An educational institution was founded in 1999, the early years of the development of downline vocational education, has achieved a certain scale, there is a certain mass base. In recent years, Internet education is hot, the company actively transformed to the Internet field. Today, it has opened nine directions of large specialties, covering a number of popular positions in Internet companies. Its main big data development education, Java education, data analytics education courses sales volume has increased year by year, and has begun to turn losses into profits.
The company's excellent growth results are due to its complete user platform and efficient user realization system. At the beginning of the company's transformation, it was not a simple online education approach, but a model of education community + online course platform. Through the education community, users are attracted to discuss career issues, and after a certain accumulation of career knowledge information, through SEO, users share discussion posts to convert netizens who have a need for career knowledge into community users. After that, through the advertising space in the community, public number promotion, outside advertisement promotion, outbound call, etc., to collect potential user leads who have a need for the course. The unified use of telemarketing team to convert the leads, so that the students through the purchase of membership to open the course platform of the small class experience function, to achieve the first step of the conversion; after that, for the students who have a large can demand, and then do further conversion, so that they buy the platform of the large class service, become the platform students. In the process of students' learning, its teaching and research team and product development team are also constantly polishing courses and optimizing products. Through years of accumulation, the platform has accumulated a high reputation, and nearly 60% of the converted students expressed their willingness to continue to take other courses or advanced courses of the company. 70% of the trainees are willing to recommend the company's courses for their classmates or friends; at the same time, the marketing team has also designed marketing campaigns that allow trainees to share information about course activities or lectures by famous teachers on various holidays to achieve the effect of fission spread. The whole self-propagation of students has a certain scale, for the company's promotion to save a lot of money.
Revenue conversion model for the education industry
The company is more concerned about the profitability of online education, the main assessment of gross profit, gross profit = revenue - cost of goods to calculate, and for the breakdown to specific goods, gross profit can also be calculated as Gross Profit = ∑ (selling price of a single item - cost of a single item) * number of single items. At this point, the selling price of a single product, the cost of a single product, the sales volume of a single product is the main factors affecting the amount of gross profit.
Online education products have the same characteristics as Internet products: they are produced once and can be sold an unlimited number of times. The more times it is sold, the lower the cost per copy of the course, which can effectively increase the gross profit amount even if the selling price of a single product remains unchanged. So, generally speaking, increase course sales, you can increase the gross profit amount.
This year's profit target of 10 million gross profit amount, from January to August gross profit amount of the situation, the completion of the situation is not ideal, only completed 3.789 million, accounting for 38% of the annual task, and then after the leadership of the discussion, the 10 million gross profit amount of the target adjusted to the existing 7 million yuan, it is expected that from September to December, the need to complete the 3.2 million gross profit amount of the target. Since September and October are the most popular months for job changes, and demand for training programs is higher during this time than in other months, 70% of the gross profit will be covered in September and October, or $2.24 million, or an average of $1.12 million per month, and 30% will be covered in the following two months, or an average of $480,000 per month.
Recently, the company has encountered a problem in that course sales have increased while gross profit has declined, which requires the data analyst to analyze the cause of the problem and provide a solution.
Second, the problem recognition and indicator dismantling
1, to confirm the problem of abnormal gross profit amount
2, to improve the gross profit amount of the program
Third, the problem solving ideas
Fourth, the application of the actual method
1, the data up and down the analysis of the variance - the problem of dismantling (logic tree)
1) statistics of the month sales and gross profit amount of line graphs, from the graph of the sales and gross profit amount of the data analysts, the data analysts can not be used to determine the cause of the problem. Amount of the line graph, from the graph to see whether the sales rose, gross profit fell
2) Calculate the gross margin and gross margin chained to each month and plot line graphs, to see whether it is a normal fluctuation.
3) Start the verification from the channel promotion, count the sales and gross margin of each type of channel and draw a quadrant graph, find the type of channel with high sales and low margin from the graph.
4) from the channel type refinement to specific channels, the same statistics of each channel sales and gross margins and draw quadrant diagrams, from the illustration to find high sales of low gross margin channels.
Introduction to Logic Tree Disassembly Analysis
Logic Tree Results
Through the month-over-month data, we know that the decline in gross profit amount is not a fluctuation in the data, the decline is 32%
Through the Logic Tree, we will be ordering data in accordance with the type of channel and the specific channels to be disassembled, and use correlation matrix analysis to find out the type of channels with lower than average sales volume and gross profit margins: Free channels. In the free channel, we disaggregate the channel name where sales and gross margins are lower than average: onsite advertising.
After that, we looked closely at the August data for the onsite ad space, and found that the anomaly was caused by the August 17th big class promotion.
Starting from the channel type, we plotted the sales and gross margin quadrant of each channel (matrix correlation analysis)
We drilled down to the segmented channels under the free channel type, and found that the August data of in-site advertising space had extremely high sales and low gross margin, indicating that it was likely that the activities related to in-site advertising space had affected the overall amount of gross margin;
Through the comparative analysis, we determined that in-site advertising space was the problem, and through the July and August data, we found that it was caused by the big class promotion.
Through comparative analysis, to determine the problem of in-site advertising space, through the July and August channel gross margin and channel sales comparison, found that the gross margin of the main decline in the channel for the station advertising space, EDM, SMS, 0 yuan experience class.
Among them, due to the station ad space channel in August 17 to do promotional activities, the day there are 20 single gross profit amount of negative orders. This resulted in a negative gross margin for the channel for the month.
While examining the contribution of different detailed channels to the free channel sales increase (that is, how much of the increase in sales of the free channel is from this specific channel.) It can be seen that due to the EDM SMS channel sales have dropped, the station advertising space sales contribution rate of more than 100%, reaching 100.69%;
shows that the August sales rise, the decline in profitability is mainly due to the station advertising space channel sales rose a lot while the gross margin fell a lot of the problems caused by. From the above two aspects, it is clear that the problem of the increase in sales and decrease in gross profit in August is mainly due to the free channel of the in-site advertising channel. The channel in August 17th enrollment appeared negative gross margins,; and because of the activities of the promotion is a new course this month, the course according to the month cost sharing is higher, so there is a negative gross margins, pulling down the overall gross margins.
2, channel ranking ---- comprehensive analysis
1), the channel of the four key indicators (ROI, enrollment conversion rate, the bounce rate, the amount of leads) through the optimization matrix for weight division to determine the weight value of each indicator.
2), the channel indicators for 0-1 standardization, to eliminate the impact of the difference between the quantitative outline.
3), the use of standardized indicators and indicators of the product and calculate the comprehensive score of each channel. Use the comprehensive score to rank the channels.
Comprehensive analysis method:
Conclusion:?
The best existing channel is the in-site advertising space, and the second best channel is Zhihu KOL.
However, due to the fact that placing ads involves a lot of factors, such as the problem of ad re-investment, the copywriting cycle, and the length of the conversion cycle, the subsequent improvement strategy needs to be further negotiated with the marketing department.
Data introduction: 9 channels ROI, enrollment conversion rate and other key indicators of 4 channels
Target optimization matrix way to confirm the weight of the indicators
3 , analysis of the training course of the various aspects of the conversion situation - product user behavior analysis (funnel analysis)
1), make funnel analysis
2), find the problem link
3), analyze the cause of the problem link
4), the experimental design of the link
5), the analysis of the target enhancement degree, the results of the improvement is estimated to enhance the amount of how much gross profit amount. This part needs to write a specific estimation algorithm.
The last analysis is specific to the analysis of each channel, but this analysis does not have channel data; it is the analysis of the overall data;
Funnel analysis:
Funnel analysis:
Funnel analysis conclusion: through the funnel analysis, found that the conversion rate of the funnel in the reserved phone link is significantly lower than the other links
From the overall final conversion is lower than the industry conversion level of 3%-5%, and the consultation to the reserved phone link is lower than the industry conversion level, and the consultation to the reserved phone link is lower than the industry conversion level of 3%-5%.
After that, we analyzed the log data of the reserved contact link and found that the user waiting time is too long.
a), the average user waiting time for manual customer service is too long
From the analysis of manual customer service response time, the average user response time is 187 seconds, about 3 minutes. Afterwards, through user interviews, found that the problem of long manual waiting time also appeared in the area of more feedback problems (the problem ranked as the fourth of the 30).
b), need to follow up ABtest to draw the corresponding conclusions
4 ), trainee value analysis - RFM modeling
1, we randomly selected 0.3% of the users in the daily activity of 100,000 people in the small class, through the RFM to distinguish between the rank of the trainees
RFM analytical method results:
follow-up plan:
p> To design a refined operation program for trainees, we first use recommendation algorithms to calculate the customer's area of interest.
For important value customers: you can use the program of directly recommending the purchase of large classes to attract customers to convert.
For important customers: we can use the small class promotion program to attract customers to repurchase, and then gradually recommend the large class to promote the conversion of users.
For important development customers: buy a large class and get a free membership to attract customers to convert.
For the important retention of customers: we can use the small class promotion program to attract customers to repurchase, and then gradually buy a large class to send members to attract customers to convert.
For the general value of the customer: you can use the large class limited time promotional program to attract users to convert.
For the general development of customers: the program can be used to attract users to convert from small class to large class promotion.
For the general retention of customers: we can use the small class promotion program to attract customers to repurchase, and then gradually recommend the large class to promote the conversion of users.
For general retention of customers: you can first do user recall, and then gradually start to attract customers to repurchase the program with small class promotions, and then gradually recommend large classes to promote.
1, August station advertising position although there is a loss activity, but January-August data assessment is still a high-quality channel, the follow-up need to focus on polishing, through analysis we found that sales rose in August and the gross profit amount of the abnormal decline is due to the station advertising channel to do discount promotions triggered. Due to the discount, the setup error led to a book loss, which triggered the decline in gross profit. The problem has been feedback to the marketing team, the follow-up will design the price, early warning system, if there is a similar price setting errors, it will be timely warning; through comprehensive analysis, the ROI of each channel, the number of people enrolled in the conversion rate, the bounce rate, the amount of leads to do a comprehensive index, calculated 19 years ago in August, the most, the best channel for the station advertising channels. We will further discuss the channel improvement program with the marketing team based on the characteristics of each channel.
2, stay electricity link for the focus of optimization, optimization completed on the overall conversion rate increase. Through the funnel analysis, we found that the conversion rate of the user conversion link is low; we worked with the customer service and product teams to develop an improvement plan. After a month of improvement, the conversion rate of the call-retention session increased from 22.57% to 44.16%, and the overall conversion rate was improved. The number of people converted to the big class increased by 330 people compared to August, calculated according to the amount of gross profit per person per capita of 1,600 yuan, it is expected to improve the amount of gross profit of 528,000 yuan (Note: the actual profit needs to be known after accounting); calculated according to the number of viewers of 16,034 people in the same month, it is expected that the number of people enrolled in the number of 586 people, it is expected that the improvement can be accomplished 937,600 yuan gross profit amount of the task;
3, the refinement of the user After the promotion of the operation model, 501 people have been converted to large classes and 3,377 people to small classes, and the model will be promoted to the 100,000 daily activity of the small class user group to get the complete customer stratification distribution, see the following chart. After the user refined operation model promotion, has been converted to large class 501 people, small class 3377 people due to the large class to complete the gross profit amount of the task has been included in the previous analysis. According to the small class per capita gross profit amount of 17 yuan to project, small class is expected to complete the task of 57,000 yuan gross profit amount.