Pumping the project from Soft Launch to $ 1 million in revenue per month

Hello, Habr! I am in charge of the traffic purchasing department for mobile and desktop projects at Mail.ru Group. Today in this post I will talk in detail about traffic for games. It is based on my presentation at The Big Deal. But first, a little about yourself. I worked with a huge number of projects on various marketing tasks, the last year and a half I work exclusively with mobile applications. During this time, he managed to work with several dozen applications, most of which were in soft launch, some of them were released in global, and several of them “shook” more than $ 1 million turnover per month. During my work, I developed a certain vision of how to analyze metrics from the point of view of project traffic and see if your project can grow, develop and pass a huge amount of traffic through itself.



I would divide this whole thorny path into 4 stages:

1. The first stage of soft launch . During it, we evaluate whether our game is played. There is a basic funnel, creatives, application stores.
   
   
2. The second stage is soft launch . We are checking monetization. We are trying to understand how paying users are held, and we do the first LTV calculations.
   
   
3. The third stage - we pass to global launch . The first test traffic purchase is completed. We continue to recount LTV and optimize traffic sources.
   
   
4. The last stage is the search for growth points .
   
   

Let's consider them in more detail.

Soft-launch # 1

Many come to us with projects, saying that they just have a hit, the whole family plays, friends play, etc. But they do not know the metrics and ask to check the project. The very first task is to understand whether other people are interested in playing your game. The first metrics are simple: understand the retention of the first and seventh day, as well as the CTR and CR that go into IR. Buying traffic is less risky in cheap countries: Eastern Europe, Philippines, CIS, Thailand. It starts with Facebook / Unity. Budget - 5-20 thousand dollars. It all depends on how many iterations will be needed for the developers. You can check the first metrics, optimize and get an idea of ​​the target audience for less money. But the developer often needs a larger audience, while still checking a huge number of product metrics.



In terms of time it takes from 2 weeks to a month or more. It is important to note that when purchasing software the question always arises: do we buy good traffic? To answer this question, you need to look at the organics - these are the users who managed to find your game in the store, install and play on their own. Their quality can be considered a certain standard: if the purchase is at the same level or higher in retention, then you are in your target audience.



Let's look at specific metrics. At the start, the purchase comes with Facebook with optimization on the installation. What are we paying attention to? Important CTR, CR, IR. CR is a conversion in the store, CTR is your creative. It turns out IR. It is very important to monitor and edit these metrics. Casuals may have more IR due to CTR, CR. More hardcore projects have a bit less. But, roughly speaking, if you have an IR of 0.44%, then this is a working draft, because you can buy fairly cheap traffic. If your IR is less, then you have big problems. Monetization should help here, or you need to edit CTR, CR.



The chart below shows the retention of players. Everyone knows the average numbers: 40, 20, 10.







So, players like everything. The application works, there are no major bugs, tracking works. Move on. The developer says: “With monetization, everything is fine. Everything is perfect. Money is flowing now. Let's test. ”

Soft-launch # 2

What are we testing in terms of monetization? There are more metrics here, let's look at the most important ones.



A cohort percentage of payers on a daily basis up to 30 days, ARPPU (average revenue per one paying) on a daily basis up to 30 days - these are the usual metrics.



CPPU is the cost of the paying user. The cost of attracting a paying user must be monitored from the very beginning of the purchase, especially in paying countries.



LTV forecast . As soon as you gain an audience of about 100-200 paying users and 500 payments within a month, you can build the first forecasts that you can focus on.



ARPDAU (average revenue per unique user who logged into the application at least once a day) regarding your key monetization by inaps (in-app purchases) and advertising. When you build this metric, it is important to understand its essence and how it is calculated by users, otherwise you can make erroneous conclusions.



Share retention of paying users . It is already important here how your paying users play.



As for countries, their choice and quantity depend on budgets and on your idea of ​​the game. There are games that, for example, do not enter Asia, and it makes no sense to test them there. According to traffic sources: Facebook / Unity will require a budget of about 30-100 thousand dollars. I would advise starting with Facebook to accumulate events that will work on your advertising account.



Dates for this stage are a month or more. There are a huge number of tasks associated with the product, which are not solved quickly, go through several iterations.







If you have indicators and IR such as on the graph, then this is a normal working situation. If IR is less, then you need to be higher than these values, if IR is greater, then you can not reach them. It all depends on the project.







An interesting graph related to LTV and the cost of a paying user. We have a sufficient amount of real data for 2 months, so we can build LTV and predict its trend throughout the year. This project will earn more - a year, two, three, depending on the preparation of the content.



During soft launch, every paying user is quite expensive for us. But at the global launch stage, there is an optimization on payments and paying users, that is, we lower the price of the paying user. A decrease of 1.5-2 times from 120 to 70, which you see on the graph, is present in each project. Therefore, you can immediately guess how much the paying user will cost.



Consider the schedule of a paying LTV user. You understand where you will be. Take, for example, the level of $ 70, it is reached in six months. According to our objectives in the campaign, we must recoup the traffic within six months, and then earn. This problem is being solved here, which is already an indicator that the project can be “uploaded”.







Let's move on to more sensitive indicators, such as ARPDAU on the main and advertising monetization. And for an example we will also analyze the real situation. There are 4 projects, three of them are seriously moving forward, and the fourth is a test one, it is yet to be released. The graph shows its main and advertising monetization. When building your primary monetization schedule, it’s important to understand the following: ARPDAU is your streaming income divided by your audience’s DAU. That is, this is an indicator of how your product monetizes your game audience every day.



Where to get traffic from? Better if it is organic. If you take paying traffic, or one of your projects is optimizing for paying users, then ARPDAU will immediately become higher. Most often I take organic matter, because it’s plus or minus random, average traffic. It is also important in organics that there are no features in application stores. As soon as they appear - you are leaked to users who do not pay, and ARPDAU again falls.



Another important point. When building ARPDAU, you have a daily audience update. If it happens at about the same volume, then the charts can be trusted, because a new audience can pay better than the existing one. It is also important to remove the effect of stocks. In advertising, everything is quite linear. You can take Tier1 and the same countries.



Here are ARPDAU examples of advertising monetization in real projects:







Developers like to say that their advertising is very well integrated, they always have everything perfect. But, having analyzed their indicators, you understand that you need to raise it again by five in order to reach the working level.



Now let's analyze paying users. It is very important to look at their retention. Often there are times when the developer looks at his retention and does not understand what is happening: it was 50, then 40, then 30. How can this happen?



Traffic can be different, but the essence remains the same: projects earn on paying users. If you want to see the real situation, you should look only at them.

A small checklist that you need to pay attention to when entering the global launch:

1. Promotional materials. We must be sure of their quality, as well as that everything is fine with our application store. We have IR either at the market level or higher.
   
   
2. Basic monetization is at the working level for its genre.
   
   
3. Advertising monetization is working.
   
   
4. For paying users, everything is in order and stable.
   
   
5. LTV built and most likely it will be more than the cost of paying users.
   
   

If for each of the listed items you can put a plus for your project, everything will work.

Calculation and application of LTV

Most often, LTV is perceived as direct income, but in fact, the formula depends on many factors: how the project starts, how it develops, what stages it enters. If you have a purchasing strategy from paying users, then the main income is inaps. Further, from above, you can generate 15-25% advertising. It is forecasted in the same way as LTV. It is necessary to divide LTV in organic terms by income and by advertising. Most often, organic matter gives about 10-20%. And organic advertising is another 5%. If you have cross-promotion, this can also add about 5% to your LTV.



It is very important to properly add organic to commercial LTV. You take it all and forecast for every day, in each country, and then erode the organic regarding commercial traffic. There are also two options:

• By volume of purchase . If you have a large purchasing team, some of them may begin to cheat - to buy traffic in which no one pays, but which takes organic matter.
  
  
• In terms of the volume of payments for commercial traffic, in this case, you give out bonuses just to those campaigns that give real paying users, both in inaps and in advertising.
  
  

Organics can be added only after reaching the payback directly, when there is a task to increase volumes, and you are ready to pay directly for a longer period. Most often, if you look at the correlation between paid traffic and organic, it will be 10-20%. It all works when your payback strategy from inaps.



But you can also try a buying strategy from advertising . When you have ads at the working level, you can afford to buy cheap users, look for bundles with good devices. At the same time, advertising will close about 50%. Another 50-60% can be added inap. Thus, you begin to go to the rate of 100+%. But important is good advertising integration.



A little more about how you can calculate LTV. There are three key approaches:

• Logarithm function . A flexible approach, every analyst can use it. For good accuracy, you need to collect about 500 payments in 3-4 weeks.
  
  
  
  
  
  
• The model for actual payments . Usually enough buying 100-200 paying users. In the future, when the project is gaining momentum, you have a large amount of data. Six months or a year passes, you build models and begin to understand how your user behaves throughout the year. Most likely, new users will behave the same way if you do not have major changes in monetization.
  
  
  
  
• A more complex approach, often works on very large projects: individual forecasts , in which they analyze not only payments, but also game inputs, training, and many different events. This is work with large volumes of data, for which specialists are needed. And not the fact that the accuracy will be higher than in the second option.
  
  

When calculating LTV, it is important to look at the schedule for all users, not just paying ones. This will help in optimizing based on what strategy you use. Also, LTV can additionally assist in business planning. The entire business plan can be built from LTV. You will accumulate user income on simple numbers without a lot of add-ons, such as retention, ARPPU, interest payers. Analysts like to wind up 20 indicators and think - “what is really happening?”



LTV already includes a huge number of nuances of user behavior, so business plans can be very accurate.

Global-launch # 1

So, we can count LTV and we know that monetization is up to standard. Have a good CTR / CR. We debugged production. New videos are released every week. We do many tests. Everything is fine, and we are confident that the paying user’s LTV will be more than the amount we paid for it.



Now you can try to exit to the global launch. Our goal is simple: to earn more on LTV of a paying user than we spent on it, and to reach daily working volumes. Testing all over the world. Sources: Facebook / Google / Unity, budget: $ 100-500 thousand, terms: two weeks-a month.



Separately, I will explain the purchase of traffic. Purchasing is not a button, after which everything is fine. This is an ongoing process of experimentation and analysis of a significant amount of data. Next is the training of platform algorithms, Facebook / Google / Unity. At the beginning of the purchase, if you have not pumped your monetization and have not the highest IR, most likely the purchase will be difficult. You will need to look for working ligaments. In the beginning you have to spend money. If the specified parameters are pumped well, then there is a chance at the start to collect cream. In this case, the costs pay off fairly quickly. The source of traffic with the most minimal risks at the start is Facebook. It is quickly optimized and has great functionality.

Optimization - Key Metrics

A little about the indicators that need to be considered.



The cost of the paying user . The most important thing is people who bring you money. You must consider how much you paid for them. This is the basis for unique “paying” metrics that are calculated over a period of 1-30 days.



The number of payments that users make in the same period. The most important are the first 7 days. During this time, most optimizations are performed. You must understand how much you are willing to spend on each payment and on a unique user who will appear from 1 to 7 days. It should be understood that after 7 days other users will appear.



It is important to know your ROAS funnel up to 7 days. It is advisable to watch it up to 30 days.

Traffic sources

Key sources - Facebook / Google / Unity, they need to take about 80-90% of the traffic. It is important to swing them at the start because of the huge opportunities for optimization. At these sites, everything is prepared in order to earn money.



When you realize that you have reached some ceiling, you can add new sources. There are a lot of them, but I highlight several: Applovin, Chartboost, Adcolony, Vungle, IronSource. In each new source, you need to collect your statistics, understand what works and what doesn’t. You will have to spend money again before you collect this data. Their optimization is not the best. Therefore, you will accumulate about 10-20% of your main volumes on them.

Instead of totals: global launch # 2

So, with the project, everything is fine. We have reached confident volumes, we believe we are preparing materials. It's time to move on to the next step. You need to understand how to increase profit by two to three times. Everyone asks this question. I highlighted several points that help with the purchase of traffic.

1. The simplest thing is to create promotional materials , static, video. Playable are now becoming very popular on Facebook / Google. Keep updating ads. If you made 5 videos, 5 pictures - this is not enough.
   
   
   
   With large volumes, it is important for you to constantly create advertising materials that will raise your IR by 10-20%, or maybe you are lucky and you will get a "star creative" that will provide you with a double increase. At the same time, each of your events becomes cheaper. The algorithm estimates the cost reduction and buys more traffic. The audience grows and immediately becomes effective for you. Then you realize that you have reached the limits of ROI. You can reduce them and try to increase volumes at times. Just the same type of video gameplay doesn’t work, look for different ways and ideas, hire people who will create promotional materials without stopping.
2. Testing new functionality in key sources . This is in most cases a gold mine. Sometimes you have to spend money on them. But when you first use the new functionality, then most often there appears a new inventory. There are fewer competitors, and you buy back a large share of the effective traffic.
3. Constantly recalculate forecasts . For example, you want to increase volumes. If your business plans and outlooks on life allow you to postpone big money for a later time, then you can take not 6 months, but 9. This can also give a transition to a new volume of purchases.
4. A significant change in monetization . Each developer is trying to get more money. Monetization is slowly changing, and so is LTV. When you recount, LTV grows.
5. Stocks . If you spend them constantly, it increases ARPPU /% pu and changes LTV.
6. The new strategy we are trying to use is payback from advertising . We buy huge volumes, we pay back 50-60% of advertising and add inaps from above.
7. Asian markets . We understand that the Asian market has a huge paying audience, but so far the volumes from public sources are small. Therefore, you need to go to local advertising networks and look for partners there.
8. It is necessary to test new sources such as Snapchat, Twitter. While the purchase of traffic there occur with varying success. I am sure that they will develop functionality, but it will also need to be tested. TikTok is a trend, in the near future they will release their dashboard for the USA and Europe. We will be happy to purchase from them.