# Analysis of payment habits (3): How to maximize the revenue by discounting PK coupons?

A game has previously done an item discount campaign, and now I want to try a full coupon delivery campaign, and I want to know which kind of activity method brings better income. The data analyst evaluated and answered the above questions by analyzing the report. For details on how to evaluate it, see this article.

01 Analysis ideas

The data analyst made the following six steps of analysis based on historical prop sales records, as shown in Figure 1.

• Analyze user consumption characteristics
• Determine the “full” base
• Determine the amount of “coupon”
• Compare the number of item purchasers in the “Full coupon delivery” and “20% off item sales” activities
• Compare the income of the “Full coupon delivery” and “20% off item sales” activities
• Determine activity plan

figure 1

02 Detailed analysis process

(1) Descriptive statistics of user consumption records

It can be seen from Fig. 2 that the user’s spending power has a right-skewed distribution.

figure 2

According to further statistics, 57% of users spend less than 3 times, and 56% of users spend less than 300 yuan, and users belong to low-frequency and low-volume consumer groups.

According to the user’s historical consumption amount, 128 yuan for the quartile, 256 yuan for the 1/2 quantile, and 584 yuan for the 3/4 quantile were obtained.

(2) User purchase frequency & average purchase amount interval distribution

According to the distribution of the cumulative consumption times of users shown in Figure 3, it can be concluded that 57% of users spend less than 3 times, 31% of users spend only once, and 15.69% of users spend more than 10 times, and 15 times The above users accounted for 10.04%.

image 3

Take a closer look at the types of props purchased by users who have accumulatively consumed more than 15 times. TOP3’s props are fashion, functional props, and grades, as shown in Figure 4.

[85,95],[125,135]These are the two peaks of the average consumer consumption, as shown in Figure 5.

Figure 4

Figure 5

(3) Top selling props TOP30:

Among the currently popular TOP30 props, 128 yuan, 88 yuan, and 70 yuan have the largest number of props, as shown in Figure 6.

Figure 6

(4) The average daily ARPPU of a single user is 100 yuan. Except for the individual activity time, the rest of the time is relatively stable, as shown in Figure 7.

Figure 7

(5) From the historical data, the average amount of items purchased by users per day is concentrated in[90,130]Interval, as shown in Figure 8.

Figure 8

(6) Effect of historical activities:

Some of the item discount activities have been done before “Sunshine Disputes”. The discount range is between 70% and 80%. The discount range of this full coupon reduction event is also expected to stay between 70% and 80%. Therefore, the historical activity data is very Good reference.

A total of 12,837 users participated in the discount purchase of props. Figure 9 is the distribution of the number of items purchased during the discount period. Table 1 shows the number of purchases of discounted items. This event generated a total of 2.35 million yuan in revenue. The average user payment amount (ARPPU) is 183 yuan, which is 47% higher than the ARPPU of a single user without full reduction activities.

Figure 9

Table 1

(7) Compare the two options

Refer to the previous 20% off activity data, that is, the full amount a, and obtain the three grade bases according to the historical 20% off (128 yuan for the 1 / 4th quantile, 256 yuan for the 1 / 2th quantile, 3 / 4 quantile is 584 yuan), then the full amount a = radix / 0.8 (respectively 160 yuan, 320 yuan, 730 yuan).

In the reference activity period, the ARPPU is increased by 47% as the target, that is, the full amount a is increased by 47% according to the current base number, then the full amount a = base number * 1.47 (respectively 188 yuan, 376 yuan, and 858 yuan).

Applying the 28 rule, 80% of the revenue comes from 20% of high-end users. The proportion of cumulative consumption greater than 730 yuan is 19.8%, and the proportion of greater than 858 yuan is 16.5%. 730 yuan is more reasonable than 858 yuan, so it is set at 730 yuan. Full third grade.

At present, the amount of mainstream props is 128 yuan, 88 yuan, and 70 yuan. If the user purchases these three mainstream necessary props respectively, the difference between the remaining costs will be easier to implement. In the second grade, 160 yuan is easier than 188 yuan users. Choose other amounts of props to achieve the full amount, and the increase rates are 20%, 18%, and 4%, as shown in Table 2.

Table 2

(8) The 7 historical activity time periods experienced by the mall, according to the current full coupon delivery effect:

The historical prop mall has experienced 7 large-scale events, and the third event is a watershed. In the first two events, the number of mall props was small, and the number of active users was at a low level. From the third event, the number of props increased, and the corresponding period was active. And the number of payers increased until the seventh time, which are relatively stable levels. Therefore, the data of the third to seventh times in history is used as reference data to determine the proportion of users with different consumption levels, the first level is 55%, the second level is 25%, and the third level is 20%. . The average payment rate of the 5 events is 14.34% as the payment rate of this event.

Table 3 shows the proportion of consumption amount and payment rate for different grades of 7 activities.

table 3

(9) Estimated income during the event: Note: By default, the activity cycle of this event and previous events are close, both are about 15 days.

a. Full coupon delivery

Based on the current number of active users in the past two weeks, and then according to the proportion of the people who are in three levels of consumption during the past 5 activities: 55%, 25%, and 20%, the number of people who purchased the three levels (n1, n2, n3) can be obtained. n1 = the number of active users * 55%; n2 = the number of active users * 25%; n3 = the number of active users * 20%.

Calculate the total income of this event, total income = 160 * n1 + 320 * n2 + 730 * n3

b. Item discount activities

If all items are 20% off, calculate the number of purchases of each item in the mall during the event.

Since some discounts on props have been done in the historical mall activities, the ratio of the number of purchasers of 20% discount items to the total number of purchasers of this item’s history is first calculated, and this ratio is used as the number of purchasers of this item.

According to historical data: the unit price of the item is less than 160 yuan, and the number of purchasers accounts for 9.98%; the unit price is[160,320] The proportion of purchasers of props accounted for 17.10%, and the number of purchasers of unit prices greater than 320 yuan accounted for 16.29%.

Number of item purchases = historical number of purchases of each item * 20% off purchase percentage during the event

Total income = unit price * number of purchasers * 0.8

Comparing the income of full coupons and item discount activities, it is concluded that the discount of props is 1.28 million yuan lower than the income brought by full coupons.

03 Analysis conclusion

According to the data analysis of the purchase of props by users in “Scorching Sun”, it is suggested that this month’s item sales plan for full coupons will be: RMB 160 for 50 yuan, RMB 120 for 320 yuan, and RMB 300 for 720 yuan It can bring in 6.3 million yuan in income, which is 300,000 yuan higher than the direct discount. The analysis steps and conclusions are as follows:

(1) Analysis of consumer consumption characteristics based on historical prop purchase records:

a. The best-selling top 30 props are mainly 128 yuan, 88 yuan and 70 yuan.

b. Users belong to the low-frequency and low-consumption consumer group. 31% of users spend only once, 57% of users spend less than 3 times, and users who consume more than 10 times account for 15.69%. Fashion is the main type of props purchased by users; 85-95 yuan and 125-135 yuan are the two peaks of the average consumer consumption.

c. Its consumption capacity is right-skewed, with 56% of users accumulatively spending less than 300 yuan and a mode of 128 yuan.

d. A single user’s daily value (item purchase amount) is 100 yuan, except for the individual activity time, the rest of the time is relatively stable.

Note: User consumption characteristics are related to item pricing and item attributes (whether they can be purchased repeatedly).

(2) Determine the “base” base:

a. The 4th quantile of the historical consumption amount is initially divided into three “full” grade breakpoints (full a): 128 yuan in the quartile, 256 yuan in the 1/2 quantile, and 3/4 The quantile is 584 yuan. Combined with the unit price of the item, it can be seen that 128 yuan is the mode and the 1 / 4th quantile is mainly due to the unit price of the mall item, and 128 yuan is within the peak range of the average consumer consumption, so it is reasonable as the first base.

b. Refer to the sales activity of 20% off the historical props, that is, the full amount a. According to the historical consumption situation, convert the 20% off the three grade bases, then the full amount a = base / 0.8 10 times the odd number, so choose an even number close to it (720 yuan).

c. The goal is to increase the ARPPU value of historical prop sales activities (up 47%), that is, the full amount a. If the current base number is increased by 47%, then the full amount a = base number * 1.47 (188 yuan, 376 yuan, and 858 yuan respectively) ).

d. Suppose that the user purchases three mainstream props (128 yuan, 88 yuan, 70 yuan) respectively. Compared with the “full amount” base of points 2 and 3 above, 160 yuan is easier for users to choose other popular items. The full amount was achieved, and the increase rates were 20%, 18%, and 4%.

e. According to the 28 rule (80% of revenue comes from 20% of high-end users), the proportion of cumulative consumption greater than 720 yuan is 19.8%, and the proportion of greater than 858 yuan is 16.5%, so it is more reasonable to define 720 yuan than 858 yuan. .

Combining the above five points, the bases of “full” are determined to be: 160 yuan, 320 yuan, and 720 yuan.

(3) Determine the amount of “coupon”:

Considering the mainstream prop amount and the coupon amount are not less than 20% off the props, the two factors are discounted, and then rounded according to the algorithm to determine the amount of coupons: 50 yuan for 160 yuan, 120 yuan for 320 , 300 yuan coupons over 720 yuan.

(4) Estimated income from “Full coupon delivery” and “20% off props” activities:

a. According to the “full coupon delivery” campaign, the estimated income during the month of the campaign is 16.3 million yuan.

i. Total income = 160 * n1 + 320 * n2 + 730 * n3. Where n is the number of items purchased = weekly active * payment rate, according to the proportion of people who spend at three levels in 7 historical activities: 55%, 25%, 20%, so the values โโof n1, n2, and n3 are obtained, and finally Total revenue.

b. According to the “Discount 20%” campaign, the estimated revenue during the campaign is 15 million yuan.

i. Total income = โ unit price of props * number of purchasers * 0.8. Of which the number of item purchases = the historical number of purchases of each item * 20% off the purchase ratio during the event

In summary, it is expected that the “full coupon delivery” activity this month will be 1.3 million yuan more than the “promotion of 20% off items”.

Analysis of MMO game payment habits (1)
Analysis of payment habits (2)

Author: Lixiang Yan
Source: Data Driven Games