Achieving quadruple ARPU: case study

Here at devtodev.com, we are working to ensure that our customers increase their revenue. And every case of increasing revenue is a new and interesting experience for us. Maximizing retention and ARPU - these are the two directions f2p-games may move towards almost indefinitely. There are no universal recipes, everybody has their own, but exploring the experience of colleagues is important and useful. And even if you don't use the recipes of colleagues directly (yet all the projects are different), it is possible to synthesize ideas, taking as a basis the methods that have already been checked.

Our friends from Vizor Interactive managed to increase ARPU 4 times, and today we are talking to Vitaly Moroz, one of the ideologists of these changes about how did they succeed.
 

Hello, Vitaly. Please tell us what kind of game released by Vizor Interactive, are we going to talk about today?

Hello. We are talking about the game, Kights and Brides - really social "social" game for knights and princesses. The genre is F2P-farm with separate gameplay for both men and women. It is now in the top of the Russian-speaking social networks with a total DAU for all networks more than 500 thousand people. At the time of the "soft launch", the reforms of which are about to be discussed, the total DAU was about 50 thousand people.

What does "for knights and princesses" mean? Is it a kind of farm with the integrated dating-service?

At the start, one of the main features of the game was conceived as different gameplay for men and women: women had to keep the house, and the men (knights) had to fight and protect women. We also planned to hold weddings and implement the co-management of the household, but because of the complexity of the technical implementation, it did not come to life. Initially, the balance of the game was adjusted so that the part of the female functional was blocked by the items that were available to earn only for men, and part of the upgrades for men visa versa was tied to the aid of women. It seemed that that would encourage them to interact and to invite their other halves to play the game.

It seemed? But did it turn out differently in reality?

In practice, it turned out that the other halves were invited to the game, but they never really become active players. In addition, we also expected that the amount of  women would only slightly exceed the amount of men, but in practice they appeared to be greater five times. At first, we tried to "pull" the players to the gameplay, push them to the desired behavior patterns by different restrictions and bonuses for the "correct" behavior - but it didn't work. "Farmers" gameplay worked for women for about three months, and men were bored from the endless battles (despite the fact that the battles were great) already in three weeks, because of that the superiority in the side of women increased more and more.

How did you understand the reasons, and what did you do as the result?

We started to communicate with the players, follow the statistics on the distribution by levels of the players (at what levels there was the churn), keep track of the virtual currency flow, and of course, play a lot at our own game - at the end, all that helped to form a complete picture of what was going on. After much thought, when it became clear that the original plan failed, we decided to turn around to face the players, and instead of "pulling the players to the gameplay", we made a decision to "pull the gameplay to the players": rebuild the game in the form that was desired by our users.

So does everybody get free gold?

Well, except for the free gold. We still have the free-to-play game.

Since there were a lot of changes to make, we planned a whole reform set to come to the desired state through the three updates:
 

How much time did this iteration take? How exactly this iteration improved ARPU and other metrics?

We had to give men a new big map, add the female version of the equipment and correct the balance, so all together it took quite a long time, more than 2 months. These changes at first not particularly affected the revenue, there was a growth but a small one (about 20%) and only within a few weeks after the update - first time players were waiting, silenced, and then began to withdraw the old stocks. And only when the first caution and stocks came out, the payments began to grow. We found out that we are on track, when we got the first reviews from the players and the gap between the number of men and women reduced (men started to stay longer in the game). By themselves, the reforms did not give immediate results, but improved the system, gave the stability to the game for further updates and revisions.

You say first players "silenced", what it's like?

Players almost always perceive any changes negative at first: many of them were shouting that we "have turned women into men," and then, when women were allowed to fight - that "the soldiers are made of princesses". Although it was all voluntary, if you do not want to dig beds - there's no digging, if you do not want to fight - there's no fighting. At first, the seething was rising, we had to go to the community and explain why and what we were changing. Players liked our openness, and they went to meet us. Very often when the game is changing a lot, the players start to wait, this is why we also tried to add some new content to the major updates.

Okay, thanks. So you turned around to face the players and equalized men and women rights. What was next?

After that, the game was on the right track, and we were able to run the upgrades, that began to show the noticeable financial results.

So what did you add? May be some new buildings or quests?

At first, we tried to make small updates, new buildings, new quests. We did not have tools for A / B tests, each update was a "leap of faith" - that is why we were experimenting cautiously.

And why there were no A/B testing?

Because the major part of the social games players is very sociable. To do an A/B test, so that no one noticed that something is more expensive than the neighbor has is difficult. On the other hand you can't check everything in such a way - the basic functional for the players should be the same, so that they can interact.

And what works best, new buildings or new quests?

Quests gave a quick effect, but  "burnt out" very fast, buildings gave more significant effect, but it was rather stretched in time. The combination of "building+quests" was more effective than each individually, but the best combo was "building+quests+new craft pyramids". Such kind of update already brought noticeable effect for a couple of months!

Please, tell us more about the craft pyramids. What is that?

Craft pyramid - is the chain of transformation of some items into the other ones. When entering the game there is a lot of basic items, and by combining several items of the first level you get a bit smaller number of second-level items, etc. Total at the start we had about 7 levels of about 15 basic items, which eventually turned into 3 top level items. All together it resembles a pyramid. To maintain long-term interest of the players, you may add new items of higher level in the chain, or you may gradually shift the focus, creating new craft pyramids. So we can control the demand for objects, making them more or less popular. Adding new pyramids gives better results than the old capacity improvement, because it devalues the accumulation of players - top players eventually move to the new pyramid and the remains of the warehouses are given away or sold to the players that are lagging behind, hitching them up.

 

At the basis of each craft pyramid there are some basic items, and on top is an item that I call fetish.

The sole of the pyramid (basic items) remains largely constant and expands slowly, while the center and the top are constantly changing.

The top of the craft pyramid leads to a new significant building for the player (ideally it should be beautiful and useful, facilitate some basic routine cycle for the player). The construction of new buildings outputs a lot of resources from the game.

The initial balance of our craft pyramid was very "narrow": to craft a new item it took only two other items, to climb the next step of the pyramid you needed to make just one basic routine cycle.                          

What exactly was changed?

We have "broadened" the pyramid - getting the item of the new level now required more (5-7) items of the previous level. That not only increased the "capacity" of the pyramids, but it was also well received by the players because some more challenge appeared.

At the top of our first craft pyramid were "spinning wheels", the receipt of which depended only on the players diligence (you could get as much as you wanted). As a  result, very soon the top players have accumulated huge stocks, without the understanding of how they should be output.

Somewhat it was the reason why we decided to add a new craft pyramid and placed on the top "the Seed of The Scarlet Flower". Player had to build a very beautiful building, which required to be completed with a very rare item, the receiving of which was tied to the player's luck. This eliminated the problem of the overaccumulation, but also reduced the motivation of the players.
 

So for the third craft pyramid we invented "onyxes," which took the best of the two previous types and got rid of their weaknesses. The possibility of getting an onyx depended both on luck and diligence of the player, the upgrade required a number of them, and the number of them in the game was ultimate: the players could get an onyx by opening the chests remaining after the cleaning from the large stones (the cleaning required to drain the energy or to regularly restart the mining).
 

After clearing the entire territory, players could get more onyxes than they needed - so some players could save a small surplus, if they cleaned more territories, and other - barter missing items from the other players. Some players bought onyxes directly, but the main effect was not given from the sales of onyxes, but from the expenses for the basic items for crafting new chains and spending on basic cycles - items and energy.

This updated with onyxes almost doubled our performance.

Doubled? Didn't you say that LTV has quadrupled?

The rest of the increase was reached by the other updates, little by little.

So how many times did you have to work over the balance?

There were two major reforms: the first one to equalize men and women's rights, and the second one to "broaden" the craft pyramids. As I already said, the players always negatively perceive any changes (perhaps, except for the Christmas gifts), so we tried to unite any global update with the large content-update.

And why didn’t you calculate everything properly in advance?

My management also asked me that many times!

Firstly, we had very little information at the beginning - there is no chance to do the best, when you do not know yet what is good and what is bad. The more experience accumulated, the better ideas appeared.

And secondly, I could not come up with a convenient presentation of data to count everything clearly. The result was something like the system of accounts,  where each account is the craft item, craft "card" is  how many craft items it includes, their price, how much the player will receive and how much will the item cost. And all these "craft cards" were sorted by the levels and timings - it went really well, but only in six months after the launch.

Did you try any other experiments with monetization besides the craft?

Of course, we did not limit ourselves to the experiments with craft only, although it proved to be the main driver of our game. In particular, there was an interesting case, after the introduction of roulette in the game - it increased the turnover of crystals three times, but the actual payments increased only by 5-10%.

For a few days we were confused: we saw the growth at the currency inflow chart, and the growth at the currency flow chart, but the payments were barely growing. Then it dawned on me, that we have almost a win-win roulette, where the player spends crystals, but instead gets something that he used to buy for the same crystals. So the turnover of the roulette is high, as the real effect is small, it turned out to be redistribution of the crystals between the different sources of goods. Ie the roulette became a kind of a black box and we had no clue what was going on inside - we could monitor the input (how much crystals the players have merged) and we could only roughly estimate what they got at the output. It took us three more updates to set up the roulette to the point where it worked right.

We also had some more interesting experiments with energy.

It is very interesting as it is often asked about the energy, and there is still no single opinion about what to do with it.

You're right. I was surprised to come across the view that "energy is not for sale" - I think that these people just do not know how to sell it, and use it as the the session limit rather than the major basic resource.

We cranked an interesting trick with energy (inspired by the idea of the article by Oleg Yakubenkov) - began to offer with the maximum package "+1 forever" - a questionable bonus. In order to get that bonus, players were required to save as much energy as it was impossible to merge, but the trick worked: after this innovation the maximum package of energy climbed from third place to the top spot.

By the way, in my opinion, energy proved to be undervalued, cheaper than the optimum, in the inelastic area of the demand - it happened historically. At first we had low demand for the energy, and when we set it up and increased it, we were afraid to raise the prices for energy since the game already had an operating history, and we didn't want to upset our users.

When released the game at the new markets, we corrected that, and the proportion of energy in the overall costs of the players there became a little higher, but the effect is difficult to be estimated directly.
 

So, if to sum up, what advice would you give?

When setting up the monetization, the most important thing is to customize the output of game currency (or what it is converted to) - if a player has a need, he will cover the deficit at the expense of payments.

And we must be prepared for the fact that the first updates will not give result - the game has a lot of local optima combination of gameplay and balance, and some of these optima are very small and settings within them do not make sense - you need to have the courage and perseverance to move the game to the new optimum with the great potential and get to it.

Thank you for the interview, Vitaly! We wish you a further increase in ARPU, say, ten times.

Say twenty, why to limit yourself!