Goodsol Solitaire Blog

Retail is Crazy

Note: This was originally posted on the old site on January 13, 2004.

During the holiday season when I was in stores I noted a trend that has been going on for quite awhile now, but seems to really be getting out of hand. Among games of the type that I make (casual games), there has been a big trend towards excessive bundling. In fact, it has gotten to the point where if you go to a place like Best Buy, about the only casual games you see are all in huge collections (1000 This, or 100 That). You cannot even go to a retail store and buy a computer card game or something like that - everything is now bundled into large collections.

This has terrible implications for developers. First, with all the games thrown together into one package for $9.99 or $14.99 or whatever, the royalties are split amongst large numbers of developers. This means that for any given product, each developer is getting only a tiny amount of money from each unit sold. With some of these collections that is literally becoming pennies or even a fraction of a penny per copy sold going to the developer of each game. There is no possible way that these collections can sell enough copies to make the royalty to the developer more than just a token payment. Second, the prices of these collections keep going down and the number of games in them keeps going up, driving the royalties downward in a death spiral.

It seems to be all about the number on the box. You have to have a huge number on the box, to say that this product has 101 games or some other ridiculous figure. So there is a drive to get larger and larger numbers on the box to outdo the box next to it. The actually quality of the games in the collection is of course of absolutely no consequence. It does not matter. The only thing that matters is to bundle huge numbers of games together to make it look like the best deal. Whether it is the publishers driving this, or big retailers like Wal-Mart, or a combination of both, I do not know.

Basically, the publishers and retail stores selling these collections are turning their games into commodities. The commoditization of the retail games market is driving prices and quality down. No developer with a quality product wants to have it sold bundled with 100 other games. Anyone with a quality game is going to get their games out of there.

And of course the customer notices. The first time someone buys one of these huge collections, he or she might think they are getting a great deal. But when they get home and discover that out of those 100 games, maybe one or two, if they are really lucky, is actually any good at all, then they will be thinking differently. The next time they go to the store, they will see the next 100 games for $9.99 package and think, "the reason it is that cheap is because it is filled with garbage", and they will be exactly right.

The whole situation is bad for both developers and customers. Developers are not making any money from these collections, and customers are not getting the games they want from these collections. If a customer goes to a store looking for a Hearts or Spades game, for example, they should be able to buy a Hearts and Spades game and not some huge collection of card and casino games, filled with games they do not want and are garbage anyway.

The only way this situation is going to change is if developers refuse to let their games become commoditized. Once every customer has purchased one of these collections and vows not to do it again, then there will be a market for quality games again.

---

On a somewhat related topic, Game Matters has an interesting post on what developers should ask for in retail contracts. Unfortunately, if you have never made a game like Duke Nukem you are not going to be able to get virtually any of these in a real contract. Only the Scott Millers of the world can get these kinds of terms.

Removing Customer Risk

Note: This post originally appeared on the old site on October 27, 2003.  Note that all links are over 2 years old.

Here are two different topics that are nonetheless very related.

Dexterity Software is running an experiment to determine which brings more sales: providing a demo (shareware version) of a game with an option to buy, or only providing buy links on a site without a trial.

On his forum, Steve Pavlina explained the experiment (link now broken), and then later, he detailed some preliminary results (link now broken).

The preliminary results so far indicate that not providing a trial increased sales by 31%. The experiment is continuing as not enough time has elapsed to make any definitive judgements.

The experiment was summarized at DIY Games and this got Slashdotted, with the usual Slashdot results. Most of the Slashdot comments showed that the commenter didn't actually read any of the details of the experiment, making all their comments worthless. But then again, when has a Slashdot comment ever been useful?

On another front, Carnival of the Capitalists #3 is up, and it has a link to an interesting article about the music industry's problems called The New Economics of Music: File-Sharing and Double Moral Hazard.

The key point of this article is that File-sharing is as much about risk-sharing as it is about the theft of value.

Fundamentally, I'm going to argue that consumers download music, as much to derive extra value from getting something for free, as they do because they want insurance against buying something they didn't want in the first place.

In effect, what he is arguing is that file sharing is used by music customers as a form of risk aversion, a way of providing themselves with a money back guarantee in an industry that provides none. Or alternatively, it is way of giving themselves a free trial of the product, to determine whether they like it or not. Try before you buy. In a word, shareware.

In part 2 of the article, the author offers some possible solutions to the music industry problem.  Emphasis is mine.

Another way is to offer listeners a contingent contract. Contingent contracts are where payment is dependent on some property of a good, like quality. You sell a contingent contract every time you order from Domino’s: if it’s not there in 30 minutes, your pizza’s free. The points is that these contracts offer another form of insurance, by matching quality to price – and so create the incentive for agents that are also good for principals. Because they make up for quality slip-ups, contingent contracts help sell goods when quality is uncertain, by reducing risk. It’s difficult to see how this can apply to information goods like music – since the price is paid before the quality is discovered.

 

He correctly states the problem - the price is paid before the quality is discovered.

But then misses the simple solution - discover the quality first, then pay the price.

Shareware.

It's a simple solution, but it does require one thing - the seller has to give up control to the customer. Which, clearly, the music companies are completely unwilling to do.

And now I come back around to my problem with Steve's experiment. I think it's great that he's testing this, because it will yield useful data. But I'm not going to do it myself.

The reason is that it would completely change my business model. My business model is centered around providing a trial version. It is completely dependent on the product selling itself by getting it into the customer's hands.

I'm certainly all for trying to get the sale without a trial download, if the customer chooses to do it that way. I've redesigned my site several times with the object of getting the immediate sale in mind. It's fabulous to get the sale before even a download. A money back guarantee is vitally necessary for this.

But giving the product away for a trial remains the ultimate customer risk insurance policy. Nothing does it better. Nothing else could convince the customer that it is a good product as well.

And it works. There is simply no possible way I could have become successful or have made as much money as I have without it. No possible way. I'll continue to dance with the one that brought me.

The way to have the best of both worlds, it seems to me, is to go for the sale on the web site, as the primary desired response from the visitor. And then provide the trial download as the backup response. That's how I have my site currently designed, although it could be made clearer.  And then go for the sale within the trial version. 

Shareware is the clearest method of removing customer risk.

Money Back Guarantee

Note: This post originally appeared on the old site on October 15, 2003.

The topic of a money back guarantee has reared its ugly head again on the Dexterity Forums (note: link now broken, now the IndieGamer forums).

My view is that if you do orders by credit cards, you are already giving a money back guarantee anyway. With credit cards on the internet, anyone can do a chargeback if they are dissatisfied. So you might as well do a refund anyway.

But on top of that, you need to do anything you can to reduce the customers perceived risk. There are still people that are very hesitant about ordering over the internet with credit cards. Even though this fear is wrong, it is still very real and you have to deal with it. One of the ways to deal with it is to reduce the risk by letting the customer know they can get their money back.

On the subject of reducing risk, I recently added some additional text to the sales pitch on my main page:

You are protected if you lose your copy. If you get a new computer or your computer crashes, just download another copy and use your registration code.

A lot of people think they have to buy again to get another copy if the lose theirs, and hopefully this will help reduce that perceived risk. I can't tell that it has had any affect on conversion rates, but it's worth putting in.

Code Requests

There's a recent thread in alt.comp.shareware called Charging to replace lost registration codes.

I have to wonder about the attitudes of those who do this. Is being contacted by your customers so repellent that you would go to such lengths to avoid it? This certainly would make sure that they wouldn't bother you, and in fact, would probably take care of the "problem" of having customers in the first place. I'm sure with a policy like this you wouldn't be troubled with the excessive burden of having too many customers.

Yes, if you use a registration code system you get people who lose their codes. It happens. Resending their codes is simply customer service. We've created web form for it.  Yes, it is time consuming to look up and resend codes.

But it's a valuable customer contact opportunity. The customer is coming to you and it gives you a chance to remind them of your products. If you have multiple products and the customer appreciates the service of your resending their code, perhaps they'll buy another product later. It's just stupid to charge for this. It won't bring in revenue, it will just discourage customers from contacting you (and, if people know about it before buying, will probably reduce sales).

I guess I'm just wondering what it is exactly that that person is doing all day that is more important than responding to the customer's problem.

The Problem

Note: This was originally posted on the old site on September 22, 2003.  Very little has actually changed on this in 2 years.

My company has a problem. It's a very nice problem to have, but it's still a problem. The company is profitable and is generating more and more positive cash flow.

The problem is how best to spend some of this money so as to expand the company (and expand cash flow even more, thus perpetuating the problem).

Over the past couple of years I have come up with some solutions to this problem, but since they are now starting to be successful, I need to come up with more.

Here are some solutions I've been considering:

1) Increasing spending on marketing of our current products.
I've been doing this one, with mixed success. It's difficult to find places to spend marketing money effectively. I'm just about maxed out on pay per click search engine spending (the only way to spend significantly more is to just raise my bids - but the ROI at this point on that is minimal). It's very tough to find web sites to advertise on that bring in quality visitors. I tried spending a bunch at download.com with absolutely disasterous results. Tucows barely generates a positive return. Web sites that actually have visitors that download and buy are very hard to find.

2) Commission development of some new products.
This has been the main thing I've done the last couple of years (see Pretty Good MahJongg and Action Solitaire) and it has worked very well. I'm definitely going to do more of this. The main problems here is that good developers are hard to find, but even more importantly good product ideas are hard to find. It's easy to come up with ideas for possible products, where it gets tough is finding the ideas that fit into our current product line. It would be easy to just develop a bunch of games of all types, but I don't want to do that. I want to stay focused and provide the kinds of games that my current customer base wants. Although an individual game can expand on that base in a certain direction, as both of the new products were intended to do. Finding the next good product idea is the real constraint here.

As a related solution, in addition to developing completely new products there is also the possibility of expanding existing products to new platforms or formats. I've been looking at the possibility of Mac versions for awhile. Other possibilities would be Windows CE and Palm versions. The problem here is that I would need to find a programmer with experience in these platforms.

3) Acquiring existing products from other developers.
This is something I have not yet done. The problem here is again finding the right products that would fit into my lineup. In this case there is then the additional problem of finding products that are of good quality available at a reasonable price. I suspect one could buy a lot of games pretty cheap, but they wouldn't be very good. The people with really good quality games probably aren't selling.

What makes Bejeweled popular?

Note: This was originally posted on the old site on September 20, 2003.

On the Dexterity Forums (now the IndieGamer forums) yesterday, someone asked what makes the game Bejeweled (Diamond Mine) from PopCap so popular.  It's a good question,  I answered it on the forum (note: link is now defunct)  but I'm going to quote my answer here.

Bejeweled is popular because it is simply a classic game. The game itself (never mind the computer implementation of it) is brilliantly structured.

I asked my wife (the casual gaming queen) why Bejeweled is a good game and she said it was because it gives you lots of small achievements. Each time you create a sequence you get an achievement and a small satisfaction. These build up as you play.

The game is very easy to learn, the concept is extremely simple. You can learn the rules in a minute. When playing you can go into a mindless state if you like, but the game is *not* mindless as there is room to use skill to add to your score. You can attempt to improve your score by looking ahead and trying to make moves that create multiple sequences of gems. It's up to the player how much thought he wants to put into it. The more effort, the higher the reward. But it can be played to pass the time for its own sake. The game itself (again ignoring the computer implementation of it) is brilliantly designed and my hats off to whoever designed it.

Good graphics, good controls, sound, music, and many of the other things previously mentioned here are of absolutely no importance as to why the game is good. Bejeweled would have been a massive hit if it had been implemented in 1980 on a TRS-80 with a black and white screen and character graphics. It is the properties of the game itself that makes it good.

That said, PopCap's implementation of the game is excellent. The graphics, sound, and user interface *are* good. But what makes it really good is how the game aids the player. It's the little touches that separate the true professionals from the amateurs. The way that the game shows you a hint if it has been a long time since you have moved is truly masterful. It's subdued but noticable, elegant but unobtrusive.

Many game developers and hard core gamers, it seems, simply don't appreciate the elements of design that go into games like this, which is why they are mystified by why these games are popular. Complexity is the enemy of good game design, unless that complexity arises naturally out of small elements of simplicity. High end graphics can be the enemy of simple elegance, which will beat the high end graphics every time. A fun game gives the player simple objectives which result in simple rewards, and it game allows the player to get back more if they put in more.

People will be playing Bejeweled (or games like it) 20 years from now, just like they are playing Tetris nearly 20 years later, because it is such a well designed game.

Hug Your Customers

Note: This was originally posted on the old site on June 30, 2003.

I just read a great book over the weekend.  The book is called Hug Your Customers by Jack Mitchell. Mitchell is the CEO of a high end clothing retailer in Connecticut. They have only two stores, but they do a stunning amount of business out of those stores (65 million a year). The reason they sell so much is not because of the clothes (which are all brands you buy elsewhere, and probably at a lower price) but their service. They go to extraordinary lengths to help their customers. In this book Mitchell lays out all the things his company does to make customers keep coming back.

Clothing retailing is a face to face business, so much of what Mitchell does relies on being actually in the physical presence of the customer. Nonetheless, there is a lot here that can be applied to an online business where you never even see the customer. Here are some of the things Mitchell discusses in his book.

The number one principle of his company is that everyone in the company comes in contact with the customers. Everyone sells on the floor at least sometime. Even the CEO. This is really important and I think an online business should be the same way. Everybody in the company should answer emails from customers. Certainly it is that way here (of course there are only two of us). It's only by that constant contact with the customer that you know what customers are thinking and what they want. Without that daily contact with the customer, it's easy to lose sight of what the company is about.

As a related point, Mitchell believes that you should know who your top customers are. His company knows exactly who spends the most money in their stores, and he knows their names and even knows a lot about them (I would kill for Mitchell's customer software system where he tracks all this stuff, I really want software like that). He knows who his top 100 customers are. Now I can't really say I know who my top 100 customers are by sales volume, because we only have a very few products, so once someone has bought them all, that is it (which is, of course, one of the problems with my business - too few products). But I do know a large group of what I'd call my "most active" customers - that is, the customers who email me often, show up on the top scores list or post on the discussion forum regularly. I don't know if there are a hundred of them that I could name, but there are a lot of them in any case. Although these people represent a small part of my sales, they represent a large part of the community of players of my games. They are the core group that I think about when I'm thinking about what kinds of new games would appeal to my customers.

Mitchell keeps track of every purchase every customer makes. The salesperson can call that up when you come into the store and determine what you might need to buy. It's amazing. Have I mentioned that I really want his computer system? He's got a database of information about over 115,000 customers. That 65 million divided by 115,000 gives you 565 dollars revenue per customer per year. Yes, there are definitely advantages to a business where you have a high average sale. I get only $25 per sale ($50 in the cases when a customer buys both major games at once). And this information is not limited to just what the customers buy. They try to find out everything about a person (by just asking in conversation). If they find out the customers dog's name, they put that down in the database too.

All this just makes me think about my dream customer software package. This software would integrate our customer orders database and our email databases. Not only could you see all the orders each customer has made, but you could see all their emails as well. It would put them together in the same database. And if someone emailed us before they ordered, it would be able to go back and pull those emails too and put them in with the order. That would be just fantastic, you could see every contact with the customer, orders and emails combined. This kind of database would just do wonders for us, since you could see the entire history of our relationship with the customer all in one place.

Mitchell also emphasizes knowing the score. By that he means that you need to know how much sales you are getting today, how you are doing for the month and year so far, and how this compares to previous years. This is something I started doing very early on, and I'm fanatic about it. I like to know how we are doing all the time. I like to know how many sales we've had in the last 24 hours, so I know how the business is doing. And I like to know it many, many, many times per day. One of the great things about RegNow, our order processor, is that since at least 1998 they've had a running 24 hours sales total on the main sales report page. So I can call up our orders anytime and see what we've sold in the last 24 hours (at least via RegNow, which is a large portion of our sales). The funny thing is that I get the feeling that this is not a big thing for other people, because several times over the years RegNow has tried to take this off their order report. One time recently they did it on a Friday evening and I was without my sales total for a whole weekend. It just drove me crazy all weekend long. I didn't know how we were doing. I yelled and screamed loud enough that they got it back up there quickly. But to me, this is one of the best things about their system, and unbelievably I've never seen another payment processor in the shareware business that does this. You gotta know the score.

Anyway, this book has given me food for thought for a long time.  The book has a web site at hugyourcustomers.com.

What is Shareware?

Note: This was originally posted on the old site on June 4, 2003.

Most of the people reading this blog right now are other shareware authors, because that's who I've told about it. But a few of my customers are trickling in from links on my sites, and this will only continue. It's possible that there will be people coming in here who don't know what shareware is.

So I think it would be a good idea to define what shareware is, seeing as how that is what this site is about.

The definition of shareware is actually a matter of some debate in shareware circles.  But it is generally agreed that shareware is try before you buy software.  It is a way of marketing software.

It is different from the traditional way of marketing and selling software. In the traditional way, the customer buys the software in a store. They take it home, open the box, install it, and then they pray that it works. This could be described as the buy and pray method of purchasing software.

In the shareware way, the customer downloads a trial version of the software. He/she can use the software free for some trial period (usually 30 days) to see whether it fits their needs. If it does, then they buy it. Shareware could be described as the play and then pay method of purchasing software.

The benefits to the customer are obvious. When you buy software in a retail store, all you have to go on is the pretty box. The pretty box tells you how great the software is. The pretty box tells you what the software does. The pretty box tells you how much you need the software. The pretty box tells you that you should take it home now!

So you buy the pretty box, take it home, install the software, and throw away the pretty box. And, as often as not, the pretty box lied to you. The software isn't that great. Or maybe it is great, but it doesn't do what you actually want.

Shareware allows you to test the software out before you buy it. You'll know if the software does what you actually want. You can try out all the different products of a certain type and find which one you like best, then buy it.

This demo approach now dominates the software industry so that even big companies like Microsoft have demo versions of their software, even if they don't market in the shareware way generally.

And there are other benefits to the shareware customer. Most shareware products offer free upgrades for at least some period of time after the original purchase (I offer free upgrades for 2-3 years). Try walking into a store and saying "Hey, I bought that last year, I see you have a new version, can I take that home with me for free?"

With most shareware products, you can get at least email technical support from the developer. Often it's directly from the programmer himself, as in my case. Again, with most retail software, you can't get any kind of support anymore.

All this makes shareware very attractive to the customer.

What is less generally known are the benefits of using the shareware method to the developer of the software. A lot of people in the software industry think that the shareware method has been a failure. Some examples - Shareware Will Rise Again (as if it fell), Who Will Pay for Software.  It's as if shareware doesn't exist to these people.

But what is known inside the shareware industry, and especially to those of us in the shareware game arena, is how much more lucrative shareware is compared to the traditional retail method for the actual developer of the software. When someone buys a piece of software in a store, how much of the purchase price do you suppose goes to the programmer who wrote it? If he's lucky, a $20 software sale will make him $1, maybe with some change. More likely, it's under a dollar. With the tendency towards bundling in retail games these days, it can be literally pennies.

When a shareware sale is made, the developer usually only pays a small processing fee for the credit card or order processing company. This fee usually ranges between 4-20%. On a $20 sale, the developer usually gets at least $16.

That makes a huge difference. But the traditional belief is that because you can sell so many more copies at retail than you sell at shareware, the developer still makes a lot more money at retail.

But I'm here to tell you that is just plain wrong. 

I have products that are sold both retail and shareware. Last fall, the retail version was one of the best selling retail card game products in the USA. Yet even with that huge success, the shareware version outsold it in actual revenue to the developer by more than 10 to 1.

There's just no doubt about it, if a developer can learn any kind of marketing knowledge at all, he can make more money selling via shareware than going through a publisher and selling retail. The reason so few people do it is that so many programmers have a fear of marketing. They don't know anything about it and are scared to try going out on their own. But it's not that hard. Joining the Association of Shareware Professionals is a great first step.

It just makes so much sense for both the customer and the developer that eventually it's bound to take over the software industry in one form or another.

How Long Does Great Shareware Take?

Note: This was originally posted on the old site on June 9, 2003.  Everything mentioned is now 2 years older.  Pretty Good Solitaire is now 10 years old.

There's a great article by Joel Spolsky called Great Software Takes Ten Years.  In it, he claims that all great software takes 10 years to create.  He gives compelling examples.

So how long does great shareware take?

Shareware is software, so the first answer is 10 years. But shareware is generally smaller and less complex software than what Spolsky gives as examples. It's reasonable to believe that great shareware would take somewhat less time. Anyway, there are very few shareware applications that are 10 years old (although anything that makes it 10 years in this business must be great).

Winzip recently celebrated its 10th anniversary.  Paint Shop Pro must be close.  Clipmate is an amazing 12 years old.  My Pretty Good Solitaire will turn 8 next month.  If you have or know of a shareware application that is 10 years or older, mention it in the comments.

So shareware probably becomes great faster, maybe at 5 years, possibly even 3. But it certainly takes years. Many shareware newcomers just don't realize the amount of time it takes. As Spolsky says:

Make a ten year plan. Make sure you can survive for 10 years, because the software products that bring in a billion dollars a year all took that long. Don't get too hung up on your version 1 and don't think, for a minute, that you have any hope of reaching large markets with your first version. Good software, like wine, takes time.

Most shareware applications are not immediate successes.  It took a month to get my first  Pretty Good Solitaire registration.  It took years before I could live on it. Spolsky says:

When you release 1.0, you might want to actually keep it kind of quiet.

This is somewhat the approach I am taking with my new product Pretty Good MahJongg. It was released last October, so it is only 8 months old. Still an infant. It's only at version 1.1, so it is still only a fraction of what I have plans for it to be. As such, I don't necessarily want everybody to know about it. Some people might try it out and decide that it isn't good enough, and then never look at it again. But Pretty Good MahJongg is really an investment for 2005 and beyond, when it should really come into it's own. Of course, that's not to say it isn't a good game, which I think it is, it just is in the early stages of its development.

Spolsky again:

Anyway, getting good software over the course of 10 years assumes that for at least 8 of those years, you're getting good feedback from your customers, and good innovations from your competitors that you can copy, and good ideas from all the people that come to work for you because they believe that your version 1.0 is promising. You have to release early, incomplete versions -- but don't overhype them or advertise them on the Super Bowl, because they're just not that good, no matter how smart you are.

Version 1.0 serves just to get the application out there. In version 2.0, you start getting the features that didn't make 1.0 in, plus some of the early customer feedback. It's only in version 3.0 that it starts getting good. It's a common joke that Microsoft doesn't get their software right until version 3.0, but the reality is that no one else does either.

Then, when you've reached 10 years, you run into a new problem.  As Spolsky says:

So, it takes a long time to write a good program, but when it's done, it's done. Oh sure, you can crank out a new version every year or two, trying to get the upgrade revenues, but eventually people will ask: "why fix what ain't broken?"

You run into the problem of what new things can you put in.  This is the problem that I will face with Pretty Good Solitaire. While I still have some ideas left, at a certain point there will come a time when there won't be anything left to add. What happens then? This is, of course, why Pretty Good MahJongg, and maybe some other future games, exist. If you are a one product company, you will eventually get to the point where the product is done. You then need some other products to keep the growth going.

My approach (which wasn't exactly planned, but it worked out this way) was to concentrate on one major product for years, building it up. Then as the revenue from it increases, use the money to fund development for a second major product. Then when the first product tops out, there will be a second product coming up to fuel growth. Then, as the shampoo bottle says, repeat.

Great shareware takes time. 

How Many Sales Does a Shareware Game Get?

Note: This was originally posted on the old site on June 21, 2003.

Somebody got to this site last week by searching for "how many sales does a shareware game get". Also, another person searched for "how much does shareware games sell". Both of these people probably left without an answer. But in the interest of giving people what they want, I'll answer the question for future searchers.

The answer is 42. 

OK, seriously, the answer is somewhere between zero and several million, with the most likely answer being something close to the zero. So 42 is as reasonable an answer as any.

There are basically two myths about shareware game sales. The first myth is that all you have to do is copy a successful game and you'll make millions. That just doesn't happen. Copies don't sell well. I imagine that the vast majority of all the shareware Tetris clones out there have sold fewer than 42 copies each (and maybe all of them together have sold 42 copies). Best selling games always do something differently, they generally aren't clones.

The second myth is that nobody makes any money selling shareware games. This isn't true either. While the vast majority of shareware games sell only a few copies, there are quite a few successful shareware game developers.

What you need is some perspective on how many copies most games of any kind sell. Most retail games don't sell many copies either. As Brad Wardell says in his article On Being an Independent Game Developer, a successful PC game sells between 30,000 and 50,000 copies. That's retail. Hit games may sell 100,000, and it's only the really big hits like The Sims that go into the millions. Most games of all types don't sell many copies, so it isn't surprising that most shareware games don't either.

So a shareware game that sells a few thousand copies is a successful one. But there are some very successful ones. While Brad Wardell's article is a good one, he does make a false statement:

I’m not aware of any games that are available exclusively electronically that have sold more than 5,000 units.

Well, I guess it's actually a true statement, it's just that he must not be very aware of the shareware game market. I have several games that have sold much more than that, and I know of quite of few others who must be selling much more than that (else they wouldn't be making a living or affording to pay their employees). PopCap claims to have sold over a million copies. It's not impossible to sell this many, it's just highly unlikely.

Shareware games are a hit driven business. The hit games make almost all the sales and most games fail. However, many of the games that fail don't have to. While a lot of games fail because they are just bad games, a lot of games fail because their developers don't know enough about marketing. You have to realize that a good, well programmed game is only step 1. It gets you nowhere except into the field. A game needs to have good marketing as well.