What is the structure of marketing expenses? What is the marketing cost in the past two quarters? Is its marketing cost efficiency really getting worse?

Editor’s note: This article is from WeChat public account “Growing Black Box Growthbox” (ID :Growthbox2), by John & Alan.

The financial report of the 2019Q2 will be released on Wednesday (August 21). This article does not constitute any investment advice.

A lot of people know how to fight, or start with its spelling link. The initial fight has been rapidly fissed by the block mode, and the WeChat acquaintance network has earned a bonus, which also allows it to maintain a low customer cost in the case of high GMV growth.

However, by 2018, the situation has changed. The marketing expenses announced by the many financial reports have risen from 3.2 billion yuan in 2018Q3 to 6 billion yuan in Q4. The marketing expenses have almost doubled, and in 2019Q1, it still maintains a high level of 4.8 billion yuan.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Source: Fighting a lot of earnings)

The secondary market began to pay attention to the sharp increase in customer acquisition costs. Especially after the release of the Q1 financial report in 2019, the high marketing expenses did not bring about the expected GMV growth, which affected the stock price.

Even analysts at Goldman Sachs repeatedly asked about the details of spending on marketing expenses in a large number of conference calls, but did not give a clear response to many aspects.

So, what is the structure of the marketing expenses? What is the marketing cost in the past two quarters? Is its marketing cost efficiency really getting worse?

The growth black box has assembled a variety of alternative data providers in China on this issue, trying to explain these problems with data.

Before the text begins, we will first refine the ideas of the full text as follows, so that everyone can read:

  • A lot of marketing expense structure

Where is the huge amount of marketing expenses spent?

What are the brand ads, performance ads, and subsidies?

Is there a lot of marketing cost efficiency?

  • A lot of marketing strategies

What is the effect of a lot of advertising strategies?

What are the products that are subsidized in the end?

1. A lot of marketing expense structure

1.1 Where is the huge marketing cost spent?

The 2018 financial report shows that the sales and marketing expenses for the whole year are 13.4 billion yuan, while the total revenue last year was only 13.12 billion. This means that almost every dollar invested is invested in marketing activities.

Where is the marketing cost of a lot of money spent?

The official financial statement is: Online and offline advertisement and promotions andbranding campaigns.

Translation is the following four parts,

  • Online advertising (also known as “effect advertising”): including search engine advertising information flow advertisements, etc., such as wide-point communication, huge engine, etc., because it can guide users to download the app directly, the effect is easy to measure, so We also call it an performance ad;

  • Offline advertising: Ads placed in places such as elevator buildings, such as some of the advertising spaces under Focus Media;

  • Brand activities: sponsorship of online variety shows and online TV shows, as well as TV commercials;

  • Promotional subsidies: Generally speaking, as long as they are sold below the market price, they can be regarded as a kind of subsidy, such as “one-person purchase exclusive to newcomers” and “billion-subsidy” activities.

We used the alternative data to predict the marketing performance of the second quarter of the fight

To make it easier to understand, we reduced the marketing cost to three parts and estimated it:

  • Performance ads (based on online advertising. About 7 billion RMB in 2018)

  • Brand advertising (ie offline advertising plus branding activities. In 2018, offline advertising is about 200 million RMB, brand activity is close to 2 billion RMB)

  • Promotional subsidies (about 4.2 billion RMB in 2018)

We used the alternative data to predict the marketing performance of the second quarter of the fight.

We have drawn the logic of estimating data and reasoning into the following figure. (At present, no research institute has dismantled a lot of marketing expenses to this granularity, and the estimation is for reference only)

We used the alternative data to predict the marketing performance of the second quarter of the fight.

The above graph uses data from two alternative data providers:

  • Too much GMV estimate for each quarter (there is no such indicator for a lot of earnings), we use it to estimate subsidy fees;

  • Statistics of the number of performance ads that we use to calculate performance advertising costs.

The multi-quarter quarterly GMV data is provided by the “Super Symmetric” team. They predict GMV by monitoring the changes in the product data of the iPad. The super-symmetric forecast is more than 2049 billion GMF for 2019Q2.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: Supersymmetric Technology)

From this data, the second quarter of 2019GMV has reached the highest level in history.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: Supersymmetric Technology)

We assume that the subsidy fee is directly proportional to the GMV. According to this logic, the subsidy in the second quarter of 2019 is about 2 billion yuan, which is bound to be a new high. Judging from the official propaganda, the company has launched the largest subsidy activity in history in the second quarter, claiming to be “a billion subsidies”. From this perspective, it is also reasonable to make a high subsidy.

The number of performance advertisements in the past one year was provided by App Growing. App Growing monitored the number of advertisements of more than 30 mainstream mobile platforms, and most of the effect advertisements were put on these platforms. There is a certain difference between the number of statistics and the number of actual delivery, but using these data to see the trend of delivery is still relevant.

We used the alternative data to predict the marketing performance of the second quarter of the fight

(Data Source: App Growing)

From the data of App Growing, it was only in the fourth quarter of 18 years that the game began to focus on online advertising. If you look at the number of placements alone, the number of advertisements in the first two quarters of 2019 will exceed the volume of the previous year.

1.2 What is the focus of marketing in 2019?Where?

According to a number of brokerage research reports, the marketing cost of 2019 is about 22 billion to 25 billion. We combined the average value of the company with RMB 23.2 billion as the market expectation for the marketing expenses of 2019.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

And the marketing cost in 2018 is only 13.4 billion, which is likely to generate more than 10 billion marketing expenses?

First of all, look at the brand advertisements of a lot of brands. Since the excellent brand advertisements are mostly locked in advance last winter, the expectations here are relatively certain. According to the feedback from senior advertisers, the brand advertisements of 2019 are basically the same as last year. At the same time, there has been no significant change in the delivery of offline advertising.

Therefore, our forecast for the brand advertising cost of 2019 is still about 2.2 billion last year, and no increase has been found here.

We used the alternative data to predict the marketing performance of the second quarter of the fight

(Data Source: Web Public Information)

So, the more than 10 billion marketing expenses that may be added this year will be invested in performance advertising and subsidies.

The actual trend also confirms this point. Just like the chart of the number of advertisements we just showed, the number of online advertisements has increased dramatically (the number of advertisements in July 19 has reached a new high, which is not shown in the figure below) . According to the current trend, the cost of advertising for this year may even double that of last year, when marketing expenses will be more exaggerated than analysts’ expectations.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: App Growing)

At the same time, in terms of subsidies, the company has launched a massive “billion-subsidy” activity in the second quarter of this year.

Know that more than 20 billion will be subsidized in 2018.

From the past news, the fight has been raging from the “billions of subsidies” since May, and was packaged into the form of “a beautiful life million group”. In June, it became the main activity of 618. As of now (in mid-August), it has lasted for more than three months, and it has already exceeded everyone’s perception of “year-old promotion.” Regardless of the cost of marketing or the monitoring of subsidized goods, the subsidies in the first two quarters of 2019 are farLess than “tens of billions.” Therefore, we will regard the “billion-dollar subsidy” as an annual plan, and it is difficult to spend money efficiently.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Source: Investment)

Overall, the marketing strategy of Putuo in 2019 is relatively clear, that is, a large number of procurement effect advertisements for APP drainage, and the use of extremely high subsidies to attract buyers to place orders, we will carry out detailed in the latter analysis.

2. Spend more marketing cost efficiency

In combination with the analysis reports of many brokers, the cost of winning a lot of customers has soared from 2018.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

Obviously, the low-cost means of relying on WeChat’s “cracking” is no longer sufficient to support its high-growth demand.

In fact, a lot of people are also jumping out of the comfort zone – try to use a large amount of online advertising with a higher amount of subsidies in exchange for the share of first- and second-tier cities.

2.1 App users have a significant growth effect, and the proportion of users in first- and second-tier cities increases

According to Aurora Mobile (NASDAQ: JG), the monthly average DAU of the iPad has continued to rise, and the growth rate has accelerated significantly since February this year.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

In terms of geographical distribution, user growth in second-tier and above cities plays an important role in all growth.

In the first quarter of 2019, 44.2% of the new users were from second-tier cities and above.

We used the alternative data to predict the marketing performance of the second quarter of the fight

In addition, according to the brokerage report, from 618 to 2019 in 618, the proportion of users in the first- and second-tier cities increased significantly. The fight also said that during the 618 this year, 70% of agricultural products orders came from the first and second lines.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

So, we can simply make a conclusion: more and more are using real money to seize the market share of the first and second lines, and have achieved certain results.

2.2 Marketing cost efficiency has improved

After spending so much marketing expenses, what is the improvement effect on the core indicators?

We can compare the ratio of marketing expenses to multiple core operational indicators to see how much marketing cost efficiency.

First of all, for GMV data, the [Marketing Cost/GMV] ratio has decreased from 2019Q1 to Q2, which represents a lower marketing cost in order to increase a certain GMV. In this way, the efficiency of marketing expenses has increased recently, but it has not returned to the level of 2018.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: Supersymmetric Technology, Growth Black Box)

For the number of active users, whether based on the monthly average DAU provided by Aurora Big Data or based on the MAU data provided by QuestMobile, the ratio of [Marketing Expenses / Active User Growth] has continued to rise in the previous three quarters. And quickly fell back in 2019Q2. It should be noted that, according to a lot of financial reports, its definition of Active User only includes App users, and the DAU and MAU data here are the same. The fall of this ratio shows thatA lot of marketing expenses have significantly improved the efficiency of app acquisition and promotion, which is obviously due to the “short and fast” effect advertisement.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: Aurora Big Data, Growth Black Box)

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: QuestMobile, Growth Black Box)

And for the number of sellers, it is also an important part of the multi-operational indicators. Although the number of active sellers has been increasing, the growth rate has slowed down noticeably. From the point of view of [marketing expenses/increased number of active sellers], the ratio of the second quarter has increased significantly, which means that the efficiency of the increase in the number of active sellers has become lower.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: Supersymmetric Technology, Growth Black Box)

Overall, from the point of view of GMV and active users, the efficiency of marketing expenses is much improved.

At this point, a brief review. By estimating the structure of marketing expenses, we have inferred a lot of marketing strategy focus – to divert apps through a large number of effect ads, and use extremely high subsidies to attract users to place orders. In this process, the expansion of market share in first- and second-tier cities is significant, and from the perspective of the improvement of GMV and active users, the efficiency of marketing expenses has been improved.

What is the marketing strategy that is packed with more details? We will explore specific marketing strategies based on performance advertising and subsidies.

3. A lot of heavy money marketing strategy

3.1 Guide App download is a key target for performance ads

We obtained 150,000 online advertisements for nearly one year through the ADBUG ad database, including the copy, landing page, delivery platform and launch date of the ad.

After the advertising data provided by ADBUG, we analyzed the landing page discovery of many online advertisements one by one. The effect of the effect advertisement mainly includes the following aspects:

  • Download App: The landing page does not lead to any product or business page, click to jump to App, App Store, other Android app stores, direct download pages, etc.;

  • Commodity promotion: Jump time limit spikes and other big promotion pages, or specific product pages and business pages;

  • Interactive Games: For users to play web games and reward them with games, prompt users to download apps or participate in promotions within the app;

And in terms of quantity distribution, the ads that lead to downloading the app account for the vast majority. It can be seen that the focus of the strategy of multi-distribution is to acquire new users and attract them to download the app.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data source: ADBUG)

We combined App Growing to count the copy of the ad landing page and found some routines they used.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Multiple ad copy word cloud, data source: App Growing)

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Jingdong advertising copywriting word cloud, data source: App Growing)

In addition to emphasizing the cheap, a lot of copywriting will also “manufacture scarcity” and use “shocked body.” Compared with Jingdong’s advertising copy, the style difference is still very obvious.

For the “management of scarcity”, we can feel a lot of advertisements from a lot of advertisements – various “only XX regions” and “limited time spikes” are endless.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Partially “manufacturing scarce” advertising copy; data source: App Growing)

The limited time zone offer is true and false, but it can’t stand the cheap, and the temptation is still very strong.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Partially “manufacturing scarce” advertising copy; data source: ADBUG)

When observing the “stunned and sighing” copy, the full screen “stunned” words came into view.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Partial “stunned” advertising copy; data source: ADBUG)

In some merchandise displays, you will be able to display one or several very attractive items, or novelty (such as drones, smart teapots), or low prices (such as 9.9 yuan). After clicking “Download”, the user will start the installation without installing the app, but will not jump to this product after the installation is completed. Even if there are already many apps before clicking, the page that jumps is just the App homepage. These ads may be testing products of interest to users or targeting stations.Search within, it is convenient to drain your own ads.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Screenshot of a lot of ads landing page)

3.2 Interactive Game Landing Pages

When counting the landing page of online advertising, we found that a lot of interactive games were set up to guide the app. Considering that the majority of the operating entities – “Shanghai Dream Information Technology Co., Ltd.” was a game company before, the gamification operation method is not surprising.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(example of interactive game advertisement landing page URL; data source: ADBUG)

At the same time, there are a lot of interactive games in the app, such as a lot of orchards, a lot of love elimination, gold pig piggy bank and so on. The rules of the game require users to interact every day and receive rewards constantly, so the games in the app pay more attention to user retention and help build a user growth system.

In addition to the application, there are many small games developed to attract users. In comparison, the games outside the App pay more attention to the instant experience, the types are casual games, such as red envelope rain, God of wealth gifts, playing the hamster. Instant sensory or reward stimuli can quickly lock the user’s attention, and then consider using reward rules to divert to the App.

We used the alternative data to predict the marketing performance of the second quarter of the fight

(Pictures for more games outside the App)

3.3 Newcomer subsidies and 10 billion subsidy conversion flows

A large number of effect ads can bring a large number of App users, but this does not mean that these users will place an order to buy. andThe many subsidies and concessions that are put together are the most direct Call to Action for these users to place orders.

As a newcomer who has just come to fight a lot, I have created a very attractive newcomer exclusive offer for you.

The information passed is also very clear: come here, it is better to spend a dollar.

The goal of this subsidy also clearly points to an important indicator of the number of active buyers (Active Buyer).

We used the alternative data to predict the marketing performance of the second quarter of the fight.

The highlight of the second subsidy is the “billion-subsidy” campaign.

A lot of practice is to subsidize branded goods and lower their prices!

For the first- and second-tier cities, the fission with the routine may not be useful. It is better to remove the routine and give the user a profit, and still have a good first impression.

We used the alternative data to predict the marketing performance of the second quarter of the fight

In order to better understand what the subsidy is and what products are more popular, we tracked the goods on the “Billions of Subsidies” page for two months, summarizing thousands of products that were successively put on shelves. According to the difference between the unit price and the subsidy price declared by the product page and the coupon amount, combined with the sales volume displayed on the page, the total amount of subsidies for the products on the page is about 880 million.

Single product subsidy = Product sales × [((Unit price – subsidized price) or coupon amount]

Total amount of subsidy = sum of all commodity subsidies

Because there is no time to cover the page on the page, there must be some omissions. Fortunately, the sales volume information is a total historical sales volume, and the sales volume that is not covered is also reflected.

From the perspective of the subsidy amount of each category, the mobile phone category is extremely difficult, and the subsidy of 440 million is far ahead, accounting for half of the total subsidy.

We used the alternative data to predict the marketing performance of the second quarter of the fight

(Data Source: Fighting a lot of “Billion Subsidies” page)

In detail, which mobile phone brand subsidies are the most? There is no doubt that Apple, the Apple mobile phone alone subsidized about 430 million. It can be seen that the subsidy for the iPhone has supported half of the activity.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: Fighting a lot of “Billion Subsidies” page)

In addition, from the price of “billion-subsidized” goods, it is consistent with the distribution of “the lower the price, the more the quantity”. Although the fight has been known for low prices, but this event did not despise high-priced goods – more than 1,000 yuan in the proportion of all SKU activities in the event more than 12%. (Most of them are electronic products, it seems to be rushing with Jingdong)

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Data Source: Fighting a lot of “Billion Subsidies” page)

One person knows that netizens have analyzed 500,000 kinds of products in the whole platform. Among them, about 3.5% of the commodity price is above 1,000 yuan, and the proportion of the “billion-subsidy” activity rose to About 12%. The increase in this proportion may mean that the target group of the event is more biased towards high-value users, which is naturally reminiscent of the argument that “to enter the first and second lines”.

We used the alternative data to predict the marketing performance of the second quarter of the fight.

(Source: Knowing – “We analyzed 500,000 pieces of multi-commodity data and came to this conclusion”)

At the same time, we also found one in the monitoring