technical field of market analysis, in particular to a brand positioning optimization method based on big data, and the method comprises the steps: collecting multi-channel brand performance data, calculating a mean value, extracting a co-occurrence relation between emotion words and labels in comments, judging an emotion and image matching structure, and analyzing competing product label frequency difference and cognitive overlap. And identifying a user portrait and emotion offset, and merging expression gaps to generate an optimization suggestion scheme. According to the method, differentiated collection and unified coding are performed on multi-channel indexes, a brand expression structure capable of cross comparison is constructed, data definition is improved, comment words and brand labels form a co-occurrence matrix, an emotional expression structure under the value dimension is extracted, competitive product analysis locates a cognitive blank on the basis of label difference and coverage degree, and the competitive product analysis accuracy is improved. The emotion deviation value and the label acceptability are introduced as dynamic parameters by user division, and finally a fusion label missing and emotion space is output, expression reconstruction and crowd adaptation are completed, and dynamic linkage and positioning updating of a brand structure are realized.

Description

Technical Field

[0001] This invention relates to the field of market analysis technology, and in particular to a brand positioning optimization method based on big data. Background Technology

[0002] The field of market analysis technology encompasses methods and tools for systematically studying and evaluating market behavior, consumer preferences, and brand performance using data collection, statistical modeling, and trend forecasting. Its core content is based on structured and unstructured data, using quantitative analysis to identify changes in target markets, consumer behavior characteristics, and competitive landscapes to support corporate decisions regarding product positioning, marketing strategies, and resource allocation. Overall, market analysis technology covers various methods such as data mining, behavioral modeling, predictive analytics, semantic recognition, and user profiling. Its development relies on the fusion and processing of multi-source heterogeneous data and the improvement of model inference capabilities.

[0003] Among them, the brand positioning optimization method based on big data refers to acquiring and processing consumer behavior data, public opinion data, and brand communication data from different platforms, and using methods such as calculating brand relevance indicators, user preference characteristics, and market competition parameters to dynamically identify and model the brand's position in a specific market. It also combines classification algorithms to perform brand feature attribution, dimensional redistribution, and target positioning optimization. Generally, it uses multi-dimensional attribute data collection, tag fusion and sorting methods, and brand mapping rule generation mechanisms to complete the reconstruction and updating of brand positioning parameters.

[0004] Existing technologies largely rely on aggregated data and static mapping. Emotional feedback loses its semantic level during structural transformation, making it difficult to interpret emotional expressions. Brand tags lack connection with user expressions, cognitive structures struggle to support emotional threads, reducing tag recognition. Competitive analysis ignores differences in perceptual spatial structures, brand differentiation is blurred, audience segmentation is based on static attributes, behavioral characteristics are not accurately categorized, segmentation fails, and strategy output lacks a linkage mechanism between expression and perception. Tag guidance becomes unfocused and fails to support dynamic positioning reconstruction. Summary of the Invention

[0005] To address the technical problems existing in the prior art, this invention provides a brand positioning optimization method based on big data. The technical solution is as follows:

[0006] A brand positioning optimization method based on big data includes the following steps:

[0007] S1: Collect sales volume, satisfaction rate, mention frequency, and comment keywords from multiple brand channels, categorize feedback content, calculate channel averages, compare and judge brand performance, and generate a brand market performance overview.

[0008] S2: Based on the aforementioned brand market performance overview, extract the co-occurrence frequency of sentiment words and tags in the comments, calculate the sentiment proportion of each value dimension, determine the sentiment and image matching structure, extract core features, and generate a brand sentiment positioning map;

[0009] S3: Based on the aforementioned brand sentiment positioning map, obtain competitor tags and their usage frequency, calculate the tag frequency difference, determine the coverage level, define the overlapping distribution of the brand and competitors in the cognitive structure, and generate a competitive brand positioning difference analysis table.

[0010] S4: Based on the brand sentiment positioning map and the competitive brand positioning difference analysis table, extract user profiles and comment tendencies, calculate cognitive offset values, screen groups with prominent fluctuations, and generate a consumer sentiment preference analysis report;

[0011] S5: Based on the aforementioned brand sentiment positioning map, competitor brand positioning difference analysis table, and consumer sentiment preference analysis report, determine the direction of group label shift, analyze acceptance differences and missing features, screen expression gaps and merge suggestions, and generate a brand positioning optimization plan.

[0012] As a further aspect of the present invention, the brand market performance overview includes sales volume distribution, user satisfaction rating, brand mention frequency, and the proportion of key comment terms; the brand emotional positioning map includes the positive and negative ratio of emotional words, emotional intensity classification, and the matching degree between brand image and emotional expression; the competitive brand positioning difference analysis table includes differences in emotional tag frequency, brand cognition overlap, and emotional coverage intensity; the consumer emotional preference analysis report includes the emotional preference intensity of user groups, the degree of cognition difference between groups, and the amplitude of emotional tendency fluctuation; and the brand positioning optimization scheme includes brand tag strengthening strategies, re-segmentation of target audiences, and supplementary dimensions of emotional expression.

[0013] As a further aspect of the present invention, the steps for obtaining the brand market performance overview are as follows:

[0014] S101: Collect data on brand sales volume, user satisfaction, brand mention frequency, and comment keywords across social media platforms, e-commerce channels, and offline terminals. Group the raw indicators according to their source channels, record sales volume and satisfaction numerically, and collect mention frequency and keywords textually. Establish standard format data entries and generate a multi-channel source data set.

[0015] S102: Call the comment keywords in the multi-channel source data set, classify the keywords according to the three preset dimensions of product feeling, purchase motivation and brand awareness, count the total number of occurrences of the keywords in each channel under each dimension, use the total number as the performance score of the corresponding dimension, and obtain the quantitative value of the channel feedback theme.

[0016] S103: Based on the sales volume and satisfaction in the multi-channel source data set, and combined with the quantitative value of channel feedback themes, calculate the weighted performance score for social platforms, e-commerce, and offline terminals. Then, summarize the scores of each channel to obtain the global performance benchmark value, compare the deviation of each channel score from the global performance benchmark value, and establish an overview of the brand's market performance.

[0017] As a further aspect of the present invention, the steps for obtaining the brand emotional positioning map are as follows:

[0018] S201: Based on the aforementioned brand market performance overview, extract sentiment expression words and brand tags from the included consumer reviews, count the frequency of co-occurrence of each pair of sentiment expression words and brand tags in the same review unit, use the co-occurrence frequency as the association strength value, construct a numerical table with brand tags as rows and sentiment words as columns, and obtain the sentiment tag co-occurrence matrix;

[0019] S202: Call the aforementioned sentiment tag co-occurrence matrix, divide sentiment words into value perception dimensions such as function and experience, calculate the total frequency of all sentiment words under each dimension, calculate the proportion of positive sentiment words, monitor the change of this proportion over time, and generate the sentiment proportion and change magnitude of the value dimension.

[0020] S203: Based on the emotional tag co-occurrence matrix and the emotional proportion and change range of the value dimension, filter the dimensions whose emotional proportion exceeds the mean and whose change range is positive, identify the emotional words and brand tag combinations with the highest co-occurrence frequency under the dimension, determine them as core emotional features, and use their intensity and dimension as coordinate axis values ​​to establish a brand emotional positioning map.

[0021] As a further aspect of the present invention, the steps for obtaining the competitive brand positioning difference analysis table are as follows:

[0022] S301: Combining the dimensional framework of the brand emotional positioning map, obtain the emotional tags, product characteristic words and user usage frequency data of the specified competitors, perform word segmentation and frequency statistics on the collected text information, and associate it with the usage frequency to establish a competitor data list containing tags, words and frequencies, and obtain the competitor perception feature set.

[0023] S302: Call the tag frequency in the brand sentiment positioning map and match the frequency of similar tags in the competitor perception feature set. For each common tag, perform the operation of subtracting the competitor frequency from the target brand frequency. Use the operation result as the difference in the volume of the two brands at a specific cognitive point to obtain the brand tag volume difference.

[0024] S303: Based on the brand label volume difference, set a zero-point offset benchmark value, define labels with differences exceeding the benchmark value as advantageous areas, define labels with differences below the negative benchmark value as disadvantageous areas, and classify the rest as overlapping areas. Summarize the affiliation, original frequency and difference data of each label to generate a competitive brand positioning difference analysis table.

[0025] As a further solution of the present invention, S401: Based on the brand emotional positioning map and the competitive brand positioning difference analysis table, extract the gender, age, consumption frequency and category ratio data from the associated user profiles, divide the users into independent groups according to these profile dimensions, count the total number of times each group mentions the brand tags in the map, and obtain the group tag interaction frequency.

[0026] S402: Call the group tag interaction frequency, and use the full mention frequency of each tag in the brand sentiment positioning map as a reference benchmark to calculate the ratio between the mention frequency of each tag for each group and the reference benchmark. Compare this ratio with the benchmark value 1 to obtain a quantitative value that represents the difference in group cognition and obtain the group cognition offset value.

[0027] S403: For the aforementioned group cognitive offset value, set a deviation fluctuation threshold, filter out user groups whose values ​​exceed the threshold, integrate their profile information, comment tendencies, and positioning overlap in the competitive brand positioning difference analysis table, present these filtered and integrated information items, and generate a consumer sentiment preference analysis report.

[0028] As a further aspect of the present invention, the step of obtaining the brand positioning optimization scheme is as follows:

[0029] S501: Based on the group cognitive bias value in the consumer sentiment preference analysis report, determine whether the bias value of each group to each brand label is positive or negative, and integrate the sign and absolute value of the value to quantify it into an indicator that represents the direction and intensity of the group cognitive tendency, and obtain the group cognitive bias degree.

[0030] S502: Call the aforementioned group cognitive bias and compare it with the disadvantage and overlap area labels in the competitive brand positioning difference analysis table. Filter out the labels with positive group cognitive bias and non-advantageous brand voice, and use them as cognitive features to be strengthened to obtain a list of brand mindshare gaps.

[0031] S503: Based on the brand mindshare gap list, group the tags in the list into core dimensions to be strengthened, and match user groups that hold a positive group cognitive bias towards these dimensions from the consumer sentiment preference analysis report. Strategically match the strengthened dimensions with the target audience to establish a brand positioning optimization plan.

[0032] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0033] By collecting and uniformly encoding multi-channel metrics, a cross-comparable brand performance structure is constructed to improve data clarity. Comment words and brand tags form a co-occurrence matrix, extracting the emotional expression structure under the value dimension. Competitive analysis identifies cognitive gaps based on tag differences and coverage. User segmentation introduces emotional offset values ​​and tag acceptance as dynamic parameters. Finally, the output integrates missing tags and emotional gaps, completing expression reconstruction and audience adaptation, and realizing dynamic linkage and positioning updates of the brand structure. Attached Figure Description

[0034] Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0035] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0036] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0037] In the embodiments of this invention, the terms "image" and "picture" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent. Similarly, the terms "of," "corresponding (relevant)," and "corresponding" may sometimes be used interchangeably. It should be noted that, without emphasizing the distinction between them, their intended meanings are consistent.

[0038] In this embodiment of the invention, sometimes a subscript such as W1 may be written in a non-subscript form such as W1. When the difference is not emphasized, the meaning they express is the same.

[0039] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0040] Please see Figure 1 This invention provides a technical solution: a brand positioning optimization method based on big data, comprising the following steps:

[0041] S1: Collect data on brand sales volume, user satisfaction, brand mention frequency, and comment keywords on social media platforms, e-commerce channels, and offline terminals. Categorize product feelings, purchase motivations, and brand awareness content in the feedback. Calculate and compare the average brand performance across channels to determine the brand's focus and relative performance, and generate a brand market performance overview.

[0042] S2: Based on the brand's market performance overview, extract the co-occurrence frequency of emotional expression words and brand tags in consumer reviews, calculate the proportion and trend of emotional words under the value perception dimension, determine the matching structure between emotional words and brand image, extract the core emotional characteristics and expression focus in consumer perception, and generate a brand emotional positioning map;

[0043] S3: Combine brand sentiment positioning map to obtain the sentiment tags, product feature words and user usage frequency of competitors, calculate the difference value of the target brand and competitors in tag frequency, judge the coverage of competitors in relevant dimensions, define the overlapping distribution of brand and competitors in cognitive structure, and generate a competitive brand positioning difference analysis table.

[0044] S4: Based on the brand sentiment positioning map and the competitive brand positioning difference analysis table, extract gender, age, consumption frequency and category ratio from the user profile, identify the acceptance strength and comment tendency of each group to the brand label, calculate the cognitive bias value and filter the groups with prominent fluctuations, and generate a consumer sentiment preference analysis report.

[0045] S5: Based on the brand sentiment positioning map, the competitive brand positioning difference analysis table, and the consumer sentiment preference analysis report, determine the direction of brand label shift among different groups, analyze the differences and missing features in acceptance, filter out dimensions with insufficient expression coverage, merge reinforcement items and recommended audiences, and generate a brand positioning optimization plan.

[0046] The brand market performance overview includes sales volume distribution, user satisfaction rating, brand mention frequency, and the proportion of key comment terms. The brand sentiment positioning map includes the positive and negative ratio of sentiment words, sentiment intensity classification, and the matching degree between brand image and sentiment expression. The competitor brand positioning difference analysis table includes differences in sentiment tag frequency, brand awareness overlap, and sentiment coverage intensity. The consumer sentiment preference analysis report includes the sentiment preference intensity of user groups, the degree of cognitive differences between groups, and the fluctuation range of sentiment tendencies. The brand positioning optimization plan includes brand tag strengthening strategies, target audience re-segmentation, and supplementary dimensions of sentiment expression.

[0047] The steps to obtain an overview of a brand's market performance are as follows:

[0048] S101: Collect data on brand sales volume, user satisfaction, brand mention frequency, and comment keywords across social media platforms, e-commerce channels, and offline terminals. Group the raw indicators according to their source channels, record sales volume and satisfaction numerically, and collect mention frequency and keywords textually. Establish standard format data entries and generate a multi-channel source data set.

[0049] This study collects data on brand sales volume, user satisfaction, brand mention frequency, and comment keywords across social media platforms, e-commerce channels, and offline terminals. First, by calling the social media platform API and setting the time period to the most recent 30 days, it retrieves 50,000 mentions of the "Shixu Technology" brand and simultaneously captures 42,000 related comment keyword text data. Then, by accessing the e-commerce platform database, it obtains 15,000 sales of the "Shixu Technology" CT-X1 smartwatch model during the same period, and 11,500 satisfaction rating records generated after user payment, with a rating of 1-. By differentially collecting and uniformly encoding indicators from multiple channels, a cross-comparable brand performance structure is constructed, improving data clarity. Comment words and brand tags form a co-occurrence matrix, extracting the emotional expression structure under the value dimension. Competitive product analysis identifies cognitive gaps based on tag differences and coverage, and user segmentation introduces emotional offset values ​​and tag acceptance as dynamic parameters. The final output integrates missing tags and emotional gaps, completes expression reconstruction and audience adaptation, and achieves dynamic linkage and positioning updates of the brand structure on a 10-point scale. At the same time, it collects evaluation text as comment keywords, totaling 12,000 entries. In addition, it digitizes 5,000 sales records from offline terminals through the POS system, 480 user satisfaction rating records collected through after-sales service questionnaires, and 350 handwritten keyword records from customer suggestion boxes. All the data from the above sources are distinguished by channel as the primary key. Sales volume and satisfaction rating are stored numerically, and mention frequency and keywords are archived textually to form standardized data entries. All entries are integrated to establish a multi-channel source data set, as shown in the table below.

[0050] Table 1 Data from Multiple Channels

[0051]

[0052] As shown in Table 1, this table summarizes the original performance data of each channel within a specified period. Among them, social platforms have no direct sales volume. Their user satisfaction was calculated by performing sentiment analysis on 42,000 keywords, and the positive sentiment ratio was 75%, which is equivalent to 7.5 points.

[0053] S102: Call the comment keywords in the multi-channel source data set, classify the keywords according to the three preset dimensions of product feeling, purchase motivation and brand awareness, count the total number of times the keyword appears in each channel under each dimension, use the total number as the performance score of the corresponding dimension, and obtain the quantitative value of the channel feedback theme.

[0054] The system retrieves comment keywords from a multi-channel data set. Specifically, it matches and categorizes each keyword from 42,000 comments on social media platforms, 12,000 comments from e-commerce channels, and 350 comments from offline terminals, based on a pre-defined keyword database. This database divides keywords into three dimensions: "Product Experience" (including terms like "battery life," "clear screen," and "comfortable to wear"), "Purchase Motivation" (including terms like "discount," "gift," and "sports needs"), and "Brand Awareness" (including terms like "technological," "high-end," and "professional"). When performing this categorization, if a comment states "Great battery life and reasonable price," the count for "battery life" is incremented by 1 in the "Product Experience" dimension, and "price" is incremented by 1 in the "Purchase Motivation" dimension. After traversing all keywords, the total number of keyword occurrences across all channels for each dimension is calculated. For example, in 12,000 reviews on e-commerce channels, the keywords related to "product experience" appeared 8,500 times, "purchase motivation" appeared 6,500 times, and "brand awareness" appeared 2,400 times. These total frequencies are directly used as the performance scores for the corresponding dimensions, thereby obtaining the quantitative values ​​for each channel on the three feedback themes. That is, the quantitative values ​​for the feedback themes on e-commerce channels are: product experience: 8,500, purchase motivation: 6,500, brand awareness: 2,400. Similarly, the values ​​for social media platforms are calculated as {62,000, 41,000, 25,000}, and for offline terminals as {310, 180, 120}. These values ​​together constitute the quantitative values ​​for the channel feedback themes.

[0055] S103: Based on sales volume and satisfaction in the multi-channel source data set, and combined with the quantitative value of channel feedback themes, calculate the weighted performance score for social platforms, e-commerce, and offline terminals. Then, summarize the scores of each channel to obtain the global performance benchmark value, compare the deviation of each channel score from the global performance benchmark value, and establish an overview of the brand's market performance.

[0056] Based on sales volume and satisfaction data from multiple source datasets, and combined with the quantitative values ​​of channel feedback topics obtained from S102, a weighted calculation weight for the performance score of each channel is first set. This weight is determined based on the user interaction activity of the channel. The specific calculation process is as follows: the total mention frequency is 50,000 (social) + 12,000 (e-commerce) + 500 (offline) = 62,500 times, then the weight of each channel is:

[0057] Social media platform weight = 50000 / 62500 = 0.8;

[0058] E-commerce channel weight = 12000 / 62500 = 0.192;

[0059] Offline terminal weight = 500 / 62500 = 0.008;

[0060] Secondly, internal weights were assigned to different indicators within each channel. Regression analysis of data from the past 12 months was conducted to verify the correlation between each indicator and market share growth. The resulting weighting coefficients were: sales volume (0.4), satisfaction (0.3), and the sum of quantitative feedback themes (0.3). Before calculation, all indicators were normalized. For example, if the maximum sales volume is 15,000, the normalized value for e-commerce channel sales volume is 1, and for offline sales volume it is 5,000 / 15,000 = 0.33. Taking the e-commerce channel as an example, its weighted performance is as follows:

[0061] (1*0.4+0.82*0.3+0.54*0.3)*0.192=(0.4+0.246+0.162)*0.192=0.808*0.192≈0.155;

[0062] Where 0.82 is the normalized value of the satisfaction score of 8.2, and 0.54 is the normalized value of the sum of the quantitative values ​​of e-commerce feedback topics (17400) relative to the maximum value of all channels (social media platform 128000). Based on this method, the score for social media platforms is calculated to be 0.456, and the score for offline terminals is 0.004.

[0063] The scores from each channel are aggregated to obtain the global performance benchmark value, which is 0.155 + 0.456 + 0.004 = 0.615. Finally, the scores of each channel are compared with the global performance benchmark value to calculate the deviation. For example, the deviation value of e-commerce channel is 0.155 - 0.615 = -0.46, the deviation value of social media platform is 0.456 - 0.615 = -0.159, and the deviation value of offline terminal is 0.004 - 0.615 = -0.611. The weighted scores of each channel, the global benchmark value, and the deviation value are integrated to establish an overview of the brand's market performance.

[0064] The steps to obtain a brand emotional positioning map are as follows:

[0065] S201: Based on the brand's market performance overview, extract sentiment words and brand tags from the included consumer reviews, count the frequency of co-occurrence of each pair of sentiment words and brand tags in the same review unit, use the co-occurrence frequency as the association strength value, construct a numerical table with brand tags as rows and sentiment words as columns, and obtain the sentiment tag co-occurrence matrix;

[0066] Based on the brand's market performance overview, from a total of 54,350 consumer review texts, natural language processing was first performed to extract predefined brand tags, such as "battery life," "design aesthetics," "health monitoring," and "value for money," as well as sentiment words, such as "amazing," "reliable," "disappointing," "cost-effective," and "smooth." Then, using a single review as the smallest unit, the co-occurrence of each pair of sentiment words and brand tags was counted within each unit. For example, in the review "The battery life of this TimeSeries watch is so reliable," the co-occurrence of (battery life, reliability) was recorded once. After traversing all reviews, the total co-occurrence frequency of each pair was used as its association strength value. For example, after the statistics were completed, the co-occurrence frequency of "battery life" and "reliable" was 2,300 times, "design aesthetics" and "amazing" was 1,800 times, and "health monitoring" and "disappointing" was 450 times. These strength values ​​were then used to construct a numerical table with brand tags as rows and sentiment words as columns. This table is the sentiment tag co-occurrence matrix.

[0067] S202: Call the sentiment tag co-occurrence matrix to divide sentiment words into value perception dimensions such as function and experience, calculate the total frequency of all sentiment words in each dimension, calculate the proportion of positive sentiment words, and monitor the change of this proportion over time to generate the sentiment proportion and change range of value dimensions.

[0068] The sentiment tag co-occurrence matrix is ​​invoked. First, the sentiment words in the matrix are categorized into preset value perception dimensions based on their attributes. For example, "reliable," "smooth," and "accurate" are categorized into the "functional value" dimension, while "amazing," "comfortable," and "high-end" are categorized into the "experience value" dimension. Then, the total co-occurrence frequency of all sentiment words associated with each dimension is calculated. Taking the "functional value" dimension as an example, its total frequency = "reliable" frequency + "smooth" frequency + "accurate" frequency = 2300 + 1500 + 1200 = 5000 times. Positive sentiment words are then identified from these words. If "reliable", "smooth", and "accurate" are all positive, calculate their frequency as a percentage of the total frequency of the dimension. That is, the positive sentiment percentage of the "functional value" dimension is (2300+1500+1200) / 5000=1.0, or 100%. Then, retrieve the data from the previous monitoring period (such as the previous quarter). At that time, the positive sentiment percentage of the "functional value" dimension was 95%. Then, the change in the current period is 100%-95%=+5%. Integrate the sentiment percentage of each value dimension and its change over time to generate the sentiment percentage and change of the value dimension.

[0069] S203: Based on the co-occurrence matrix of emotional tags and the emotional proportion and change range of value dimensions, we screen the dimensions whose emotional proportion exceeds the mean and whose change range is positive. We identify the emotional words and brand tag combinations with the highest co-occurrence frequency under this dimension, determine them as core emotional features, and use their strength and dimension as coordinate axis values ​​to establish a brand emotional positioning map.

[0070] Based on the sentiment tag co-occurrence matrix and the sentiment percentage and change range of value dimensions, the mean of the sentiment percentage for all value dimensions is first calculated. Assuming "functional value" accounts for 100%, "experience value" for 85%, and "price value" for 70%, the mean is (100% + 85% + 70%) / 3 ≈ 85%. Then, dimensions with sentiment percentages exceeding this mean of 85% and showing a positive change are selected. In this example, the "functional value" dimension (100%, +5%) and the "experience value" dimension (85%, +2%) meet the criteria (assuming the change range of experience value is +2%). Next, under these selected dimensions, the sentiment words with the highest co-occurrence frequency are identified. In terms of brand tag combinations, the most frequently co-occurring combination under the "functional value" dimension is (battery life, reliability), with a frequency of 2300 times. Under the "experience value" dimension, the most frequently co-occurring combination is (design aesthetics, stunning), with a frequency of 1800 times. These two combinations are identified as the core emotional characteristics of the brand. Finally, these characteristics are visualized and mapped. For example, in a two-dimensional coordinate system, with the value dimension as the X-axis and the association strength (co-occurrence frequency) as the Y-axis, the points on the "functional value" dimension are marked at the coordinate (functional value, 2300), and the points on the "experience value" dimension are marked at the coordinate (experience value, 1800). This method is used to establish a brand emotional positioning map.

[0071] The steps to obtain the competitive brand positioning difference analysis table are as follows:

[0072] S301: Combining the dimensional framework of the brand emotional positioning map, obtain the emotional tags, product characteristic words and user usage frequency data of the specified competitors, perform word segmentation and frequency statistics on the collected text information, and associate it with the usage frequency to establish a competitor data list containing tags, words and frequencies, and obtain the competitor perception feature set.

[0073] Based on the "functional value" and "experience value" dimensions established in the brand emotional positioning map, we first initiated a data collection program targeting the competitor "Eternal Wristband." This involved acquiring user review texts from the same channels and within the same time period, totaling 48,000 reviews. These texts were segmented, and the frequency of occurrence of words identical to those in the brand tag vocabulary (e.g., "battery life," "design aesthetics") and product characteristic words (e.g., "heart rate sensor," "NFC payment") was statistically analyzed. Simultaneously, user activity data from e-commerce platforms and other channels was obtained as proxy indicators of usage frequency. For example, the "design aesthetics" tag for "Eternal Wristband" appeared 3,200 times, and the product characteristic word "round dial" appeared 2,500 times. Its average daily user interactions were 80,000. These collected and statistically analyzed tag, word, frequency, and usage frequency data were integrated into a structured competitor data list to obtain a set of competitor perceived characteristics.

[0074] S302: Call the tag frequency in the brand sentiment positioning map and match the frequency of similar tags in the competitor's perception feature set. For each common tag, perform the operation of subtracting the competitor's frequency from the target brand frequency. Use the result as the difference in brand tag volume between the two brands at a specific cognitive point.

[0075] The system calls upon the total frequency of tags in the emotional tag co-occurrence matrix upon which the brand emotional positioning map is based, and matches the frequency of similar tags from the competitor perception feature set obtained from S301. For example, the total frequency of the "battery life" tag for the brand "TimeSeries Technology" is 4500 times (derived from "reliability" 2300 times and other functional words), and the total frequency of the "design aesthetics" tag is 2800 times. Meanwhile, the frequency of the "battery life" tag for the competitor "Eternal Wristband" is 2500 times, and the frequency of "design aesthetics" is 3200 times. For each common tag, the system performs the operation of subtracting the competitor's frequency from the target brand's frequency, calculating the difference in volume for "battery life": 4500-2500=+2000, and the difference in volume for "design aesthetics": 2800-3200=-400. The result of this operation is taken as the difference in volume between the two brands at a specific cognitive point. After traversing all core tags, a series of brand tag volume differences are obtained.

[0076] S303: Based on the difference in brand tag volume, set a zero-point offset benchmark value, define tags with differences exceeding the benchmark value as advantageous areas, define tags with differences below the negative benchmark value as disadvantageous areas, and classify the rest as overlapping areas. Summarize the zoning, original frequency and difference data of each tag to generate a competitive brand positioning difference analysis table.

[0077] Based on a series of brand label volume differences, a zero-point offset baseline value is first set. This baseline value is set with reference to the average of all absolute differences. Assuming the calculated average absolute difference value is 600, the baseline value is set to 25% of this average value, i.e., 150. Labels with volume differences exceeding +150 are defined as advantageous zones, such as "Battery Life" (+2000). Labels with differences below -150 are defined as disadvantageous zones, such as "Design Aesthetics" (-400). Labels with differences between -150 and +150 (e.g., "Health Monitoring" with a difference of +80) are classified as overlapping zones. Finally, the data on the division of all core labels, the original frequency of the brand, the frequency of competitors, and the final volume difference are summarized to generate a structured table, namely the Competitive Brand Positioning Difference Analysis Table, as shown in the table below:

[0078] Table 2 Analysis of Competitive Brand Positioning Differences

[0079] Label Frequency Competitor frequency Volume difference Positioning partitions Battery life 4500 2500 2000 Advantageous areas Design Aesthetics 2800 3200 -400 disadvantage area Health monitoring 1500 1420 80 Overlapping area Cost-effectiveness 3500 100 1400 Advantageous areas

[0080] As shown in Table 2, the table clearly lists the relative positions of "Time Series Technology" and "Eternal Wristband" on key consumer perception labels, forming a competitive brand positioning difference analysis table.

[0081] The steps to obtain a consumer sentiment preference analysis report are as follows:

[0082] S401: Based on the brand sentiment positioning map and the competitive brand positioning difference analysis table, extract the gender, age, consumption frequency and category ratio data from the related user profiles. According to these profile dimensions, divide users into independent groups, count the total number of times each group mentions the brand tags in the map, and obtain the group tag interaction frequency.

[0083] Based on the brand sentiment positioning map and the competitive brand positioning difference analysis table, we first extracted the relevant user profile data of the commenters from the user database, including gender, age, purchase frequency in the past three months, and the proportion of purchased product categories. Based on these profile dimensions, users were divided into independent groups. For example, Group A was defined as "male, 18-25 years old, purchase frequency greater than or equal to 5 times", and Group B was defined as "female, 30-40 years old, fashion accessories category proportion greater than 30%". After the division, we counted the total number of times each group mentioned the core brand tags (such as "battery endurance" and "design aesthetics") identified in the brand sentiment positioning map in all comments. For example, in the 12,000 comments generated by 8,000 users in Group A, "battery endurance" was mentioned 2,500 times and "design aesthetics" was mentioned 800 times. These values ​​are the interaction frequency of the group tags.

[0084] S402: Call the interaction frequency of group tags, and use the full mention frequency of each tag in the brand sentiment positioning map as a reference benchmark to calculate the ratio between the mention frequency of each tag for each group and the reference benchmark. Compare this ratio with the benchmark value 1 to obtain a quantitative value that represents the difference in group cognition and obtain the group cognition offset value.

[0085] The interaction frequency of group tags is used, and the total mention frequency of each tag in the brand sentiment positioning map among all users is used as a benchmark. The ratio of the mention frequency of each tag for each group to the benchmark is calculated. Taking the tag "battery life" as an example, its total mentions among all users are 4500, while group A mentions it 2500 times. The total number of users is 30000, and group A has 8000 users. First, the benchmark mention rate is calculated: 4500 / 30000 = 0.15, then the ratio is calculated... The mention rate of group A is 2500 / 8000 = 0.3125. Then, the ratio between the two is calculated as 0.3125 / 0.15 ≈ 2.083. This ratio is compared with the baseline value 1 to obtain a quantitative value representing the difference in group cognition, namely, the group cognition offset value = 2.083 - 1 = 1.083. This positive value indicates that group A pays much more attention to "battery endurance" than the average level. The cognition offset values ​​of all groups for all labels are calculated in this way to obtain a complete set of group cognition offset values.

[0086] S403: For the group perception offset value, set a deviation fluctuation threshold, filter the user group whose value exceeds the threshold, integrate their profile information, comment tendency, and positioning overlap in the competitive brand positioning difference analysis table, present these filtered and integrated information items, and generate a consumer sentiment preference analysis report.

[0087] To address the group's cognitive bias, a deviation fluctuation threshold is set. This threshold is determined through historical data analysis. When the absolute value of the bias exceeds 0.5, the corresponding group preference is considered statistically significant. This value was determined through a review of the past five marketing campaigns, which found that groups with a cognitive bias exceeding 0.5 had ad click-through rates more than 20% higher than the average, providing clear guidance. Therefore, user groups with an absolute bias greater than 0.5 are selected. For example, Group A's bias for "battery life" is +1.083, and Group B's bias for "design aesthetics" is +0.65. Both groups are selected. Subsequently, the complete profile information of these groups (male, 18-25 years old, etc.), the main positive and negative sentiment words in their comments (comment tendency), and their positioning of corresponding attention tags in the competitive brand positioning difference analysis table (e.g., "battery life" is an advantage area, and "design aesthetics" is a disadvantage area) are integrated. This filtered and integrated information is presented in an itemized and structured manner to generate a consumer sentiment preference analysis report.

[0088] The steps to obtain a brand positioning optimization plan are as follows:

[0089] S501: Based on the group cognitive bias value in the consumer sentiment preference analysis report, determine whether the bias value of each group to each brand label is positive or negative, and integrate the sign and absolute value of the value to quantify it into an indicator that represents the direction and intensity of the group cognitive tendency, thus obtaining the group cognitive bias degree.

[0090] Based on the group cognitive bias values ​​in the consumer sentiment preference analysis report, the cognitive bias degree of each label for each group is quantified. This process directly integrates the sign and absolute value of the cognitive bias value. Specifically, the group cognitive bias degree is the group cognitive bias value itself, because it already contains the direction (positive or negative) and intensity (absolute value). For example, if group A's cognitive bias value for "battery life" is +1.083, its cognitive bias degree is +1.083, which represents a positive cognitive tendency with an intensity of 1.083. If group B's cognitive bias value for "design aesthetics" is +0.65, its cognitive bias degree is +0.65, which represents a positive cognitive tendency with an intensity of 0.65. If another group C's cognitive bias value for "design aesthetics" is -0.7, its cognitive bias degree is -0.7, which represents a negative cognitive tendency with an intensity of 0.7. Through this step, the cognitive bias degree matrix of each group for each label is obtained.

[0091] S502: Invoke the group perception bias and compare it with the disadvantage and overlap area labels in the competitive brand positioning difference analysis table. Filter out the labels with positive group perception bias and non-advantageous brand voice, and use them as cognitive features to be strengthened to obtain a list of brand mindshare gaps.

[0092] The generated group cognitive bias is used to compare with the disadvantage and overlapping area tags in the competitive brand positioning difference analysis table of S303. A filtering operation is performed, and the filtering rules are as follows: find a tag that is not in an advantageous position when facing competition (i.e., it belongs to the disadvantage or overlapping area), and at the same time, at least one user group holds a positive cognitive bias (greater than 0) for the tag. Taking the tag "design aesthetics" as an example, it is in the disadvantage area in the competitive analysis, but the cognitive bias of group B for it is +0.65, which meets the filtering conditions. Therefore, the tag "design aesthetics" and its associated group B information are extracted as cognitive features to be strengthened. Similarly, if the tag "health monitoring" is in the overlapping area and there is a group D with a cognitive bias of +0.4 for it, then "health monitoring" is also filtered out. All such filtered tags are collected to obtain a list of brand mindshare vacancy.

[0093] S503: Based on the list of gaps in brand mindshare, the tags in the list are grouped into core dimensions to be strengthened, and user groups with positive cognitive bias towards these dimensions are matched from the consumer sentiment preference analysis report. The strengthened dimensions are strategically matched with the target audience to establish a brand positioning optimization plan.

[0094] Based on the list of gaps in brand mindshare, the first step is to group tags such as "design aesthetics" and "health monitoring" to identify the core dimensions they share. For example, both tags can be grouped under the core dimension of "lifestyle and aesthetics" that needs strengthening. Then, from the consumer sentiment preference analysis report, groups B (women, 30-40 years old) with a positive cognitive bias towards "design aesthetics" and group D with a positive bias towards "health monitoring" are precisely matched. The core dimension of "lifestyle and aesthetics" to be strengthened is strategically paired with these two target groups to form specific action instructions. For example: "For the 'women, 30-40 years old' group, strengthen communication on 'design aesthetics' through collaborations with fashion designers and content placement in lifestyle media; for the other group, strengthen professional communication on 'health monitoring'." All such pairings of "strengthening dimension - target group - strategic direction" are summarized to establish the final brand positioning optimization plan, as shown in the table below:

[0095] Table 3. Summary of Brand Positioning Optimization Scheme

[0096]

[0097] As shown in Table 3, the table clearly matches the identified mindshare gaps, target groups with positive preferences, and specific optimization strategies, forming an executable brand positioning optimization plan.

[0098] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto.

[0099] Not limited thereto, any person skilled in the art can explore the technical scope disclosed in this invention.

[0100] Any variations or substitutions that can be easily conceived within the scope of this invention should be included within the protection scope of this invention.

[0101] Therefore, the scope of protection of this invention should be determined by the scope of the claims.

Claims

1. A brand positioning optimization method based on big data, characterized in that, Includes the following steps: S1: Collect sales volume, satisfaction rate, mention frequency, and comment keywords from multiple brand channels, categorize feedback content, calculate channel averages, compare and judge brand performance, and generate a brand market performance overview. S2: Based on the aforementioned brand market performance overview, extract the co-occurrence frequency of sentiment words and tags in the comments, calculate the sentiment proportion of each value dimension, determine the sentiment and image matching structure, extract core features, and generate a brand sentiment positioning map; S3: Based on the aforementioned brand sentiment positioning map, obtain competitor tags and their usage frequency, calculate the tag frequency difference, determine the coverage level, define the overlapping distribution of the brand and competitors in the cognitive structure, and generate a competitive brand positioning difference analysis table. S4: Based on the brand sentiment positioning map and the competitive brand positioning difference analysis table, extract user profiles and comment tendencies, calculate cognitive offset values, screen groups with prominent fluctuations, and generate a consumer sentiment preference analysis report; S5: Based on the aforementioned brand sentiment positioning map, competitor brand positioning difference analysis table, and consumer sentiment preference analysis report, determine the direction of group label shift, analyze acceptance differences and missing features, screen expression gaps and merge suggestions, and generate a brand positioning optimization plan.

2. The brand positioning optimization method based on big data according to claim 1, characterized in that: The brand market performance overview includes sales volume distribution, user satisfaction rating, brand mention frequency, and the proportion of key comment terms. The brand sentiment positioning map includes the positive and negative ratio of sentiment words, sentiment intensity classification, and the matching degree between brand image and sentiment expression. The competitor brand positioning difference analysis table includes differences in sentiment tag frequency, brand cognition overlap, and sentiment coverage intensity. The consumer sentiment preference analysis report includes the sentiment preference intensity of user groups, the degree of cognition differences between groups, and the fluctuation range of sentiment tendencies. The brand positioning optimization plan includes brand tag strengthening strategies, re-segmentation of target audience, and supplementary dimensions of sentiment expression.

3. The brand positioning optimization method based on big data according to claim 1, characterized in that: The steps for obtaining the brand's market performance overview are as follows: S101: Collect data on brand sales volume, user satisfaction, brand mention frequency, and comment keywords across social media platforms, e-commerce channels, and offline terminals. Group the raw indicators according to their source channels, record sales volume and satisfaction numerically, and collect mention frequency and keywords textually. Establish standard format data entries and generate a multi-channel source data set. S102: Call the comment keywords in the multi-channel source data set, classify the keywords according to the three preset dimensions of product feeling, purchase motivation and brand awareness, count the total number of occurrences of the keywords in each channel under each dimension, use the total number as the performance score of the corresponding dimension, and obtain the quantitative value of the channel feedback theme. S103: Based on the sales volume and satisfaction in the multi-channel source data set, and combined with the quantitative value of channel feedback themes, calculate the weighted performance score for social platforms, e-commerce, and offline terminals. Then, summarize the scores of each channel to obtain the global performance benchmark value, compare the deviation of each channel score from the global performance benchmark value, and establish an overview of the brand's market performance.

4. The brand positioning optimization method based on big data according to claim 1, characterized in that: The steps for obtaining the brand emotional positioning map are as follows: S201: Based on the aforementioned brand market performance overview, extract sentiment expression words and brand tags from the included consumer reviews, count the frequency of co-occurrence of each pair of sentiment expression words and brand tags in the same review unit, use the co-occurrence frequency as the association strength value, construct a numerical table with brand tags as rows and sentiment words as columns, and obtain the sentiment tag co-occurrence matrix; S202: Call the aforementioned sentiment tag co-occurrence matrix, divide sentiment words into value perception dimensions such as function and experience, calculate the total frequency of all sentiment words under each dimension, calculate the proportion of positive sentiment words, monitor the change of this proportion over time, and generate the sentiment proportion and change magnitude of the value dimension. S203: Based on the emotional tag co-occurrence matrix and the emotional proportion and change range of the value dimension, filter the dimensions whose emotional proportion exceeds the mean and whose change range is positive, identify the emotional words and brand tag combinations with the highest co-occurrence frequency under the dimension, determine them as core emotional features, and use their intensity and dimension as coordinate axis values ​​to establish a brand emotional positioning map.

5. The brand positioning optimization method based on big data according to claim 1, characterized in that: The steps for obtaining the competitive brand positioning difference analysis table are as follows: S301: Combining the dimensional framework of the brand emotional positioning map, obtain the emotional tags, product characteristic words and user usage frequency data of the specified competitors, perform word segmentation and frequency statistics on the collected text information, and associate it with the usage frequency to establish a competitor data list containing tags, words and frequencies, and obtain the competitor perception feature set. S302: Call the tag frequency in the brand sentiment positioning map and match the frequency of similar tags in the competitor perception feature set. For each common tag, perform the operation of subtracting the competitor frequency from the target brand frequency. Use the operation result as the difference in the volume of the two brands at a specific cognitive point to obtain the brand tag volume difference. S303: Based on the brand label volume difference, set a zero-point offset benchmark value, define labels with differences exceeding the benchmark value as advantageous areas, define labels with differences below the negative benchmark value as disadvantageous areas, and classify the rest as overlapping areas. Summarize the affiliation, original frequency and difference data of each label to generate a competitive brand positioning difference analysis table.

6. The brand positioning optimization method based on big data according to claim 1, characterized in that: The steps for obtaining the consumer sentiment preference analysis report are as follows: S401: Based on the brand emotional positioning map and the competitive brand positioning difference analysis table, extract the gender, age, consumption frequency and category ratio data from the related user profiles, divide users into independent groups according to these profile dimensions, count the total number of times each group mentions the brand tags in the map, and obtain the group tag interaction frequency. S402: Call the group tag interaction frequency, and use the full mention frequency of each tag in the brand sentiment positioning map as a reference benchmark to calculate the ratio between the mention frequency of each tag for each group and the reference benchmark. Compare this ratio with the benchmark value 1 to obtain a quantitative value that represents the difference in group cognition and obtain the group cognition offset value. S403: For the aforementioned group cognitive offset value, set a deviation fluctuation threshold, filter out user groups whose values ​​exceed the threshold, integrate their profile information, comment tendencies, and positioning overlap in the competitive brand positioning difference analysis table, present these filtered and integrated information items, and generate a consumer sentiment preference analysis report.

7. The brand positioning optimization method based on big data according to claim 1, characterized in that: The steps to obtain the brand positioning optimization plan are as follows: S501: Based on the group cognitive bias value in the consumer sentiment preference analysis report, determine whether the bias value of each group to each brand label is positive or negative, and integrate the sign and absolute value of the value to quantify it into an indicator that represents the direction and intensity of the group cognitive tendency, and obtain the group cognitive bias degree. S502: Call the aforementioned group cognitive bias and compare it with the disadvantage and overlap area labels in the competitive brand positioning difference analysis table. Filter out the labels with positive group cognitive bias and non-advantageous brand voice, and use them as cognitive features to be strengthened to obtain a list of brand mindshare gaps. S503: Based on the brand mindshare gap list, group the tags in the list into core dimensions to be strengthened, and match user groups that hold a positive group cognitive bias towards these dimensions from the consumer sentiment preference analysis report. Strategically match the strengthened dimensions with the target audience to establish a brand positioning optimization plan.

Images