Selling software: Where have all the business models gone?
As traditional desktop and notebook PC sales slow, hardware manufacturers are not the only ones feeling the impact of a shrinking market. Traditional PC software vendors, those who sold what were once installable software products, are seeing their markets evaporate.
Taking the place of traditional software products so common during the Windows era are two relatively new categories of offerings: mobile apps and cloud-based software-as-a-service (SaaS) applications.
In this article, we're going to take a look at the traditional software application market and scope out which of the newer markets might have a better fit for the traditional PC independent software developer.
A few weeks ago, a reader asked me about how he might build an app business. I answered him in some depth in "So you have an app idea and want to make a bajillion bucks." I cited a number of mobile app developers who were having difficulty generating viable business revenue from the app store model.
That got me thinking: What are valid business models for today's software product companies? Back when I ran a software company, we sold packaged software products in retail shops; we duplicated disks, printed manuals, and shipped boxes all over the world. That was a long time ago.
The problem with the app market is that prices have been pushed down so much that you can buy an app for less than you'd pay for a can of soda.
Today, with the exception of console video games, the packaged software market -- where you put software in a box and physically ship it -- is obsolete to the point of ridiculousness.
But what's the best new model, especially for smaller, independent software companies? At the high end, of course, there's Oracle and Microsoft and SAP and such. But, for decades now, there has been a vibrant market of PC software vendors developing unique applications, tools, utilities, and tweaks -- and making a living off of it.
These days, of course, the desktop PC market is cratering, and with it are going the revenues of PC software developers. Back in the eighties and nineties, when I ran a software publishing business, we sold packaged products for prices ranging between $49 (for an icon editor) and $395 (for an expert system development environment). In today's dollars, that's roughly $87 to about $700, depending on how you calculate inflation and relative value over time.
No matter. The point is, we rarely ever buy packaged unit software in that price range anymore. Even Adobe and Microsoft have mostly given up on the packaged software model and are selling most of their application-level software through subscription services.
In addition, the growth of iOS and Android cut the bottom out of the price points for packaged software. Granted, the app store model also cut out a lot of the middleman and production costs as well.
Let's discuss the app store model for a minute. Back when I sold software through retail outlets, we had to discount our software 60-70 percent for sale to distributors.
So, using 1990 money, if we sold a $49 package, my company would only get about $17. If we sold a $395 product, we'd only get $138. The distributors and retailers would get the rest of the package list price.
The 30-40 percent of list that we earned had to be used for per-unit production costs (disks, printing, duplication, packaging, shipping) as well as contribute to the overall company administrative and software development expenses. These were non-trivial expenses, especially in terms of cost-of-goods-sold and upfront production costs prior to shipping the first unit.
By contrast, most app stores today take only 30 percent, rather than 60-70 percent. As a result, software vendors keep a lot higher percentage of the take than they did in the old retail software days.
But the problem with the app market is that prices have been pushed down so much that you can buy an app for less than you'd pay for a can of soda. For the app software business to be profitable, then, you need to sell as many apps as the soda companies sell cans of soda. For most app developers, that's not happening.
In fact, most smartphone owners never, ever download an app. According to a study reported by our own Adrian Kingsley-Hughes, "the entire app ecosystem is being driven by about one-third of smartphone owners, with seven percent of owners downloading nearly half of all the apps."
I'm part of that seven percent. I buy a ton of apps. But I can get 10-20 apps for $25, far less than my olden-day customers paid for a single copy of my old software products.
The economic model is different, of course. Let's dive into it next for apps, and then we'll move on to SaaS-based products. Next: Comparing packaged products to mobile apps.
By the way, I'm doing more updates on Twitter and Facebook than ever before. Be sure to follow me on Twitter at @DavidGewirtz and on Facebook at Facebook.com/DavidGewirtz.
For the moment, let's ignore the relative value-of-the-dollar-over-time issue and just look at what it costs to produce software, comparing packaged products to mobile apps.
Both revenue models had software development expense. You can argue we have better development platforms today, or you can argue we have more development expenses because of the need to pay to produce high-quality media, but for our purposes, both models had software development expense. We can call that a wash for our rough comparison.
Packaged products had all the production and shipping cost involved in producing physical goods. That was risky and expensive.
Let's say you had three products. Without a doubt, one product would have sold better than the others. That's just the way of things. In the packaged goods days, you had to pay to produce a minimum quantity of units, and you often wound up producing the same size production runs, regardless of whether one product vastly outsold the others.
The idea of freemium software-as-a-service is that you get to use a product for free, but if you want to take advantage of additional features, you have to purchase the service.
That was wasteful and expensive.
Today's app business beats that completely. There's no physical production cost risk whatsoever. Of course, the dollars-per-product sold are different, but so is the size of the total available market.
On the other hand, there's the support problem. The app model doesn't support support. There's just no way you can spend 40 minutes on a phone call with someone who bought a $3.95 product for which you get all of two dollars and seventy six cents.
As a result, while app vendors often provide some level of online customer support in the form of FAQs or forums, there's just no money (as in no way to afford to pay someone) in providing quality customer support in the app business.
While support was always an uncomfortable expense in the packaged software business, it was at least something that sales of the goods could generally support.
Tech Pro Research
Then there's marketing cost. When I started in the software business, there was no Web. We bought ads in magazines and in the retail stores. That was not cheap. On the other hand, you could buy ads in three or four magazines and saturate the market.
Today, getting visibility in the app store is extremely hard, and so is getting visibility online. All the SEO in the world won't make your app discoverable if you're not picked as a showcase by Google or Apple. Yes, coverage on various sites, affiliate programs, and cross-app advertising will generate awareness, but it's still very difficult to stand out in today's app market. There's just no way to throw $50-100K at the problem and know you'll be seen by everyone who matters.
On the other hand, let me also point out that most retailers didn't like to pay their bills. One thing you can say for the Apple of today and Google is that they regularly pay out for app store sales.
Old school software retailers and distributors often didn't. It could take 90 days to six months to get paid, and some retailers just didn't pay at all. Collections agents and the expense of collections was another expense, as was factoring (pre-selling your payments) for larger orders. The problems of collecting from distributors were legendary among most mid-sized software companies.
Now that you've got an overview of the market differences, lets enumerate the characteristics of old-style packaged products vs. apps, and then move on to SaaS freemium products.
| Packaged Products | Apps | |
| Development cost | Yes | Yes |
| Physical production costs | Yes | No |
| Physical shipping costs | Yes | No |
| Risk of over-or-under producing inventory | Yes | No |
| Per-unit-sale revenue | Much higher | Much lower |
| Unit sales revenue helps cover tech support cost | Generally | Not even possible |
| Percent of selling price received by software vendor | 30-40% | 70% |
| Time to payment | 90+ days or longer | 30 days or so |
| Ability to reach total prospective customer base with marketing | Expensive, but possible | Impossible |
Both business models were as much art as science, but while apps reduce the barrier of entry considerably, they reduce the barrier of entry considerably (to the point of almost zero), wildly increasing the competitive noise.
As I discussed back when I wrote True confessions of a former iPhone developer (my most recent and least serious foray into the software business), there are schmucks out there who will even copy a DaysTo Christmas app, which is not exactly the top rung of app development quality.
All of this brings us to the software-as-a-service (SaaS) model of software product marketing. I'm not going to get into the big-scale variant of this market, like Salesforce or Office 365. Instead, I'm going to look at SaaS as another business model for producing the equivalent of packaged software products.
The idea of freemium software-as-a-service is that you get to use a product for free, but if you want to take advantage of additional features, you have to purchase the service. Next, we'll take an in-depth look at the benefits of the freemium business model.
Stay tuned for the next article: why the freemium cloud model may be your best choice.
By the way, I'm doing more updates on Twitter and Facebook than ever before. Be sure to follow me on Twitter at @DavidGewirtz and on Facebook at Facebook.com/DavidGewirtz.