Defender’s Quest: By the Numbers
by Lars Doucet [Business, Social/Online]
Today I’d like to talk a little bit about the marketing strategy and sales results for Level Up Labs’ latest game, Defender’s Quest: Valley of the Forgotten. Though our game was only released this January, we have already received significant critical acclaim and brisk sales beyond our wildest expectations, and wish to offer any information from our experience which might help other independent game developers succeed.
今天我要說明有關於我們團隊Level Up Labs最新作品Defender’s Quest: Valley of the Forgotten的市場策略與銷售結果。雖然我們的遊戲今年一月才剛釋出，我們就已受到顯著的喝采，持續成長的銷售也超過我們最好的預期。因此我希望能提供我們的經驗給其他的獨立製作者參考。
But before we get started, let me introduce myself — my name’s Lars Doucet, and I’ve spent the last six years as a freelance game developer and writer. I’ve worked on serious games for the MacArthur Foundation, Texas A&M, Rice, and Wake Forest Universities, as well as a (so-far unreleased) Facebook game for a major publisher.
在開始之前，讓我先自我介紹我叫做Lars Doucet，最近六年自詡一位獨立遊戲開發者與作家的身份。我曾作過MacArthur Foundation，Texas A&M，Rice，and Wake Forest Universities的教育遊戲，以及一個才剛釋出的Facebook遊戲。
My game design articles also appear frequently among Gamasutra’s expert blogs. Our current team at Level Up Labs includes Anthony Pecorella, our lead designer and an employee of Kongregate, James Cavin, our writer and character artist, and Kevin Penkin, our musician.
我的遊戲設計文章也發表在Gamasutra上。目前我在Level Up Labs的團隊中還包含主企劃Anthony Pecorella，Kongregate的一個成員James Cavin，他擔任劇本與角色設定，以及我們的音樂Kevin Penkin。
Defender’s Quest is a departure from my previous work making educational games like Super Energy Apocalypse and CellCraft. For this game, our team has done its best to combine the storytelling, melodrama, party-building, and metagame elements of a tactical JRPG with a tower defense-style battle system.
Defender’s Quest是從我之前一個教育遊戲像Super Energy Apocalypse與CellCraft脫離出來的。這個遊戲整合很多敘事手法，音樂劇，團隊建立，以及一個日式RPG的塔防戰鬥系統的一些元素。
However, instead of ubiquitous “towers," players summon to battle individual characters from their party, each with their own personalized names, stats, equipment, and skill trees. Like a traditional RPG, characters level up between battles and gain strength. The player can also specialize each party member’s skills to fill different strategic niches.
We imagined a short, three to six month project, but after a few months, it became clear our ambitions for the project would take a lot longer. This was a big problem, due to the economy behind Flash games. Most revenue from Flash games comes from a primary sponsorship deal, where a sponsor (usually a games portal) pays to have their logo and a link to their site in all versions of the game, stripping all other outgoing links, in order to capture all the game’s traffic and attention for the sponsor’s benefit. Essentially, a sponsorship is an offer to buy your game’s audience.
As the game spreads across the internet, it drives traffic to the sponsor — so the better the game, the higher the traffic, and the more the sponsor will pay. No matter how good the game, however, every sponsor’s payment has an upper limit. Practically, this means the longer you work, the less you earn per hour spent. Extra cash can be earned from ads, contests, secondary sponsorships, and microtransactions, but even with all these extras, the pay for this labor intensive work usually amounts to peanuts.
It’s worth noting that microtransactions work best in multiplayer games, or Facebook games, but according to Anthony’s experience at Kongregate, they generally do poorly on single-player games on Flash portals.
Andy Moore’s famous SteamBirds: By the Numbers article and others like it reveal that top-quality games can expect a sponsorship of around $25,000, with a theoretical (albeit virtually unattainable) maximum somewhere around $50,000 if the game is literally the best Flash game ever, AND you’re willing to accept draconian terms. And even if a team could expect these amounts, they would still be forced to cut corners to keep development time as short as possible.
Andy Moore有名的文章SteamBirds: By the Numbers說明了高品質的遊戲可以得到一個25000美金的贊助(折合台幣74萬)，若是名列前矛的遊戲可以再提升到50000美金(折合台幣148萬)。即便是這樣的收入，團隊都必須縮衣節食，想辦法減少開發時間。
This was our original plan — create a high-quality game worthy of a top sponsorship in a short enough amount of time to make the effort worth it. But after sinking countless hours into the project, our only shot at this strategy was to slash features. Even then, we’d still have to be the best Flash game ever to nab a large enough sponsorship to provide a decent hourly wage for all four team members.
Instead of hoping to make minimum wage for our efforts, Anthony and I doubled down on Defender’s Quest. We decided to upgrade from a free Flash game to a full commercial release. This was risky — commercial games are obviously judged by a higher standard than free games. Furthermore, Flash has the stigma of an “amateur" platform. Among the larger independent game development scene, there is also the sentiment that success is difficult or impossible without the aid of gatekeepers like Steam or the Humble Indie Bundle.
After much toil and sweat, we released the game on January 19th, 2012. This was a full 20 months after we started, more than three times longer than our original worst-case estimate. We spent the extra time polishing the game’s story and battle mechanics, features particularly scrutinized by RPG fans.
We’ll go into greater detail about our marketing strategy later in the article, but the basic approach was to sell the game on http://www.defendersquest.com, post a free demo, and then upload the demo to all the free Flash game portals, driving traffic back to our site.
The initial numbers have rolled in, and now I’m here to report on the results.
Let’s start with sales numbers. We went on sale January 19th. As of this writing, we’ve been on sale for about two and a half months. All dollar values are in USD.
毛利 70716 13846
Defendersquest.com 52900 9973
FastSpring 抽走8% -4232
Kongregate.com 17816 3873
*This does NOT represent a profit of 46.7K. Altogether, we put in several thousand hours of labor that still needs compensating.
We sold the game on http://www.defendersquest.com by partnering with FastSpring, a popular online payment provider, which takes 8 percent of our gross revenues. At first, everyone who played the demo on any Flash portal was directed to our website to buy the full game.
On February 7th, we started letting players on the portal Kongregate buy the full game with “kreds," the site’s virtual currency. Even though Kongregate takes 30 percent of kred revenue, we decided to try it because the system was easier for players, which would hopefully lead to more overall sales.
We started with the special introductory price of $6.99 on our site, and 50 “kreds" (about $5) on Kongregate.
Why two different prices?
1. The Kongregate version is online-only, and has compressed graphics/audio. The download version has a few extra features, can be played anywhere, and has higher-quality assets.
2. 50 kreds is a better impulse purchase price than 70, since the smallest amount of kreds players can buy is 50.
Though the game is a complete, finished experience, we plan to add more content later in an upcoming Gold edition release. With the bonus material included, the final price for Gold edition will be $9.99. Everyone who has already bought the game will receive free upgrades for the additional content.
Next, let’s look at a chart of our sales over time:
Blue represents gross revenue from our own site, and red is gross revenue from kreds purchases. I’ve also marked various points on the graph with the following press and other events:
Source Event or Article Date
新聞 Destructoid Tower Defense meets RPG in ambitious Defender’s Quest 1/20
新聞 Joystiq The Joystiq Indie Pitch: Defender’s Quest 1/21
新聞 Jay Is Games Defender’s Quest 1/22
新聞 Rock, Paper, Shotgun Hours Of Towers: Defender’s Quest 1/23
發售 Event Launches on Kongregate 1/27
新聞 Destructoid Review: Defender’s Quest: Valley of the Forgotten 1/28
發售 Event Launches on Newgrounds 1/30
進入首頁 Event Front page of Kongregate and Newgrounds 1/31
下架修改 Event Newgrounds goes down for redesign 2/6
開始收Kreds幣 Event Enabled kreds on Kongregate 2/7
進入排行榜 Event Featured Game on Kongregate 2/8
新聞 Rock, Paper, Shotgun Second Wave: Defender’s Quest Free Gold Update 3/22
Mentions in the gaming press (Rock, Paper, Shotgun, in particular) correlate to single-day sales spikes. The largest spikes, however, as well as sustained sales over time, correlate with launching on the major Flash game portals (particularly Kongregate and Newgrounds). So the question is, which made a bigger difference: Flash portals or good press? For further analysis, we turned to our coupon data.
We made several different coupon codes available, as one of our many metrics for tracking sales and virality. We gave people who had signed up for our newsletter prior to release a special coupon for $2 off. We gave reviewers coupon codes named after their sites, which were good for $1 off. Our first batch of coupons was good through the end of January, but we later added some that were good through February 14th, specifically the Kongregate and Newgrounds coupons.
Here’s a chart of sales by coupon:
You’ll also see some other coupons on the chart. I gave the coupon code “ZEBOYD" to fellow indie RPG developer Robert Boyd, who sent it out to his followers over Twitter and posted it on various forums around the internet. The “MONA" code was posted by our lawyer, Mona Ibrahim, and I posted another in my previous Gamasutra blog post about our launch.) The “SPECIAL" code was the $2 coupon from our newsletter.
When we released on Newgrounds and Kongregate, we were not sure if a demo with an up-sell would anger players, since those sites are free-to-play portals. Anticipating the worst reaction, we gave players on each site a $2 off coupon towards the game. After about a week on Kongregate, we replaced the link to our store with a link to buy the game for 50 kreds.
Coupon codes correlate fairly strongly to sales sources, but they’re not exact. For instance, many people from a referring site will not use the provided coupon. Coupons have a tendency to leak throughout the internet, so obviously people outside the original target group end up using them. Also, if the coupons differ in value, word eventually gets out and people start using the higher-value coupon more. In the above example, the newsletter coupon (SPECIAL), was worth $2 off and only advertised in our newsletter. Many newsletter subscribers bought the game at full price anyway, and a lot of people who used this coupon were not on the newsletter.
Finally, all the coupons eventually expired, so sales from these sources after the expiration dates aren’t reflected in this chart.
But even with these caveats, I think it is fair to say that these sites were among our highest sources of revenue so far:
Based on coupon redemptions alone, the two flash portals, Kongregate and Newgrounds, represent over 75 percent of sales.
Given that this chart only represents sales made with a coupon from our FastSpring store, and leaves out kreds sales altogether, the real number for Kongregate is surely signficantly higher.
For now, we are only selling on our site with FastSpring and on Kongregate through kreds. We plan to expand onto other portals soon (Desura/Impulse/etc.), and of course, if Steam or the Humble Indie Bundle come calling, we’ll topple off our seats.
But here are some things we did to increase sales.
I don’t know that much about web development, so I whipped up something very simple from an old css template. It’s just static HTML with flat colors, images, text, links, and embedded videos.
The website was designed to achieve two things as quickly as possible:
1. Give the visitor everything they need to make a decision
2. Make it very easy to buy the game
The first thing you see when you go to http://www.defendersquest.com is this:
We do have a really nice quote from Rock, Paper, Shotgun, with a link to the review they did for us, but we didn’t launch with it — Adam Smith didn’t get his review copy until after launch, so his article came out a week later. Prior to this, we used a positive quote from Robert Boyd of ZeBoyd games.
在Rock, Paper, Shotgun我們確實有一些好的評論文章，但是事實上這篇評論並不在我們銷售計畫中，甚至比我們發布遊戲還晚了一週才出現。在此之前我們引述了Robert Boyd在網站ZeBoyd games的評論文章。
Right below the heading is our launch trailer. We put this together using free screen-recording tools and iMovie. Captioning the video is our elevator pitch, and our technical selling points — “cross platform and DRM Free." Finally, we have our BUY/TRY buttons, which are on every page of the website.
We have the usual information pages: about, support, faq, etc. But the two most important links are FAQ and demo, so I made those stand out in the navigation bar.
I’ve already written about our free demo strategy, but I’ll summarize here. We have a free, lengthy, web-based demo that visitors can play without downloading or installing anything. We also have downloadable demos for each platform, through direct link and BitTorrent.
When the player finishes or exits, the demo directs them to an up-sell page that links to the store and tells them how to export their save file to the full version. We hoped the demo would answer most of the player’s questions. Those wanting a richer experience with uncompressed assets and fullscreen support can install the downloadable demos, which also demonstrate whether the full game will run on their system. If the demo works, the full version should work too.
We also considered the fact that some visitors might be parents, so we included a page for them. Standard ESRB ratings aren’t very helpful, and as low-budget independent developers, they weren’t really an option for us. So to address the ratings issue, this button accompanies every BUY/TRY pairing:
I have observed several e-mails and forum postings from parents specifically mentioning this page as a reason for buying the game.
Next, I’ll talk about our conventional marketing strategy.
First, we created a press release. You can find the full text on our media page as part of our press kit, but I’ll share a snippet here:
Level Up Labs is proud to announce their new Tower Defense / RPG hybrid Defender’s Quest for Mac, Windows, and Linux!
The kit then goes on to give a brief game description, list of features, quotes from reviewers, links to reviews, along with additional pertinent information. However I felt the most important thing was the first five lines. I know reviewers are busy and appreciate us getting right to the point. In the first line, we give our company name and elevator pitch, and mention that the game is cross-platform.
This is followed by our “three magic links." These links contain absolutely everything a reviewer needs, even if that was the whole press release. The first is a video, as it’s easy to follow that link, sit back, and watch. If that catches their interest, they’ll want to know more about the game and perhaps give it a spin. The playable demo should give them hands-on experience instantly while the media press kit gives them all the material they need for an article. Below is all we included in our press kit:
The fact sheet contains the game’s basic info — price, technical settings, features, etc, and is included in three different formats to make it easy for the reviewer. “Box_art" contains, well, box art — which makes for pretty pictures to accompany an article.
“Icons_and_thumbnails" includes pictures of the game in all shapes, sizes, and image formats — these are mostly for Flash portals that want to host the demo, but can also be used for a reviewer, for example, to accompany a short news post link where space is limited. “Press_release" and “screenshots" are pretty self-explanatory.
We created a dedicated media page that has all this information, as well as links to all the press we’ve received. There’s a link to the press kit in big, obvious letters, and anchor links to the rest of the page’s content to minimize scrolling.
We’ve been updating both the press kit and this page after launch to add new quotes and reviews as they come in. We also tried a few things to announce the game prior to launch, which I’ll touch briefly on now.
Many developers use Games Press, a clearinghouse for video game press releases, to contact the press. We first used it to announce our public demo a few months prior to launch. Andrew Smee of Rock, Paper, Shotgun wrote a lovely preview of our game, and told me his editor probably discovered our game on Games Press. Using Games Press again after launch had less success — it seems to have resulted in many mostly blank, auto-generated profiles for the game on sites like GameFAQs and Metacritic, but no actual reviews.
很多開發者使用Games Press，一個影片發布的網站來與媒體溝通。一開始我們透過它來發布展示版本。Rock, Paper, Shotgun網站的Andrew Smee寫了我們遊戲的評論。他說他們編輯就是從這網站知道我們遊戲的。但是發布後在透過此網站就沒有什麼進展。有些自動產生的連結，但是沒有真正的評論文章。
Prior to release, we also sent emails to Joystiq and Jay Is Games. Joystiq accepted our query and ran an article just after the game’s launch, and Jay Is Games did the same, both correlating to sales spikes (see chart, above),
發布前，我們寄送了電子郵件給Joystiq與Jay Is Games。如前面折扣券的地方所說，他們都幫我們寫了評論。
I must confess a total lack of experience with social networking. Here’s a few things we tried.
We put up Facebook and Google+ pages for our game which link to our site, and we put “like", “+1″ and “Reddit" buttons on our front and demo pages. The most attention we seem to have gotten is just from my friends, family, and followers by directly talking about the game and posting links to reviews on my own personal Facebook, Google+, and Twitter accounts.
我們網站也建了對應的like +1 reddit按鈕。大部分都是我們親朋好友來幫助我們透過Facebook， Google+，Twitter分享出去。
I’ve had the most success with Twitter, perhaps because it’s so simple. You tweet stuff and if people care, they follow you, etc. That’s pretty much it. It’s also a great way to connect with people that might otherwise be difficult to meet.
We got a lot of attention from Reddit threads (particularly ones sharing our coupon codes). I posted a $1-off REDDIT coupon when I saw the activity, which generated a decent chunk of sales (see chart, above). I also started an “Ask Me Anything" thread, which generated a lot of interest.
Surprisingly, most of the attention our game has gotten through social networking has come from people sharing an blog I wrote, “Piracy and the Four Currencies." As soon as I saw this, I added prominent links to Defender’s Quest in the article.
Free-to-Play Flash Portals
Front page of Kongregate! W00t!
Although most of our players never notice, Defender’s Quest is a Flash game. We compiled it with Adobe AIR, which let us employ some tricks to get the standard “desktop quality" game features, like fullscreen resolution switching, to work. We also used Flixel, which gave us great performance (for a Flash game) by circumventing Flash’s normal, unoptimized, display logic. Using Flash made it easy to develop a browser demo, which we could distribute through flash portals.
As I mentioned in my previous article, rather than paying to host the demo on our own servers, we uploaded it to Mochi Media, turned off advertisements, and embedded it on our site. Then, we enabled distribution, which means that Mochi would send the demo to lots of flash portals across the internet.
After seeing a truly astonishing number of visitors to our site from http://www.minijuegos.com of all places, we decided to go after the big boys — Newgrounds, Kongregate, and Armor Games. We launched on Newgrounds and Kongregate, but Armor Games is invitation-only and we weren’t able to reach an agreement with them.
當看到從www.minijuegos.com連過來數目驚人的瀏覽者，我們決定繼續發布在Newgrounds，Kongregate與Armor Games。不過由於Armor Games的遊戲是由該網站所主動邀請，所以我們一直未能通過該網站的審核。
Our plan was to create a special version of the browser demo for each major portal, implement each one’s achievement APIs, and create a unique demo landing page for each offering a $2-off coupon. Since I had gotten complaints that our store’s coupon field was hard to find, I made the “buy" button on each landing page pre-apply the coupon to the order. These coupons were worth more than our usual $1 off, but because we were afraid of “demo backlash" we figured the magic $5 impulse-buy price couldn’t hurt.
We started with Kongregate and immediately had our worst fears confirmed. The game’s page was quickly flooded with flames and one star reviews. Our score was around 3.4 and dropping steadily, which means our work would soon be consigned to video game oblivion. Many players complained they had been tricked, and that demos were unwelcome on free game sites.
We were also accused of being “greedy," “evil," and — my favorite — “illegal." Against my better judgment, I logged into my personal account and argued back, but my comments were swiftly down-rated by angry players, and my grumpy words (mercifully) disappeared in minutes.
Anthony acted much more rationally than I did. He jumped onto our developer account and instantly started damage control. He changed the name of the entry from “Defender’s Quest Prelude" to “Defender’s Quest (lengthy!) DEMO". He updated the description fields to say “THIS IS A DEMO!" right at the top, rather than tucked away after the game’s feature list.
If people aren’t expecting this, it makes them very angry!
He started responded courteously to each negative comment and within a day our review score had started turning around. Fans started buying the game, and a lot of them went to bat for us in the comments. After a few days we cracked the four star barrier, and were featured on the front page.
We’re still getting the occasional one star hate comments (who doesn’t?) lambasting us for being a demo, but the score climbed steadily. All of the metrics suggest Kongregate is overwhelmingly responsible for the majority of our total sales.
After things settled down, we got many comments like, “Glad you changed the description. I don’t mind demos, I just wanna know what I’m getting up front." as well as, “I never buy based on a demo, but I bought this." So although our demo will never get a top rating, changing the context to manage expectations seems to have helped the perception.
A vocal minority will likely continue to down-rate us, but that comes with the territory.
We later launched on Newgrounds, using the experience from Kong to guide us. We made it absolutely clear in the title and description that the game was a demo, and included links to the RPS and Destructoid reviews of the game. Newgrounds received a similar $2-off coupon, just like Kongregate.
I also created some medals (achievements) for the game. Kongregate badges are all hand-crafted by Kong employees (mostly Greg McClanahan), but on Newgrounds you can create them yourself and submit them to Newgrounds chief Tom Fulp for approval. I created some for the usual stuff like beating levels and gaining party members, as well as a set of easy ones just for visiting the various option menus.
My hope was that players would spend the first five minutes of the game exploring the options menus to rack up some quick medals and discover all the options and customization in the game. Hopefully players likely to complain about “too much reading" would quickly see they can turn tutorials and dialogue off, and those demanding “more hotkeys" would discover that not only does the game have hotkeys, but they are all customizable. The entire game can be played with nothing but the keyboard should they so desire.
Newgrounds’ feedback system is less noisy than Kong’s, and oriented towards reviews rather than comments, so the reviews are dominated by players who have bought the full game. Furthermore, Kongregate’s system only allows a user to vote once, but Newgrounds lets users vote every day, giving higher voting power to established users.
The upshot is that on Kong, the score depends more on the first impressions of a large audience, trends downwards over time, and is more vulnerable to one star hate-reviews. On Newgrounds a devoted fanbase of site regulars can gradually raise the score and easily offset drive-by down-ratings. Shortly after launching, our score on both sites was around 4.11. It’s plateaued on Kong, but it’s up to an unbelievable 4.40 on Newgrounds!
Shortly after I submitted my medals to Tom Fulp for approval, he responded to my message with great news. Not only had he approved the medals, he’d also placed the game on the front page! Unfortunately, Newgrounds went down for their long-delayed redesign on February 6, and when it came back up, we were just below the threshold for front page status.
Even so, coupons suggest we’ve made a lot of sales from Newgrounds — even more than from Rock, Paper, Shotgun, making them our second-largest source of sales. Together, Newgrounds and Kongregate probably represent about 80 to 90 percent of our total revenue. This is a completely unexpected outcome; we thought a demo up-sell model was a long shot for Flash portals, and anticipated only modest sales.
折扣券顯示我們有很多銷售數字來自Newgrounds，比Rock, Paper, Shotgun還多，名列銷售來源第二名。Newgrounds與Kongregate可能包辦我們8成到9成的獲益，這是未預期的結果，我們本以為這方法在Flash入口不會管用，而且只會貢獻小部份銷售，
People playing games on free Flash portals don’t like spending money, and they also don’t like going to another website to do so. Given we were asking them to do both, our success took me by surprise. We estimate that our demo-to-sales conversion rate is somewhere between 1 to 5 percent. Moving to kreds on Kongregate seems to have helped — on-site currencies let players buy the game without going to another website or pulling out their credit cards.
Since we knew that many smaller flash portals “steal" game files from Kongregate and Newgrounds, we wanted the Kongregate version to work even on other sites. This is how Flash games spread virally. To be clear: we want portals to do this so that our demo gets more plays, and hopefully, more sales. The Kongregate version had special code that detected when it was running on a site other than kongregate.com, which would make it revert to the normal format, directing players to visit our website to buy the game.
So, that’s everything we’ve done, and our sales numbers so far. I’m not prepared to draw broad conclusions as to what worked and what didn’t in terms of driving sales. It’s still a bit too early in the process for that. But I will say that I think there are some pretty clear spikes correlated with reviews and portal promotions. Our steady daily sales seem like they’re coming from Kongregate, Newgrounds, and small flash portals.
Our last and most important strategy has been to focus on customer service. Anthony, James, and I all try to keep the backlog from building up. We do our best to stay abreast of glitches and compatibility problems, and to surmount the limitations of our less-than-perfect storefront. So far customer feedback has been positive.
We hope to move on to other stores like Desura and Impulse, etc, soon, but first we want to iron out the major bugs, solve compatibility glitches, and finish the bonus version so we can put our best foot forward on those platforms.
I hope all this information helps someone.