Indie Game App Store Gap

The advent of mobile platforms has been an amazing boom to independent developers. The app economy has been a virtual gold rush, in all the best, and worst ways. Since the advent of Steam, iOS, Android, Humble Bundle etc., each with their own developer ecosystems, there has been an explosion of channels for indie game developers. I’ve dabbled in some of these since their inception, but in the last year I’ve become more involved through helping a friend, a gifted designer, with their efforts to survive, and thrive as an independent game developer.

I’ve learned a few things helping him, and watching others. The sporadic, and growing habit among indies to publish their exploits has provided an interesting text-based, episodic form of reality series on life as a modern solo-to-small company game developer. Like stories of other pioneers, and explorers, many don’t end well. These post-mortem lessons are sometimes hard to read, let alone hard to take.

A big problem isn’t that they don’t always end well. The problems, as I’ve encountered it, both helping others, and working toward a few releases of my own, is that for all the activity going on in independent games, let alone media, and independent software development, there is a lack of actionable insight, and a dearth of deconstructed models and systems for the reader to learn from. Instead, we’re left to cobble together bits and pieces of information, hack through false axioms, and superlatives, and deal with amazing declaratives that are the false-idols of advice.

A recent article on Gamasutra offered the analysis of the top three reasons games failed: Quality, Positioning, Visibility.

Let’s be clear, there are many games that aren’t fun, and many pieces of software that have defects they should not. Lots of games that lack visibility, but if you’re at a point where advice like Quality, Positioning, and Visibility are meaningful, you’re in a heap of trouble. And if you’re dispensing such advice, you’re not helping anyone who is otherwise capable, but is thirsty for something actionable.

With all this in mind, and a lot learned, it seemed worthwhile to try and present something I wish was more representative of what game developers should be sharing with each other.

Models Needed

For me, one of the biggest issues going has been to try and understand things beyond the disparate parts. To actually construct models, and written descriptions of what constitutes useful strategies for indie game development. By trying to stitch together various ideas into something more whole, in written form, it forces truth above fiction. It requires digging in, because at the end of the day you need enough parts, enough truths, enough evidence, and insights, to be able to cobble together something complete, and decently detailed.

Set The Game Aside for Now

Recently my friend launched a game on the iOS App Store. As I said, he’s gifted, and so it’s doing well. By all accounts it’s a decent hit, and the hand-crafted nature of his efforts means it’s profitable. So a successful game launched it is, but it’s not without faults and interesting lessons, and deep within the fault lines of an otherwise successful game I think lies some important discussion for independent game developers operating in today’s app store dominated world.

For my friend I work on three core tasks:

1. I play early versions of games he is developing. I offer my own thoughts both as a designer/producer, but also someone trying to focus one-or-two steps ahead in terms of how to sell, and earn produce revenue.

2. I participate in launch, providing help with organization of PR, and other tasks around visibility, messaging, and merchandising. I suggest ideas, keep a punch list, and help on the ones that seem doable, and fit his vision for how he wants to sell his game. This also includes pitching in on possible biz dev ideas when they come across his email.

3. I analyze resulting data, and look at data and evidence from other developers (game and otherwise) that offer insight as to what to do next immediately, and for future games.

All of these efforts, and his efforts to build the game, require constant questioning of the value of effort, time, and out-of-pocket costs. Small developers have unusually high opportunity costs that can trap them. They can only do so much due to time, attention to more important things, and pure capacity. At the end of the day, getting the game done, or updated, always wins out, which lowers energy and time spent on other tasks. Such limited capacity means every decision is seen as a risk more than an requirement; evaluating opportunities to understand what they represent can be a task left unattended.

One task I’ve been working on is trying to build a model around the opportunity for paid media in promoting his latest game. This analysis also serves as a test-case for eventually promoting all his work, past and future. There are two reasons I’ve been pursuing this:

1. Making paid media a net-positive contributor to a sustainable business is in itself a major milestone. Think for a second how many businesses you see who can survive the expense of even modest advertising? By setting this as a goal, it forces many downstream decisions and discussions. Even if advertising is never undertaken, let alone proven sustainable, considering it seriously is helpful as a developer.

2. Investing in paid media forces you to build a model of your games sales and life-cycle, because advertising blindly, is a fool’s errand.

A great element of modern-day game development with app stores is great access to daily data. With services like Sensor Tower, Appfigures, and AppAnnie, you can start to look at your own data, as well as competitors data. The problem for independents with data works as follows:

1. Better levels of data are hard to come by. Competitive intelligence data from some of the services previously mentioned sources is usually housed behind paid expensive levels of service. Some of the social network competitive intelligence tools don’t even list pricing — you can guess what that means. Paid sales people on the other end of the line aren’t selling $99.00 yearly subscriptions. Thankfully you can avoid most of these higher paid tiers. So far I’ve paid for some services like AppFigures, or Twitonomy, I’ve spent less than $200.00 for extra reporting and insight systems.

2. It’s easy to get OCD over your daily, let alone hourly sales figures, and feel this urge to act in response. Sometimes this is good, motivation can come in many different forms, but often, it can lead to a distractive, even negative, read-and-react nature.

3. It doesn’t necessarily tell you things that are actionable. Data can be confusing, and deeper analysis, often requires more of it, and stronger analysis skills than most independents have. For what it’s worth, the things I will explain, were arrived at with what I describe to my friend as nothing more than good high-school level math skills, and “course” forecasts.

At the same time, the more I work with data just exploring it without the drive to find “actionable” data, the more it actually results in insights that do drive actionable discussions.

For example, I did the graph in Figure 1.1 to show him how much the overall average daily decline in sales was improving — meaning more long tail! It didn’t necessarily tell him something he couldn’t sort of see just looking at the daily sales/ad data, but it was a reassuring story to see, and it offered additional perspective and conversation over how each update, and promotion was strengthening the title’s long-term horizontal flow toward steadier long-term sales.

Figure 1.1 Looking at the rate of decline in sales as it slows over time providing more likelihood of long-tail sales

4. Building systems to combine various datasets, and create customized work and reporting environments is tough. It’s an area I want to improve, but for now we’re just doing some basic importing of different data-sets into Google Sheets. At some point using something more rigorous, and automating it more with scripts, is the goal. There are some higher-end tools focused on these via a SaaS model, but the expense and time to devote to them makes them temporarily prohibitive.

Independents have to make do with what they have vs. larger concerns. We’re not all MBAs working at P&G or Unilever, two of the most data-driven consumer goods companies you can find, with unparalleled access to data and analysis capacities. So to make use of this, we must build models, often with gaps, but ideally with evidence, and basic logic, and then merge those with strong opinion over what the goals are, and what the ethics will be.

Ranks, Revenues, and CPMs

As of this writing I’ve got some 50+ days of sales data to look at, some other older title sales data, and a bunch of low-level competitive intelligence data.

Lets first understand what the data says, and what it tells you about app stores.

1. Everything is about narrow funnels.

It’s no secret the draw back to app stores is everything goes through a narrow funnel of getting someone to notice your game, pushing them on some device to an app store landing page, and then hoping they click download, open the app, play, and ideally play it enough.

For paid titles the initial download includes revenue, but for free-to-play titles, a download is the first part of a longer journey of monetization using advertising, in-app purchases, or both.

The app store landing page is the first critical juncture, creating three elements of a revenue-yielding value-chain:

Element 1 is pre-landing page, it is everything that generates attention and ideally planned behavior to play, and ideally make a purchase, ahead of the requirement to download.

Element 2 is the landing page itself, and other tasks such as SEO and app store SEO to fish additional players from the serendipitous audience finding your game among any specific platforms’ customer base.

Element 3 is everything you within game, and amongst your downloaded audience (which may be out-of-game) to continue to engage them and generate additional revenue per user, and word-of-mouth/social activity that generates person-to-person promotion for your game.

NOTE: The A in ARPU means average, but I think additional is better in terms of what you should be doing. Rising ARPU is the results of additionalrevenues per user, after initial contact, be it a free or paid user. Additional also signifies that games which earn key revenues from advertising, that this additional revenue is a direct result of sustained engagement which delivers more ads over thecourse of thatengagement vs. specific purchase decisions by the audience. Advertising is passive ARPU, derived from engagement only, not just engagement + transactional behavior.

The funnel in its best form is more of an hourglass. Going in are Elements 1 & 2, through the narrow channel, then coming out the other side is everything from Element 3 above.

2. The data has huge gaps and different sensitivities.

Unless you have time and set up additional, sometimes costlier systems, your data is going to have large gaps, and different sensitivities to understand. For example, Apple provides you data on what direct referral links have led to revenue, usually from specific articles where the publisher uses referral links to generate revenue from their coverage. Others are from third-party app directories. The reports are useful, but they only capture the small slice of users who directly click over, and many users don’t do that. Instead users often choose to open the app store on their device and search for the name of a title they just read about earlier. Lost in that moment is the direct link between PR or other referral efforts and your game.

Revenues for the day will include late settlements, refunds, etc. So it’s not precise either. Sometimes aggregation services like AppAnnie or AppFigures can have hiccups importing correct data from iTunes (which also gets fixed, and may not be their fault) so any given day isn’t really what it is.

In comparing days with similar results say on app page visits, you will find different results in terms of downloads. Some days ad revenue is higher because eCPM was higher, not because your users played more. For country-by-country advertising data the second largest segment of performance is Unknown country. While we can assume it likely breaks down similar in percentage to the known countries when a report has Unknown as its second biggest category that is a gap in your confirmed knowledge base.

There are ways to deal with these problems. One is looking at internal analytics from the app to see what it tells you about the app vs. what store and ad network data tells you. Another simple method is to chart a rolling 7 day average for key metrics as that will help smooth out mini-spikes and dips. It’s also good to potentially remove your biggest launch period out for some basic analysis since, almost nothing will be as big as your launch.

As data expands I also look for “comparables” to learn from. For example, compare days that had the same number of downloads, or the same number of in-app-purchase events. These are interesting because the can help to spark conversations and investigations into your game, marketing, etc. Most of all they give you insight into what’s going on. For example, during a period where we had unusually high conversion rates for IAP, we traced it to a specific review. That may not necessarily give us anything especially great to use, but I do believe the more we understand things the better.

Another tidbit of late. The game started gaining traction in China. However Chinese consumers are not big spenders, preferring to grind out on free-to-play modes. The sheer volume of page views generated by interest from China coupled with lower IAP conversion means that our IAP conversion metrics ended up radically skewed. Hence, we now have to remove China-based downloads from our download-to-IAP conversion formulas to otherwise have a more accurate depiction of our IAP conversion performance where culturally it’s more of a behavioral fit.

Given all this data I did a few things:

1. I took all pageviews on the app store, download statistics, each ad networks’ revenues and impressions, and rankings from AppAnnie and began logging them daily. I created several simple statistics from this:

A. Conversion rates : page views to downloads, downloads to IAP, etc.

Not only did I do singular points of these statistics, I also created “available to convert” statistics both in total and on a 7 day rolling average.

Available to convert is basically:

All downloads-IAP Purchasers. Since the game only has a single IAP, once you buy it two things happen. First, you no longer see ads, so those customers should be removed from the impressions/user category. Second, you can’t be converted twice. Thus, what we want to know is not only on a daily basis how many people converted, but what was the total population of un-converted users left to convert them from.

Some of these stats could be improved through better in-app analytics. For example, the statistic of current day conversion being current day downloads/current IAP income is not great. With the model we’re using it’s likely 50% or more of the IAP purchase from any one day is earned from a user obtained prior to that day. However, I’m not using in-app analytics because of several issues. First, depending on the user, while I know when they purchase, the app store might not settle their payment for several days, second they might take a refund, third, it might be affected by exchange rates. By only using app store provided sales data, and then using some cumulative and rolling averages, I can still get a strong gist of what I want and I’m not mixing two different analytic feeds at the same time.

Remember: independents, less capacities, etc. Yes, someday, over time I want to merge the two systems of in-app, and app store end result analytics, but for now, keeping it simple isn’t the worst outcome.

B. Impressions per user and eCPM: The current game uses four (and possibly soon five) different ad networks, and two different ad products. The nature of ad networks is an entirely other article, but for now understand we’re analyzing the following as it relates to in-game advertising:

I. Gross impressions per-day per network.

II. Per-user impressions on average: total impressions/ad-exposed users

III. The Effective CPM: total ad revenue / (total impressions / 1000)

IV. The Average eCPM per day, per all ad-exposed users: Total Ad Revenue / Ad Exposed Users

V. Requests/Fill-Rate per-day /per country: An indication of where we’re missing out on possible revenue due to poor in-country coverage.

Ideally we should also be measuring subsets of our ad exposed users, specifically Ad Revenue/DAUs for that day, but again, that’s a different set of data than the gross data I’m working with. By dividing by our entire user base, even those who drop out, we’re doing two things. First, we’re looking at lifetime expectations for any user. Second, we’re penalizing ourselves a bit for users who drop out from playing, and not forgetting that bringing them back to the game, can have a great impact on our overall audience availability, and ideally ads served, and revenue generated.

For all these numbers we compute each individual ad network for comparisons, and the total for all ad networks.

C. Total revenues & revenues per X: Once all revenues sources are available, I look at not only revenues per day, but most importantly revenues per download, and revenues per page view. Ultimately, if all revenues must start with someone actively hitting the app’s download page, the ability to convert from that first-visible metric is key. If page-views rise but revenue doesn’t rise in relation to that, each page view will be worth less and less. If it’s the opposite then it might indicate the viability of blindly getting more people to that page whatever means possible.

D. Cumulatives & Rolling Averages : Given the day-to-day ups and downs of app stores and ad networks, not to mention general 7 day retention, I’ve found a simple rule-of-thumb is using 7-day rolling averages for most daily stats results in a better set of numbers to work with. It’s a simple smoothing process, and helps to create more honest comparisons of day-to-day points when looking at comparable periods. I also look at the overall average since inception, or some other long period of time, so I can compare what any given day is doing vs. what has been the overall average— is it above or below? Finally, I look at every new day, and mark (using conditional formatting features) if this day is bringing the average down or up overall.

The goal is to have simple data points that help with identifying positive or negative anomalies daily, or in short periods, because those are times where you can learn more. Ideally you can trace these moments to something that happened which might lead to further actions you can take or learn from.

E. Curve fit forecasts : A key issue for any developer and their total revenues for a title is the notion of sustainability and the long tail. Much has been written about the long tail, which I won’t recreate. Instead, let’s use sustainability, best described in the following contextual terms:

The long term capacity for your game to return meaningful daily revenue for as long as possible with less and less tangible investment to maintain an acceptable level of sustainabile daily revenue.

In fact, what I’d argue is the goal for any independent game developer (and really any independent app developer) is that this really is the ultimate floor. If you can earn $20/day from a title for four or more years with minimal effort after its biggest sales are behind it, that’s the result of a successfully achieved sustainable position — especially if you’ve been very careful with your total input — which unfortunately for many indies may not be the case.

What this says about what your inputs should be is also, an entirely different article, but the short version is here:

My opinion is most indies fail to achieve sustainability because they don’t plan for what it looks like at the outset and don’t know how to achieve it in the end. This most often manifests itself in the form of too much expense in terms of developmental hours, loss of other income, and team size/tactics.

In general I can’t imagine broad swaths of the indie/hobbyist community resting its survival on sustainable titles that require long tails in excess of $20-$50 a day in daily income. There is for now, no data, I’ve seen showing that there are enough slots for games earning in excess of $200-$500 a day to support the population of indie/hobbyist developers as it exists today.

In terms of my friend’s game, what I have done is taken the first 60 days of revenue and charted it out looking at daily revenues and number of users, and the day-to-day raise or decline. With that chart, and some other futzing, I’m able to calculate an equivalent daily-average-decline rate long term for the title. This provides a course ability to project future revenue and over time it’s updated. While it would be nice to have a title that hits a self-sustaining life-cycle in the top ten of its category, the reality is people move on to other titles, and App Store featuring isn’t a permanent position for those on the outside of top 20 category charts. So plotting out your own game’s downfall is humbling but important.

Hopefully, the success of the title so far, sees the bottom end of the curve flatten out much more as time goes on, and settles into that $20+ level of sustainability for a long time.

Another important reason to do this, isn’t just ideally peace of mind, but the great confidence you have in your unaltered projected revenues, the more you can decide if you want to risk some of that future revenue on boosts to your marketing that results in better near term revenues and ultimately overall as well. It’s a risk, but clearly by definition, a calculated one.

2. I Created some Models for Sustainable Advertising

As said earlier a key goal is to look for evidence, and build a model that justifies the use of paid-media to achieve sales, revenue, and ultimately profit growth for the title.

To do this I look at a combination of the sales data, and data gathered from other sources. I have constructed two approaches to look at:

Approach #1 : Looking at Per 1000 Revenues

Approach number one is the total amount of income per download (aka user), and then total income per page view on the app store. Both are different derivatives of ARPU (Average Revenue Per User) or even ARPDAU (Average Revenue Per Daily Average User). I then calculate that to a CPM metric so it’s easy to understand the relationship of a CPM ad rate, and a RPM revenue per 1000 users rate. It’s a nice 1:1 comparison.

There are a some issues with this approach worth discussing before getting to additional conclusions:

1. I’m looking at advertising costs on a pure CPM basis. CPI (Costs-Per-Install) rates are high because of three things: first, direct CPI purchasing offers a certain level of security that larger companies which have games or businesses that generate a higher ARPU can afford to pay for. Second, there is an over-abundance of demand overall, fueled in part because of cheap money flowing through VC and private equity paying for higher user acquisition costs. You’re not just competing for CPI with Clash of Clans, but Uber, and others. Third, to justify CPI, you need to do additional engineering to track it all properly.

This means for developers who want to advertise, competing on direct CPI buys is really hard, unless you can see an ARPU that justifies it, and that’s only likely if you’re engaging in chasing down revenue through some of the less ethical elements of FTP (free-to-play). Instead, smaller developers, those with titles that aren’t engineered for large average customer revenues will have to consider the risk of extracting installs through CPC and CPM purchases.

Since the assurance of install success with these campaigns drops, and inventory is more plentiful relative to demand, there is opportunity to purchase advertising at much smaller CPMs. As such, you’re counting on the results of your funnel to generate the assurance of converting users at a profit that a CPI ad buy would give. If your funnel’s monetization is not good enough, then it will fail.

Given the cost of CPI relative to my friends revenue model there really isn’t any other choice.

2. Extracting the costs of advertising so we can consider various forms, because advertising cost data is hard to come by. This is not only because of the seismic movements in the adtech world that constantly take place, but also because of the normalized nature of what essentially is a global daily, hourly even, auction system around advertising.

Furthermore, much of this auction pricing is opaque, hidden behind ad marketplace systems you have to sign up for, and then individually understand. All of them have a bevy (understatement!) of choices for targeting, customizing, pricing, and sometimes investigating result forecasts.

To this day, I still don’t feel I fully understand how to optimize an ad buy on Google. Facebook’s system is less complex, but still has a ton of options. Apple’s iAd is much easier to use, but its forecasting for anything other than banners isn’t clear up front. I’ve yet to unpack other systems that exist.

From an indie perspective purchasing advertising to support your title is a daunting task, and that’s even before you decide to even begin to consider the risk of laying out actual dollars.

NOTE: Most of the previously mentioned indie stories I’ve seen, have tests of advertising early in their marketing campaign, and it usually is less than $500.00 spent. The results aren’t promising, but it’s important to state, when I’ve seen this discussed, the buy is pretty blind, has been mostly adwords based, and likely too small a sample size.

Looking at the data of the most recently launched title it is earnings $XX per 1000 page view on iOS App Store. Per 1000 users the game is earning ~$X00. The multiple of these two numbers is essentially 100/conversion rate of views to downloads.

What this means is that for advertising to work and be of value to the title the following formula has to be true:

Cost of 1 CPM Unit of Advertising < Resulting Revenue from Users

This is where the performance of the game, and the performance of advertising conversion rates all begin to align, and the funnel nature of digital businesses becomes all-encompassing.

According to Adivity the average CTR (click-through-rate) in the U.S. for moPub is 2% for iOS and .8% for Android. For other ad networks it changes. This is a composite of all ad types so again, it’s a bit opaque. Mobile Interstitials are reported to have a 5.5% CTR. Video ads would ideally be higher as well, but likely cost prohibitive. Let’s assume then that mobile CTRs run the gamut from .5% CTR to 7% CTR depending on many factors including audience, creative, targeting, competition, time-of-year, and what color the sunset was last night. Assume for a moment that you could assemble a mix of advertising that performed at a 2% CTR.

To get 1000 impressions to the App Store would require 50,000 impressions. That means I need to purchase 50 CPMs and for that, if all else held true we’d make $XX back. The average eCPM on MoPub in the U.S. is $2.09 according to recently published data, so $50 CPMs is $104.50. Now I don’t have to even tell you how much my friend is earning for you to figure out that $104.50 minus any two-digit revenue number is a loss.

Looking at it from the revenue per downloads perspective, to achieve 1000 downloads requires hitting a much larger multiple of page views to the App Store. If we even hit a 20% download conversion rate from the App Store page we’d need 5000 app store page views to generate that would take 250,000 CPMs assuming a 2% CTR. 250,000 CPMS is going to cost $522.50. Currently my friend is making some level below that per 1000 downloads.

Approach #2 Looking at Rank to Revenue Relationships

With App stores you get rank data. Given enough time, and with the fact that an update, plus some App-Store promotions helped to push the title back in-and-out of higher ranks, I was able to use some basic regression analysis to rough out numbers showing the relationship between rank in the key games subcategory for the game, and resulting revenue. There is expected variance, but using some averaging, and an exponential curve fit, the confidence level for the final set of data was pretty strong.

The data shows the value of moving up from a rank of 200 to 100, or 160 to 80, or 80 to 40. By looking at this data it indicates that if several thousands of dollars successful improved the rank of his game from 160 to 80 the resulting rise in revenue would likely be recouped. This assumes the comparison is a weekly one. The difference in being ranked around 160 for a week, and being ranked 80 for a week, in the game’s specific App Store game sub-category was just under $5K. If the move resulted in going from a sustained ranking of 160 for a week, to a sustained ranking in the 50s in the U.S. App Store, the gain in revenue would be closer to $10K-$12K.

NOTE: These figures are just for the U.S. iOS App Store which is the #1 market in the world in terms of total gross revenues. They are also for sub-categories within the games category and shouldn’t be confused with the recently published data on costs to reach top tier status overall — as those numbers are much higher. Aside from the U.S. App Store we could also look at some individual countries and see if some pockets outside the U.S. are worth calculating similar values for, but it’s likely no more than five have a sizeable enough aggregate audience to make this work useful.

Keep in mind that the goal for this analysis was to determine how much a sustained period of ranking was worth. This doesn’t mean the ranking itself achieves the results for you. Yes within the top 25, or 50, I am pretty certain it is true the higher you get, the more that rank brings with it increasing organic visibility on app stores. Overall, my sense is rankings outside the top 50 or so are likely to be indicative of what you’ve done with your game, and its related visibility/word-of-mouth and advertising, and not the result of self-fulfilling app store presence. There may be some, as some consumers plunge through the top category charts in search of titles that catch their eye, but so far I have no sense of how much each rank within an app stores larger top 200, 500, 1000 contributes-by-itself to future sales.

When discussing moving an app from 160 to 80 in rank, my goal is to use that to forecast what that likely brings in future downloads, installs, and revenues so we can see if the cost of getting there equals the return.

I also have no idea of the cost of getting from 200 to 100, 160 to 80, or higher. I can assume it has some level of exponential curve correlated to the activity exhibited by titles at those ranks but I have no certain sense what that number is. And as I said earlier, the only public forecasts are for much higher tiered categories and even then, those numbers are under dispute. Furthermore, issues like creative, ad size, demographics targeted, etc. will all affect the resulting outcome of any ad campaign.

So Approach #1 shows a likely financial shortfall in its model, and the rank raising model shows the possible return but not the definitive cost. Both have enough uncertainty, and unanswered questions. So it would seem that we’re coming up short, we should end the analysis, and the conclusion is don’t do any advertising. But the reality is this:

The gap between your game’s revenues, and revenues that enable some level of sustainable advertising can fall into three categories:

With so many games being released everyday, and so many other non-game applications, media, and social media efforts vying for attention of users, there is only so much attention any given release can hold, for so long. While some developers are inventive enough to find new ways over time to keep their game visible, almost every such effort other than advertising requires approval of people other than the user, be it a journalist, or some other gatekeeper. Advertising, on the other hand, has less restrictions, and if successful, can extend the sustainability of a title, raising its audience, and profitability for some time beyond the capacity of free-media, and updates.

Achieving a game whose profitability outstrips your probable cost and risk of advertising makes it a more favored candidate for longer term success at higher levels of revenue and ideally profit.

Making the Leap

Advertising by itself it is a risk. Despite years of clickstream data, and the rise of deeper analytics over response rates, there are still no guarantees. Many elements could conspire for/against any particular developer. In the cases where creative is great, and targeting is spot on, and those reached by the advertising are receptive, you may be able to be successful with a gap that isn’t quite closed or slightly closed mathematically. In negative cases, even if your title mathematically seems well positioned for advertising it may still not return on its investment.

Making the leap for advertising, especially for smaller developers, is always going to be risky, and difficult, but ultimately, as stated previously, this goal provides a level of rigor to shoot for even if you never pull the trigger.

Everyone wants their games to do better, to reach the most audience, to generate impact among their players, and ultimately to profit well from privind that value. My search for an overall sustainable model including advertising just tries to put some numbers to it.

In the case of my friends game there are no some very concrete goals and hypothesis to consider as follows:

The first hypothesis is good — it may be proven wrong, but it’s a solid hypothesis created from observable data with reasonable confidence.

The second goal, is a question of work and faith. So far my friend and I have done some optimizations to the game’s App store landing page, improved social sharing on the game, and continued to look for PR opportunities. There are other in-game actions planned to try and improve IAP conversion, and currently we’re learning how to optimize ad revenues both overall and on a country-by-country basis.

What’s most important about everything I’ve explained up to now is we actually have some semblance in numbers and models to guide what to do, why to do it, and ultimately what it might be worth. Do I wish it was way better, given how many gaps I’ve explained? Yes. However, given everything we’ve cobbled together to get to this point, the learning and perspective has been tremendous and will help his current title and follow on works. This work breeds higher levels of certainty, despite gaps, and deeper levels of perspective, that ultimately should make things less paralyzing.

I don’t foresee overcoming the gap between forecasted revenues and forecasted costs of advertising entirely, the game just does not monetize any individual user at super high numbers like a Candy Crush Saga does — and that’s by choice. What’s left to ascertain is, how much flywheel effect can come from advertising?

Assume several thousands of dollars was spent on a CPM-style campaign to acquire users. The campaign would use either low-effect banners, or slightly better static/animated interstitials, video ads are too expensive.

The question will be “does the advertising move the needle enough to produce four pools of revenue that we can measure, and once we do, it outstrips the spend?”

What we’d be looking for are the following:

In all these cases, and others I have discussed, there are decent hypothesis that represent reason to believe that a close gap between revenues and cost-of-advertising might be overcome.

Already some improvements, born from looking at the same data, and seeking to quantify the risk-return of advertising have resulted in excellent improvements. This includes the game, its marketing, and general understanding of the systems that control who plays his game, for how long, and at what financial benefit.

By optimizing everything we can, within reason, the stage is set to advertise.

Depending on your corporate structure consider that the risk of advertising, at least in trial form, is a shared risk. Assuming you’re reinvesting some level of incoming revenue post-launch, those funds spent are not net-of-taxes. So if your tax rate for pass-through income is, for sake of argument, 40% (don’t forget self-employment tax in the U.S.), then risking $5K on advertising is really “only” risking $3K in personal income, net-of-taxes.

For many indies, income net-of-taxes is the true sense of what’s at risk, provided your overall business obligations are low — which by default, as an indie, they should be.

I can’t say yet what the results will be. The decision to spend is not with me, but ideally I’ve made that a less stressful decision because there is a greater sense of the risk and possible return. if/when he does, I promise to bring back the results, and explore further what worked and didn’t.

NOTE: For “I told you so” sakes, I predict that the advertising, provided the creative is decent enough, will break even in the short-term or just fall short, but that it will measurably improve the game’s overall revenues over a longer period of time, as those new breakeven users add long term to the games word-of-mouth related revenues. Initial returns on advertising will be mixed, to break-even, and we’ll potentially recycle those revenues for a slightly longer period of time, and then see a bit of a pro-longed lifecycle over the succeeding weeks and months as those new users, bring in new users through word-of-mouth. Depending on how much of a boost that last bit is, will determine if the effort to advertise overall was worth it.

The positive x-factor will be if rank growth turns into a super-positive upside flywheel, and the negative x-factor will be if general apathy toward display ads (banners, or interstitials) in mobile apps is just too hard to escape.

App Store Gaps or Something More?

While so much of this essay has been about advertising, the true title of the essay, is about the gap that exists on app stores for indies.

The specific gap I’ve outlined is around the bigger question of can advertising help my friends title earn more money, for longer periods of time, and ultimately help it have a more sustainable existence on app stores?

But the larger question that underlies this all is the overall gap that exists between most indies today, and their own confidence, and sustainability.

True long-term return is best achieved through sustainable approaches. Punching a lottery ticket can still happen, but if all you’re around for is the purchase of a single ticket, you need to be infinitely more lucky than the developer with 5–6 sustainable titles who gets to try again, and again.

When app stores were fresh, visibility was easier. As more products entered the fray, and stores didn’t necessarily respond with a bevy of systems to improve the chances to breakout beyond the narrow windows of visibility afforded by those same stores, indies have arrived at a point where it’s much harder to flourish on game alone. Sure it’s possible, and that’s one of the charms of the games business, and the attraction we heap upon brilliant indies, but this is but a small swath of the total industry. And it’s not the subset of the industry I’d want my kids hanging their hat on should they tell me they’re going to do what Notch or Jonathan Blow did.

Today, modern game development, even for indies, requires having not only a good game, and a clear audience target, but you need systems that measure many data points, and means to quickly adjust and tweak related revenue systems. You need models by which to make continuous opportunity cost decisions, you need in-app systems to cross promote titles, systems to mediate multiple advertising networks to optimize revenues amongst real-time advertising markets, and much more. These are not the sexy things that we associate with indie success but they are the reality now.

While I applaud most indies for not jumping down the ethical trap many FTP+microtransaction titles fall into, it’simportant to know implementing those systems effectively means a CPI model (which is expensive but has more certainty) requires expensive systems and work to do — let alone ethical landmines to navigate. It’s as much hard to do, as hard to stomach.

For many indies, development has becomes publish your game, send out some PR, and hope for the best. There have been many talking about the so-called Indie Apocalypse. The only Indie Apocalypse I see concerns approaches, not indies themselves.

There is a real Indie App Store Gap. It lies between big companies, and smaller ones, who are operating with some, or more of the systems and models that let them chase more certainty, and sustainability, let alone high-level success, and those indies who aren’t doing it, or even trying.

My advice here isn’t only about advertising or other higher-order marketing efforts. My advice is to identify, understand, and be realistic about the many gaps in knowledge, data, and capacities that exist for smaller game developers. Carefully build models, and test them, with the goal of increasingly gaining control over systems that govern sustainable success for you (not just your games) over the long term.

As you do, please share some of this, so we those that are more deeply and rationally committed beyond the dream of a hit game, can benefit as well. Together maybe we can close some of the gap, so everyone has a slightly better chance at sustainable success.