27
Deconstruct Clash of Clans Andrew (Drew) McInnes [email protected] 408 455 1090

Drew McInnes_Deconstruct_Clash of Clans

Embed Size (px)

Citation preview

Deconstruct Clash of Clans

Andrew (Drew) McInnes [email protected]

408 455 1090

  2

TABLE OF CONTENTS

OVERVIEW 3

THE GAME 4

THE CORE LOOP 4

WHAT WORKS WELL 5

FUN PAIN CREATES TOP GROSSING GAME 5 SUPERIOR REVENUE STRATEGY 8 APPOINTMENT MECHANIC (THE GOOD) 9

WHAT DOESN’T WORK 11

APPOINTMENT MECHANIC (THE BAD) 11 VIRALITY 13

PRODUCT PRIORITIES 14

FEATURE RECOMMENDATIONS (SOCIAL) 15 FEATURE RECOMMENDATIONS (OTHER) 17

NEW FEATURE EVALUATION 19

CHOOSING WHICH FEATURE TO IMPLEMENT FIRST 19 EVALUATING IMPACT OF SELECTED FEATURE 20

  3

OVERVIEW Clash of Clans (CoC) is my favorite game to play. So much so that my wife says that I am addicted to it. To which my credit card statement says that she’s right (hate when that happens – lol). But why is it that I’m on the fast track to become a “whale” of a spender in this otherwise innocuous free to play game (FTP)? Welcome to what I look forward to learning in exploring:

• Why the game is successful • How could the in-game virality be made stronger • What are the pros and cons of the appointment mechanic • Why the game’s revenue strategy is superior to that of other games • What should be the next five priority features on the product roadmap

Each question speaks to what works in the game, and what can be better. Yes, I said it. Even though CoC is the best in class strategy game out there today as measured by its top billing among all grossing games, I believe that it’s not without its flaws. At least, when I look at it through the eyes of a product manager and behavior designer, but more importantly as an avid gamer both since the 1980s (think Intellivision) and now as a “true fan” of CoC (level 50). Let me explain what I mean by first speaking to what works well in the game. Then to what doesn’t work so well. Before talking about some possible features to include in the roadmap. One of which I’ll drill down on to understand it’s potential impact on revenue, retention, and engineering. First up though let me touch on what the game is all about and how the core loop works in order to frame in the discussion as a whole. Please turn to the next page.

  4

THE GAME As you may know, CoC is a FTP, real time strategy game. Where you harvest resources, build an army and a village, and direct your troops in battling with other players or the computer. Battles afford you the opportunity to secure additional resources for building your village: Screenshot of My Village

THE CORE LOOP When you play the game, you engage in one of three core activities: collecting resources, building and training, and battling. Not all of the activities are equally important though. The importance of each activity is influenced by your ongoing goal in the game. Creating different game-play styles.

Image Credit: Deconstructor of Fun

  5

WHAT WORKS WELL

Fun Pain Creates Top Grossing Game Ultimately what I believe makes CoC top grossing is the game’s use of behavior design to create a fun pain that drives very solid retention and a whale economy. Where ARPPU of retained players rises as average purchases increase in response to rising costs from making progress in the game. The game’s emphasis on playability over everything else also makes CoC accessible to casual gamers without alienating hardcore ones, who will really appreciate the depth beneath the surface. Thanks to a complexity of layers created through the game’s balanced loop, compelling graphics, and smooth gameplay that allow a broad audience to create an emotional attachment to the game. If interested in understanding what I mean by fun pain, I offer up an explanation below. Before walking through an example of how CoC creates fun pain to breed retention and revenue success. Saving my thoughts on mechanics/features that I like for driving playability for a future discussion. What Fun Pain Is As you know, in order to have a successful FTP game, you need to get players into the game (acquisition), keep them in the game (retention), and get them to enjoy the game enough that they spend money on things that they value (monetization). All of which is hard to do without fun. But you can have a fun game that doesn’t monetize. Look no further than FishVille. Where 60% of the players found it fun to feed their fish everyday, but never felt any pain in doing so. As a result, they never purchased anything. Forcing Zynga to shut the game down this past December. On the flip side, you can also have a game that creates pain, but doesn’t monetize because it’s not fun. If only there were a way to create what Roger Dickey (Mafia Wars Founder) calls a fun pain in order to make it easier to acquire, retain, and monetize players. Something that is simultaneously entertaining yet a little bit annoying. Like how Farmville players need to click squares to plant or harvest crops. Giving Zynga a

  6

chance to upsell players on pain-reducing items like tractors, which save players from a lot of clicking and provides Zynga with a significant amount of revenue. How to Create Fun Pain that Breeds Success Now incorporating fun pain into a game is both art and science. That CoC takes to a whole another level as seen in the game’s battle play mode. Where CoC uses four features to create, elevate, and remedy a fun pain that I believe is what drives the game’s retention and revenue success. Starting with the use of 3D-View and Fight-Replay features, CoC creates an extraordinary fun pain by allowing a player to watch his battle unfold from start to finish. Knowing that when a human being can watch any form of competition that he’s an active participant in unfold from start to finish; anticipation, anxiety, fear, and elation come bursting out uncontrolled with every twist and turn in the competition. Creating an emotional rollercoaster that is both exciting and unpredictable, and stems from higher levels of dopamine now surging in the player’s brain in anticipation of a reward. Whether that reward for winning a CoC battle will be more trophies, gold, elixir, or a combination thereof is unknown to the player, and this uncertainty is intentionally baked into the game. Make no mistake about this. Designers of CoC and other games know that when you introduce a schedule of variable rewards in any form of competition you can maintain a player’s dopamine levels at an elevated state, but this is where CoC’s designers innovated on this standard metal hack in games. By allowing a player to see every twist and turn in his battles using 3D Views and Battle Replays, CoC’s designers can actually elevate, not just maintain, that player’s dopamine levels. Providing for an even greater multiplier effect of this standard mental hack in games than what you already get from introducing a schedule of variable rewards. The result of which is longer periods of player engagement with this core gameplay and other gameplay modes. As we know, longer periods of engagement often lead players to create preferences, and preferences generate in-app purchases. In CoC’s case, players just can’t help themselves but to stay engaged because the “anticipation is killing them.” When a player says this it signals to a behavior designer that the levels of dopamine in the player’s brain surpassed the point of chemically triggering the brain to enter a semi stressful state called desire. Where the brain is now in a frenzied hunting state that suppresses areas most

  7

closely associated with judgment and reason, while activating parts more closely associated with want and desire. This is where the Next Battle button now takes over the job of driving retention. Let me explain by saying that now that the player’s brain is immersed in this “stress of desire,” the player’s natural instinct is to seek relief from this fun pain. Leading the player to jump on the Next Battle button in search of the perfect village to plunder, but this doesn’t actually provide relief. If anything, taking this action actually increases his stress of desire. You see our dopamine system doesn’t work to provide us with rewards for our efforts, but to keep us in search mode. So before our hypothetical player knows it, he’s spent 20+ minutes and $10K in gold clicking the Next Battle button again and again. Digging only a deeper hole that he will have to crawl out of to escape this fun pain, and only made deeper when he finds a perfect village to attack, yet loses the battle. Welcome to the final aspect of CoC’s fun pain. When a player loses a battle or fails to seize the reward that he anticipated winning, his stress of desire is fast replaced by another wave emotion that washes over him. That emotion is literally a tinge of physical pain. Activated by the same area of the brain associated with physical loss and now triggers the brain to both feel the pain and rid the body of this mental agony of defeat. Great just when the player escapes the clutches of the stress of desire he’s greeted by physical pain. Fortunately, CoC offers the perfect feature for vanquishing this final element of the game’s extraordinary fun pain, and from a behavior economic perspective, it’s what I believe drives a fair portion of CoC’s revenue. Finish Training is the fastest way for the player to get the troops that he needs in order to get back into battle and win one. Winning a battle after a defeat turns off the part of the brain associated with physical loss and allows the brain’s natural “fight or flight” mentality to take over. Willing to do anything to admonish the agony of defeat. Even if it means slapping down $4.99 for 500 gems in order to use the Finish Training feature. The brain will do whatever it takes to get off this emotional rollercoaster – well at least until it’s ready to experience this extraordinary fun pain again. Now this is how you manufacture desire that keeps free players engaged and provides room for spending by whales. Bringing me to your question about monetization strategy.

  8

Superior Revenue Strategy I believe that CoC’s revenue strategy is superior to that of other empire-building games because it’s progression based. Meaning that prices only increase as a player makes progress in the game. Thereby creating a situation in which increases in ARPPU go hand in hand with retention. Please read on for a more detailed discussion of why I think this. Taking a player progression driven approach to monetization is superior because it helps CoC to better retain players for months and months and even years. Especially with the game’s decision to only make new content available to loyal players. Providing for, you guessed it, another fun pain. In which a player’s dopamine system sets the brain aflutter again with anticipation of what new content (i.e. variable rewards) that this player will be able to unlock and interact with next. Compare this to a game like Kixeye’s Backyard Monsters. Which is quite successful in monetizing on the desire for revenge, but some players say that their desire for revenge starts to wane knowing what the likely outcome will be. Leading them to stop playing the game altogether having only picked up the game two months prior. This is code for their anticipation “stopped killing them.” Another reason that this progression based monetization strategy is superior is because it doesn’t cripple the experience for those players that don’t pay like other empire-building games do. Zynga I’m looking at you. Where in Coasterville I find it really tough to progress beyond level five without having to part the “Red Seas” of my wallet. To which I say that if you’re making a FTP game, make sure it’s really free. I recognize that making a FTP game requires a change in mindset from that of the bygone days of say paid console games, but if you can make the shift, you stand to capture the consumer and producer surplus you’re leaving on the table by not making your game truly FTP. Granted that in CoC there is a lot of waiting around that can cripple the experience. As you wait for new armies to be trained or upgrades to finish once you get to level twenty. At that point, the game truly revolves around the concept of “time is money,” but if you have time-consuming actions, the game never forces you to buy gems to speed up their completion. Instead, CoC allows you to do other things and sends you a notification when your actions are completed. This in turn makes the players who I talked to feel less anxious about

  9

playing the game, perceiving that no one is trying to swindle them out of their money. Nice way to create “true fans.” Nonetheless, let’s take a closer look at this “time is money” concept by looking at the appointment mechanic in the game next.

Appointment Mechanic (The Good) In my experience with the game, I find that the appointment mechanic does some things well and other things not so well. Please read on for the “good,” and see the next the section for the “bad.” We know that appointment mechanics are a powerful way to retain and engage players, and these mechanics work best when players can schedule their own appointments. Players who can choose their own appointment schedule are much more likely to come back each day and remain engaged with the game. But how do you retain and engage players when the time in between appointments grows longer and longer because of the game’s use of a progression based monetization strategy? Welcome to the question that CoC addresses in quite a meaningful way. Let me explain how. Meaningful Downtime Mechanic As an avid gamer I typically find that games relying on appointment mechanics as part of their way to keep me engaged have trouble addressing the downtime that arises in between appointments. Struggling to figure out how to best keep me entertained during these periods of time given that the most engaging gameplay and core mechanics are intertwined with these downtime-creating appointment mechanics. For instance, in Tiny Monsters, the only thing that I can do as I wait for my tiny monsters to breed and hatch is to re-arrange my village – really shallow gameplay. In this sense, CoC is no different, and actually faces a greater challenge in keeping players like me engaged during what are now longer periods of downtime. All because of how the game punishes us through its progression based monetization strategy, which dictates that players who are further along in the game have to wait longer and longer periods of time for upgrades of

  10

buildings and troops to be completed. Where the amount of time that I once waited to say upgrade my cannon was fifteen minutes, but now I’m at a stage in the game where upgrading the same cannon takes two days. Leaving me to twiddle my thumbs for those two days wondering what else can I do. How CoC addresses this even greater challenge in quite a meaningful way is by making the composition of my village not only vital to my success, but also a downtime session for moving resources around. So that I can focus on better defending against attacks from other players who are constantly exposing holes in my latest and greatest layout, and have access to tools for doing so. Like using Fight Replay to watch and learn from these defeats. As well as ask other players for suggestions about how to make my defense layout better. Feeding inside of me an obsession with creating an airtight defense to protect my stores of gold and elixir. This isn’t just a fantastic downtime mechanic, but a fantastic gameplay mechanic period! In that I’m now playing additional sessions of the game that are even longer and more fun. As I invest several hours at a time engaged in this meta-level of gameplay where I’m revamping my village literally one stone at a time. Heck, this mechanic has even gotten me to make in-app purchases. After I see that having a level seven cannon will not be sufficient in my new layout, I paid to accelerate the upgrading of the cannon to level eight to fend off attacks that are sure to occur while I sleep. What great care and attention that went into creating this mechanic, but the appointment mechanic can be better. I discuss how so in the next section on the next page.

  11

WHAT DOESN’T WORK

Appointment Mechanic (The Bad) Again, appointment mechanics are a powerful way to retain and engage players and work best when players can schedule their own appointments. I find this difficult to do in CoC, however. Now that I’ve reached a higher level in the game where upgrades of buildings and troops require several days to complete, and no longer can be done in minutes and hours. Making me feel frustrated by my inability to play the game on a daily basis like I once did and enjoyed doing very much. Yes, I too have spent my downtime in between multi-day upgrades obsessing over how to perfect my defense layout (see Appointment Mechanic (The Good) under What Works), but I can only obsess so much in this meta-level game play while waiting for my buildings and troops to finish upgrading. Before I become anxious to find something else to do in the game to pass the time. Ultimately, I want to get back to engaging in the game’s fun pain of attacking other villages (see Fun Pain under Successful Game under What Works), but what inevitably happens when I attack other villages as an alternative to obsessing over my defense layout is a huge waste of time and effort. In that after I invest my precious time to attack a village someone attacks me (not a revenge attack). Seizing the gold, elixir, and trophies that I secured in my raid. Putting me back at square one, where I’m still waiting for my upgrades to finish. This is truly frustrating – my motivation to play the game is so high, but playing the game on a daily basis is now way too hard to do! One idea I have for ridding others and myself of this frustration is to make the game’s appointment mechanic better. By improving upon the current “all or nothing” upgrade functionality. Where a player today must accept having to wait say the four days required to upgrade his Wizard Tower from level three to four, or not. If “not,” then he can’t upgrade his Wizard Tower at all. What if however we incorporated 1-2 other features so we no longer punish players for playing? Boost

  12

The first feature could be a standard “boost” mechanic. Where the player can shorten his upgrade time for that Wizard Tower from four to two days. By paying 400 gems to only have to wait two days instead of 800 gems to avoid waiting four days altogether.

Pay As You Go What would be even cooler is if the player could pay as he goes. That is to say, he could pay for the four-day upgrade in say twelve-hour increments. So instead of paying the full 1,280,000 in gold required to upgrade that Wizard Tower to level four, the player would make eight installment payments of 160,000 each at anytime of his choosing, but for game-balancing reasons, he would not realize the increase in “damage per second” and “hit points” from his upgraded Wizard Tower until he completes the entire upgrade. Why this would be very cool is that in the matter of five minutes twice a day I could enjoy the fun of attacking a village and then the sense of progression from plowing my winnings into upgrading my buildings and troops on a time scale that works for me. Where I play five minutes in the morning while on the train and five minutes in the evening before heading to bed. Thereby allowing me to get back to attacking and upgrading on a daily basis, and as part of a set routine when I play mobile games. Where I’m like everyone else in the sense that I mostly play my mobile games on public transport in the morning and right before bed in the evening as this survey shows. It’s really a tragedy that CoC is not capitalizing on this golden opportunity to help higher-level players like me to get back into playing the game on a daily basis like we want to be doing. Now I should mention one important design and development caveat here. My proposed Pay As You Go feature would likely require a greater amount of design and development resources than those that would be required to deliver my Boost feature. Nevertheless, my hypothesis is that the return on Pay As You Go would more than justify this greater amount of required investment. As measured by the increase in the number of retained and monetized users at later stages of the game. That is to say, CoC would be able to drive higher volumes of revenue from now having an appointment mechanic that can actually capture the game’s producer surplus of “minnows” and “dolphins” like me. Who are willing to pay something to reduce upgrade times because we want to get back to playing the game on a daily basis, but are less likely to pay the “full upgrade cost” like a “whale” of a player would in order to keep on playing the game everyday.

  13

Virality Now another aspect of the game that could be made better includes the game’s virality features. Where I see room for improvement in the clan and friend features that CoC uses to drive in game virality, and others and I find to be disappointing in making our game experience more fun and pleasurable. We want to have better ways to play together in the game starting through the clans. Clans Today there’s no real upside for becoming a member of a clan. Unless you count getting a few low-quality reinforcement troops for your next attack or defense worthwhile. Or you’re one of the top ten performers in the one clan that gets a reward each week for wreaking the most damage among all clans. So I guess clans could be fun for you, but not for the vast majority of other players like me. What I believe would make clans more attractive to join for the vast majority of us “non top decile” performers is providing for more interaction between clans. Say through clan tournaments, or even quest systems. Where two clans are paired together to achieve another goal other than amassing as many trophies as they can. Like destroy 1,000 builder huts in a week and receive 20 gems. This could require the clans’ members to interact on a daily basis in order to coordinate everyone’s efforts toward achieving the goal. Anything that would get more interaction going on between clans sure would be a step in the right direction from an in game virality perspective. Friends Even CoC’s team acknowledges the importance of increasing interaction between players, having said, “the most important reason CoC players keep on returning to the game is other players.” Players want more ways to play together. So you can imagine how disappointed a player is when he connects to the game through Facebook or Game Center, only to be able to see which friends are playing the game and visit their clans, but can’t attack, message, and gift any friends. Nor invite non-playing friends to play. What kind of in game virality is that – definitely room for improvement! Map

  14

Or how about including a map in the game? Like the one that Kixeye added to War Commander last fall after having success with it in Backyard Monsters. As this article says, “adding a world map in [War Commander] led to a major lift in retention and monetization.” Makes sense to me. Having used the map in Backyard Monsters to create a few neighboring rivalries for myself when I played that game and that wouldn’t have come about if not for a map like the one below. The rivalries sure kept me playing and spreading the good word about the game to several people. Here’s an example of the Backyard Monster map.

So why not test this feature in CoC as another way to increase interaction among clan members. Where they can now create rivalries by picking which clan to do battle with as part of a tournament. Or if CoC ever lets friends attack other friends and rivals, then this map is sure to get me going too. Now that we have a sense of what doesn’t work well in the game as well as what does work, let’s talk about how to make the game better.

PRODUCT PRIORITIES Without knowing the business needs of CoC, or having access to data to understand these needs, I believe that the game would stand to benefit from delivering a better system of social rewards through the game’s clan feature. Bridging the gap between the rewards of competition and cooperation the feature only scratches the surface in delivering today and CoC’s acknowledgment that the most important reason players return to the game are for the rewards of the tribe.

  15

Before continuing on, I should say that how I design and prioritize new features typically starts with understanding the needs of the business. Then translating the business needs into product needs. Before working with design to translate the product needs into features; and in the case of games, I would also work with design to fit these new features into the game’s fiction space. So they make sense to players. Finally, I would take this second cut of priority features and work with the team to finalize their priority based upon the cost and time to develop each feature.

Feature Recommendations (Social) Now I took this approach in coming up with three priority features for delivering a better system of social rewards, even though I had to improvise at certain stages in the process: Clan Challenge Through this first feature, two clans can challenge one another to a one-week competition. Where only the clan leaders can accept or decline the challenge. If accepted, then whenever members of the clans were to select the attack button, they would attack a village from their rival clan. Now whichever clan has the greater number of trophies at the end of the week would be declared the winner. With each member of the winning clan receiving 100 XP and 50 gems from the losing clan. Of course simulations would need to be performed to ensure proper game balancing. This feature marries the tribal rewards of competition and cooperation nicely in driving higher levels of retention and monetization. Providing another economic sink for gems that could produce revenue from losing clan members who must buy gems to cover their portion of the collective lost. Not to mention any gem purchases made by either team during the competition, even if the amount purchased exceeds that of the reward. Never underestimate the behavioral economics of rivalries. Finally, different metrics could be used to iteratively test and build this feature, which can reside in the clan chat page and leverage the same functionality that already exists for donating troops.

  16

Clan Map Incorporating a map of all of the clans could then drive even higher levels of retention through the Clan Challenge feature. By allowing clan members to track their collective progress toward conquering neighboring clans on the map in their quest for more purposeful, tribal gameplay. With the map also helping to drive increased amounts of social interaction on the Clan Chat page. All of which further feeds the player’s sense of rivalry and camaraderie that in turn drives retention. Dueling Members Clan members also want to interact more around how to build defenses or pursue attacks, but doing so today is just way too hard. Given that the only tool clan members have to engage in this kind of interaction is the Clan Chat. Yet providing this kind of advice strictly by text just doesn’t work. Players need to incorporate visuals into this form of advice giving in order to be effective. The best solution that exists today for engaging in this kind of interaction is YouTube, which offers countless videos on how to defend and attack your CoC village. The problem though is that these videos are only a one-way communication medium. Making it difficult for a player to gain clarity around why whoever created the video made the defense and attack decisions that he made. The good news though is that this YouTube activity shows that people clearly want to engage in this type of interaction. Demonstrating a high likelihood that if CoC were to develop a feature that provided two-way communication, the return on development dollars would be meaningful. One way for improving on the YouTube solution is by developing the Dueling Members feature. In which a clan member can challenge another member of the clan to attack his village, and if the challenge is accepted, the attack is recorded using Fight Replay. Then the clan member who was attacked can watch the video and ask questions afterward using the Clan Chat page. Dueling Members is a step in the right direction. Delivering the tribal rewards of competition and cooperation that clan members want to have with each other as

  17

well as higher levels of retention. Particularly among players who are about to quit the game because they’re always losing in their attacks or defenses. But would stick around if an experienced clan member could coach them.

Feature Recommendations (Other) This now brings me to my two final priorities on the game’s product roadmap. Yes, I still believe that improving upon the game’s social experience through the three features proposed above is “mission critical.” But I don’t want to leave out two other priority features that I believe can also help to take the game’s retention and monetization success to the next level. Boost The first feature would be the Boost feature that I proposed in Appointment Mechanic (The Bad) in What Doesn’t Work. Where the player can cut the amount of time in half that it takes to upgrade buildings and armies when the upgrade times exceed two days. I believe that this feature is worth building and testing. As a solution for helping mid to later-stage players remove the upgrade friction that makes it tough for them to play the game on a daily basis. Like they once did and want to get back into the habit of doing. The results of this initiative can then be used to gauge the viability of iterating to the Pay As You Go function that I also proposed in the same section as Boost.

Wall Mover Finally, one of the best fun pains for a player in the game to engage in beyond the core gameplay is re-arranging his village. The fun lies in the promise of winning more trophies through more successful defenses, while the pain lies in having to move the walls of his village. Having to move the walls of his village one stone at a time becomes excruciatingly painful to do for mid to later stage players with anywhere from 100 to 200 stones to move. Made even more difficult by having to move the stones on a tiny mobile device screen. Eventually leading some players to say, “why bother to re-arrange my village anymore, the hassle of moving the walls is too much.”

  18

From a behavior design perspective, this is well known signal of a user that is likely to engage less and less with the product going forward. Not good for driving retention of a player who is just now at the point of the game where he’s invested enough of his time that he will start to spend money in order to preserve the investment of his time. So why not hit two birds with one stone. I propose developing a Wall Mover feature that not only makes it super easy for a player to move his walls, but also in making his first purchase in the store a no brainer. By selling the player this Wall Mover for $0.99. The objective of this no brainer first dollar is to get the player to spend for the first time, not to maximize revenue. Making it incredibly easy for the player to spend the $0.99 that he’ll perceive as an excellent value investment for preserving his time invested in the game to date. Now let’s talk about which feature to implement first. Please turn the page.

  19

NEW FEATURE EVALUATION

Choosing Which Feature To Implement First Taking a Goal Oriented Approach Say that we observe a declining trend in the game’s ARPPU over the last three months. In peeling back the layers of this onion, we uncover an actionable insight. The game is drawing fewer players along its revenue curve (see below) from Point A to Point B because of an increasing rate of churn among the minnow spenders in the game. Creating a trickle down effect that adversely impacts the conversion rates of minnows into dolphins and dolphins into whales (Point C). That is to say, the continued ability of the business to profitably scale CoC net of marketing and development costs. Drilling Into the Problem

Revenue Curve

Product wise, we might then determine that this rate of higher churn is attributable to a decrease in the “win/loss” ratio among level 25 to level 30 minnows that spent $4.99 to buy 500 gems, and used the gems to accelerate the training of troops for raids or upgrades of buildings for defenses. Only to lose more often than previous cohorts did as measured by their aggregate “win/loss” ratios. Leading us to next perform another statistical analysis. That reveals the decline in this “win/loss” ratio across the last 12 monthly cohorts of level 25 minnows is positively correlated to the challenge that these minnows face in finding a combination of troops or buildings that will generate wins in their first five raids and defenses after they gain level 25 status. What this means is that from both a business and product perspective, our goal is to help more level 25 minnow spenders in the game get over this first major

  20

progression hump. In order to stem the decline in ARPPU and preferably without tinkering with the balance of this well balanced game. Design Solution (Dueling Members) One solution for drawing more of these minnows along the revenue curve is the Dueling Members feature (see Product Priorities). In that this feature builds on a behavior that these minnows already engage in today. Only that Dueling Members makes it far easier to solicit and receive high quality advice from fellow clan members than the two clan-based features that players can use today to do so. Thereby providing the minnows, and heck their fellow clan members too, with a tool that they would truly value. For honing their abilities to wage effective raids and laydown better defenses, and more importantly, removing the frustration that leads these minnows to stop playing. Placing them on a path to become long term, high-spending players, without charging them for the feature.

Evaluating Impact Of Selected Feature What I mean is that CoC employs a progression based monetization strategy as discussed in What Works in the game. Meaning prices for items and upgrades only increase as a player progresses in the game. Creating a situation in which ARPPU goes hand in hand with retention – more progress the player makes, the higher his average purchase will be (for the same items and upgrades he acquired before). Meaning retention rates are truly the bellwethers of success in this game. Retention So let’s first evaluate Dueling Members’ impact on retention – by architecting an analytics system to capture and monitor the effect of this new feature on the retention rate of our target minnows. How we would build this system is by collecting the telemetry data in the game and translating the data into metrics for understanding Dueling Members’ impact on the retention profile of our target minnows. First mapping the retention profile of level 25 minnows over the last 12 monthly cohorts by retrospectively indicating engagement and retention metrics. Before next evaluating how three features in the game today impact this retention profile and its correlation with the “win/loss” ratio. Thereby providing us with a proxy against which to evaluate

  21

the impact of Dueling Members in moving retention. Given that Fight Replay, Clan Chat, and Village Visit target the same important yet unsatisfied job that Dueling Members is targeting and can do a better job of getting done. Retention and Engagement Metrics to Track Primarily measured by DAU and its correlation with the “win/loss” ratio. I believe that DAU retention by days is the king of all general metrics in the game and the “win/loss” ratio is the most important in game activity metric to track for moving the needle forward on the retention of our target minnows. Before stating all metrics that we should track, let’s first contextualize their importance through the diagram on the next page. Retention and Engagement Flows

Primary Metrics to Track (For Answering the Questions that Follow)

• DAU Retention by Days o What is the current retention profile of our DAU minnows by

calendar days o By how much does the new feature improve this current retention

profile

  22

o By how much does the new feature improve the rate of reactivated users

o How do these improvements compare to those brought about by existing features

• DAU/WAU

o What is the current engagement profile of minnows o By how much does the new feature improve this current

engagement profile o How does this improvement compare to that brought about by

existing features

• “Win/Loss” Ratio (In Game Activity Metric) o How well does this ratio predict churn among our target minnows

More Granular Metrics to Track More granular metrics that we would track, analyze, and optimize would be determined by running A/B tests. Analyzing the conversion of minnows that use Dueling Members versus any combination of the three aforementioned features in the game today, or choose not to use either feature set at all. In turn allowing us to segment our target minnows by those that are most likely to convert to using Dueling Members based upon attributes like level of play, hour of day, etc. All of which we need to optimize the feature to further drive retention via the feature’s in-game virality mechanic. With this said, our A/B tests are most likely to reveal the need to track these more granular metrics:

• DAU/WAU by Feature o Existing Features (Fight Replay, Clan Chat, and Village Visit)

§ How frequently do our minnows interact with these features (all combos)

§ How do these interactions change as minnows make progress in the game

§ How do these interactions affect retention by days, sessions, session times

o New Feature § How frequently do our minnows interact with Dueling

Members

  23

§ How do these interactions change as minnows make progress in the game

§ How do these interactions affect retention by days, sessions, session times

§ How do these effects compare to those of the existing features

o Neither Feature Set § What percentage of our minnows use neither set of features

• Playtime

o How many more minutes of play does Dueling Members generate § By level, day, and time

o How does this effect compare to that brought about by the three existing features

• Drop-off Rates

o By what level have 80% of our target minnows stopped playing the game

o What effect does either feature set have on reducing drop-off rates by level

o How does this effect compare to that of the three existing features

• Reactivation Rates o What effect does either feature set have on improving rates of

reactivation o How does this effect compare to that of the three existing features

Now let’s assume that the Dueling Members feature is moving retention of our target minnows in the right direction. That is to say, the feature is helping them to win more raids and more defenses. To the point where their frustration no longer leads our target minnows to stop playing the game. This does not necessarily mean that the feature is drawing our target minnows to Point B on the revenue curve. The feature may be doing so, or may not be doing so. We won’t know until we take a look at a few key metrics for evaluating the impact of the feature on the game’s revenue. Revenue

  24

Let’s begin by benchmarking the absolute revenue generated on a daily basis at both Points A and B today. Measured as a function of: (DAU) * (Buyers/Players) * (Revenue/Buyers). Where we insert the number of minnows or dolphins where “Buyers” is indicated. Why also benchmark absolute revenue at Point A is to see whether Dueling Members has a “middling” effect. That is to say, the feature generates more minnow purchases at Point A versus more dolphin purchases at Point B. It’s useful to know that the typical minnow and dolphin purchases at Points A and B are $4.99 for 500 gems and $9.99 for 1200 gems, respectively. Now let’s move on to diagramming the flows that affect absolute revenue at both points A and B. Using metrics that will allow us to measure how much Dueling Members contributes to absolute revenue at each point and are spelled out in the following diagram: Revenue Flows

Next up then is benchmarking the impact of two in-game activities on absolute revenue: upgrading defenses and accelerating upgrades. Measured by our target minnows’ use of Village Visit, Fight Replay, or Clan Chat, or all three, right before these players spend gems purchased at points A and B, and assuming that these two in-game activities account for the majority of gem spends. Of course, we’re also assuming that statistical analysis validates our hypothesis that changes in the usage of these three features (as well as Dueling Members when implemented) are a strong predictor of changes in the frequency of our

  25

target minnow’s defense upgrades and accelerations, which again directly impacts our minnows’ “win/loss” ratio that keeps them playing the game. Let the fun begin now. We would run A/B tests. Analyzing the conversion rate of our target minnows into dolphin spenders right after using Dueling Members. Comparing that conversion rate to those conversion rates attributable to any combination of use of Village Visit, Fight Replay, or Clan Chat, as well as to minnows that convert into dolphins without using any of the features. We would also run tests to determine the impact on revenue by minnows that use all four features. Much like we would do in evaluating the impact of Dueling Members on the game’s reach. Reach Let’s think of Reach as the Princess in our royal family of game metrics, without which, Retention as our King and Revenue as our Queen wouldn’t have loyal and profitable subjects to preside over. What makes for an attractive Princess of Reach is strong in game virality and game viralization metrics. Both of which Dueling Members can drive nicely. Here’s why. When our target minnows duel with fellow clan members, that’s definitely something worth sharing, win or lose. And we’re not talking about posting “brags” on the minnow’s Facebook timeline. We’re talking about posting the videos of the duels themselves. So friends who also play the game can see how to lay down better raids and defenses, while non-playing friends can pique their interest in giving the game a try. Oh, let’s not overlook those friends who stopped playing the game altogether. They may reactivate themselves as players after seeing how to improve their “win/loss” ratio through the videos. No matter the virality scenario. How we would measure the impact of Dueling Members on the game’s Reach is by tracking the metrics in the following function: [Posts * (Receivers/Post)] * (Click Through Rate). In order to answer the question of how many DAUs do Dueling Members videos generate per post. We can then optimize for the number of DAUs generated per post of a Dueling Members video by targeting the dialogue and clickers in the posts. Meaning that we should also track the breakdown of the click through rate on an “installs” versus “reactivations” basis.

  26

Now that we have a way for evaluating the impact of our new feature priority on the game, we need to evaluate the cost of building the feature and getting it into the game. So let’s wrap up our discussion by evaluating the impact of Dueling Members on the game’s engineering. Engineering Before evaluating the impact of integrating Dueling Members into the game, please know that I’m not an engineer by training, only a product manager with some programming skills. So take what I say with a huge grain of salt, and correct me where my thinking goes askew. It’s how I learn best. On the surface, Dueling Members leverages engineered assets that are already in the game today. Namely, the game’s ability to record and replay fights. What this means is that we may be able to use much of the code base for the Battle Log and Fight Replay features to record and reply duels. In terms of new engineering, we will need to replace the Revenge button for each battle logged and recorded with a Share button in order to allow both playing and non-playing friends to watch these recorded duels on Facebook and within their clan areas of the game. Swapping out the button from a frontend and backend integration perspective is not trivial. Mostly because of the need to post the videos in the clan chat area. Even though posting videos to a player’s Facebook timeline should be more straightforward upon reviewing the literature for Facebook’s API for posting video content. How we would ascertain the scope and scale of these integration requirements as well as identify other development costs is by asking the development team for a cost estimate. In the form of five bullet points that we would ask the team to spend no more than an hour or two to complete, and exclude technology debt and maintenance, which we’ll assume accounts for 10%-15% of cost We would also ask QA for a cost estimate, which will take time to complete. Given that our target minnows are highly unlikely to be the only players interacting with this feature. The majority of the player population is sure to interact with the feature. So QA will have its hand fulls.

  27

End of Document