Mobile Evolution - Thoughts, insight, and experiences related to the mobile software industry

Developing for mobile platforms such as iPhone, Android, and Blackberry, is hot right now. The choices have exploded in recent years. For micro ISV's (independent software vendors) like ours, Creative Algorithms, many positive things have been happening in the mobile space. Barriers to entry have lowered on some platforms, so getting personal with carriers is unnecessary; on-device application stores have become standard (with lower commission rates, increased customer awareness of apps, and ease of installation and purchase); upfront fees or memberships have become reasonable, and the market size for smartphones has been increasing. These improvements, however, have made it more difficult to choose which platform is the best fit, or which has the greatest potential for supporting oneself. To make the choice more difficult, the numbers keep changing. For example, the numbers have already changed since the start of research for this post--sales volumes for Android handsets have increased, 30 new Windows Mobile phones are now predicted, and 20k more apps have entered the Apple AppStore.

An important part of the decision is the numbers, but each mobile platform also has their negatives--Apple's submission practices, increasingly difficult discovery in the AppStore, and penny-candy pricing for apps. Android has few released phones with its platform, which means low volumes, plus its Market can be difficult to find on its phones. The new Palm Pre (webOS) has very low initial volume (as compared to its competitors) and has only just now opened its online store for submission of paid applications. Blackberry World must be installed on the device before use and the installation of apps isn't streamlined. Windows Mobile's look and feel is outdated, and its new new app catalog, Windows Mobile Marketplace, is not open, just currently taking submissions. Symbian is downright confusing--too many options, too many phones, and entry pricing is complicated and expensive. The Ovi store is promising, but consumer awareness for apps needs more promoting.

Each platform also has varying developing environments, but that is not the focus of this post. What business-side information can help small developers determine which path to strike? Is it possible to earn a living as a mobile software developer and on which platform is this goal easiest to achieve? This post will provide a valuable platform comparison and a foundation on which to determine the path for reaching self-employment goals as an independent mobile software developer.

Cost of Entry
If a mISV is bootstrapped, as many are, the cost of entry can be important. Costs vary from registration fees, signing fees, and equipment required to develop. For purposes of this discussion, general business expenses will not be covered here, such as incorporation fees, internet service, etc. in order to focus on the differences between the major platforms. Contracted development and design costs will also not be covered, as this can vary, plus the subject of this post is geared towards a do-it-yourself developer. The post will provide a basis for a single salary, but if you contract out your development and graphics, you should be able to adjust the goals to your personal situation and recalculate.

Costs specific to each mobile platform will be discussed. The focus of the comparisons will be on the six major platforms that are hot right now: Google's Android, RIM's Blackberry, Apple's iPhone, Nokia's Symbian, Microsoft's Windows Mobile, and Palm's webOS. A complete summary of the data used for the analysis, plus information on the smaller (mostly Linux-based) platforms, is included in our Mobile Development Platforms matrix. Fixed costs are defined as expenses unaffected by sales volumes. These costs remain without any sales, and become negligible as volumes increase. Variable, or incremental costs, are directly related to sales volumes. Each sale incurs a new expense per unit.

Company Registration Fees
Right now, Symbian is the only platform that requires an upfront company registration to develop on its platform. According to one developer, Tam Hanna, the cost can range from 200-2000 Euros per year, depending on what type of company. The rest of the platforms are free--you can download and use their SDK without paying a fee. However, some platforms, such as Apple, require a fee to sell in their AppStore, which is the only sanctioned distribution channel for selling iPhone apps (jailbroken phones are the alternative). Some platforms also charge "certification fees" which is verification that the app has been tested. For example, Symbian charges 200 Euros/year for certification, plus you also have to pay for third-party testing of your application. Paying these certification fees can gain you access to certain device features, but is not generally required to sell an app, so these costs are not included in this analysis.

Signing and Registration Fees (Fixed Costs)
Signing is a digital signature, a method of authenticating digital information and its source. Some platforms require signing in order to install your applications on devices. Some make it optional, but warn users extensively that the app has not been signed and could be malware. Some only require signing to get onto their on-device store. For most purposes, if you want to reach the frightened masses, apps should be signed. Some platforms allow for self-signing, some require you purchase 'keys', others get more involved in the process. For many of the platforms, signing is very convoluted. Let's look at the costs involved for signing in order to be on the on-device stores (AppStore, Android Market, etc.), since a presence on this distribution channel is essential for good sales volumes. Here's a breakdown on the costs for signing:

Android is the least expensive and Symbian is the most expensive (the minimum company registration fees are included in its total). From the above chart, the first year and first app costs cannot be the only consideration. Blackberry is costly to start, but they do not charge additional fees until you've established a large library of applications, and by that time, if you've made good sales, the cost should be negligible. On the other hand, Windows Marketplace charges per app, and per year fees, so the costs to sell could add up quickly as time goes by, even though the initial fees are palatable. A small recurring fee can be a good thing--it gives the OEM's a way to weed out the outdated titles from their stores. The classic ESD's (electronic software distributors) have very dated catalogs, but have no clear, fair method for cleaning house. Android is the only modern platform that does not charge a recurring fee to sell on its on-device store.

One more consideration to note for fixed costs is the cost of a Macintosh computer, which is required to compile your application for the iPhone. Mac developers will already have a Mac, but anyone new to developing in Objective-C/Cocoa will most likely not already own a Mac; it's estimated that only 8.9% of computer owners own a Mac over a PC. A refurbished Macbook can be estimated at $900, which could be a cost barrier to a new, non-Mac developer.

Commission (variable costs)
Sales commissions for selling on the on-device stores is more uniform than the fixed costs. Apple set the standard and most OEM's have followed. One exception is Blackberry, which only charges 20% commission. However, both Blackberry and Palm webOS pay their developers via PayPal, which charges 2.9% + $0.30USD per transaction. This method may be good for non-US developers, where wire transfers can be costly, but it does increase expenses for US-based developers.

Breakeven Points
The breakeven point (BEP) is the volume where net revenue (price minus variable cost) equals fixed costs: ((Price - Unit Variable Costs) x Volume) = Fixed Costs. Fixed costs and unit variable costs have been determined; now pricing must be considered to complete the equation and solve for volume. All stores have a minimum price point of 99c except for Blackberry World, where the minimum price is $2.99. A quick analysis of the top selling titles can provide a general average price for selling software in each store. Let's do a quick study at how much volume needs to be generated to cover selling expenses, both by using the minimum prices and average prices in each storefront:

The volumes over two years to recover costs are quite low. In general, the results show little risk of recovering initial fixed costs for selling on the on-device stores, especially at average pricing. The iPhone is the highest hurdle to climb, but that is mostly due to the lower average pricing. Even if we add in the Macbook, the volume to recover the cost is two sales/day of a 99c app over a two year period. The Symbian apps require more volume at 99c prices, but the average price is much higher, so more likely the BEP volume is closer to the volume with the average price.

Earning a "Paycheck"
The next step in the analysis is to look at what volume is required to bring in a decent salary, equivalent to a software engineer's day job. This salary will vary, but we'll use $100,000 for this study. Using a similar method for break even analysis, we'll now compare the volumes to achieve a yearly salary of $100k. The formula per year is Volume = (Fixed Costs + Salary) / (Price - Unit Variable Costs). Let's look at the results:

These results are more interesting. It's now very obvious that the fixed costs are negligible when you're actually making a living as a mobile developer, since the salary is a much higher cost per year. Selling one app at the minimum price in any store is going to require about 150,000 sales per year, with the exception of Blackberry, where the minimum price is higher. 150k sales per year amounts to 410 sales per day. Keep in mind that the number of sales per day must be maintained over time. We'll evaluate if this volume is really possible by considering required marketshare in the next section. An interesting observation unfolds with this analysis--the concept of the average price. Without considering selling factors and market conditions, Symbian is the most likely platform to produce the desired salary, followed by Windows Mobile, mostly due to higher average selling prices. Blackberry, Android, and Palm webOS volumes are a bit higher, with about 33 sales per day at average pricing. The iPhone race-to-the-bottom pricing is the most concerning--even at the average price, 157 sales per day are required. The required volume is almost 5 times as high as any other platform. However, we need to take a look at the big picture to determine if it is possible to achieve the desired volumes, considering each platform's market size.

Mobile Platform Market Size
Each mobile platform has a market size for its established base, which is changing daily as each hardware manufacturer strives to increase its marketshare and also to grow the total market. The mobile market for smartphones has escalated in the past two years as manufacturers have moved from simply PDA's to smartphones and as new players, such as Apple and Google, have entered the marketplace. The iPhone's success has raised public awareness of smartphones and more importantly for developers, third-party software. Smartphones are quickly becoming mainstream and replacing the standard feature cell phones at an accelerated rate. Quarterly data is much easier to determine, but a good estimate of the established base is available if you search hard enough and extrapolate. Despite the high variations and room for error, there is enough information available for a quick comparison analysis of market potentials for individual developers. Here's a snapshot of the size of the current user base, number of competing applications on the on-device stores, and a ratio of applications to users for each of the mobile platforms:

This chart provides interesting insight to the markets. iPhone and Symbian smartphones are leading with the largest user base and magnitudes above the fledgling Android and Palm webOS product lines. Blackberry and Windows Mobile have healthy volumes around 30 million handsets. The lower volumes of Android and Palm webOS put a cap on what is achievable for each title, which does not spell good news for developers. What is really telling is the number of apps, compared to other platforms, of which the AppStore is the obvious leader. However, a higher number of apps is also not a good thing for developers--it means more competition in the marketplace for the user base. Discoverability becomes an issue and the need for a higher marketing budget becomes evident.

When you combine the two data points and look at the ratio of users to apps, some interesting numbers shake out. Using this ratio, Blackberry and Symbian are clear winners in the discoverability column. However, one must keep in mind other factors that affect sales. For example, Blackberry World must be installed manually on the devices, so the actual base volumes will be lower. For Symbian, the existence of apps is not well-marketed, so that fact will have an effect. Palm webOS has a high ratio of users per apps, but its low volume caps the maximum potential for volume. Windows Mobile actually has a nice ratio and user base, but its interface has become dated, so its need for a facelift will affect future growth until rectified. Android has a comparable ratio of users to apps as iPhone, but its user base is low in comparison, so its market is also capped. However, a large market increase is projected with the release of new hardware in 2010, so this fact may soon make Android more appealing.

Finally, let's take our market comparison and look at what percentage of the market is needed to achieve our desired salary.

What amount of marketshare is possible? Many factors affect what is achievable. How mainstream is your application? What kind of competition do you have? Many anecdotal stories are available online. Some look at the top successful apps and estimate 3% as a cap in marketshare, others are less optimistic. Exact attainable marketshare requires further research to determine. What the above chart does is to compare the platforms and to highlight any platforms where marketshare is impossible to achieve. At this point Palm webOS does not look promising. Android is also magnitudes above the other platforms. Windows Mobile, Symbian, and Blackberry are comparable in required marketshare to achieve the salary goal. The numbers for iPhone are in the middle of the group, but still reasonably possible.

Conclusions
Earning a living as a mobile software developer is possible, but how successful you are will also depend on your additional expenses. Many factors will determine which platform is best for you, but running the numbers highlights some key factors. Pricing is important--pricing closer to the average (or above) will greatly affect success.

The maximum user base is a factor--Android and Palm webOS will be platforms to watch--if their volumes increase, they will become viable platforms for serious developers. These two are just not there yet. Some of the older mobile platforms, such as Symbian and Windows Mobile have good potential, but each has its own obstacles to topple. Symbian is a difficult program to get into--from higher startup costs and approvals, to an ancient fragmented operating system that needs a complete overhaul. Windows Mobile is and easier platform in which to get started, and has good tools for development, but its interface and associated hardware is dated and needs a refresh, similar to the old PalmOS platform.

Blackberry has great potential, from a good sized user base, an achievable marketshare, and a less-than crowded app store. However, the mechanics of delivery and awareness of the apps is necessary to help the third-party developer achieve its goals. iPhone is the most modern, growing platform, but it's massive growth has severely limited discoverability of applications on its single-source of distribution, its AppStore. If its user base volume starts to level off, but its application numbers continue to climb, plus pricing doesn't improve, the potential to maintain a living will begin to decrease. It's hot now, but Apple must continue to monitor the market to maintain its desirability to developers.

Each developer, depending on the size its organization, will have its own goals, additional expenses to factor in the equation, and other non-quantifiable factors that will affect the choice of platforms. Each platform has its idiosyncrasies, but the numbers can provide a good starting point. The data highlights the viability of the platform for supporting a mobile software developer who is serious about making a living in this field.