How much does CDROM development cost? Now for the nitty-gritty. What is this little project really gonna cost? Different sources will give you different guidelines, but the process of creating a CDROM product involves the following parts, one way or another: Product design Rights acquisition Data conversion Software acquisition/indexing Premastering Mastering Packaging Marketing Recognize that there are no hard and fast rules for any of these processes. Even those portions where vendors have price lists are frequently negotiable. In those places where I can quote a range of prices, remember that your project may be much more or much less. For better or worse, CDROM development - as a total process - is not yet a commodity business, and business plans cannot be made without a detailed analysis of the particular project at hand. Product design Since this is the topic of most of my Developer's Corners, I am not going to explore this item in detail here. Suffice it to say that this is the stage where all of your cost factors are created and defined. Rights acquisition The most efficient scenario is the one in which you own all the data you wish to publish. If this is the case, then these costs have already been absorbed by other projects. If you must negotiate for the rights to all or part of the content you wish to publish, however, this item can be very significant. Keeping costs down is an exercise in creative negotiation. Content owners typically have a very high opinion of their intellectual property, and try to price it accordingly. Your challenge is to try to keep the up-front costs as low as possible. Licensing costs are generally of two types: up-front licenses and royalties. The up-front type can take at least two forms: an outright licensing fee, or an advance against expected royalties on the sale of your final product. In the print world, content owners seem to have calmed down to more realistic expectations from CDROM publishers than they had several years ago. While most will want something up-front, I find that more and more are willing to work with the CDROM producer to keep those costs down; in a sense, sharing some of the risk (except they are risking only future revenue, not out-of-pocket money). These arrangements are as unique as each deal, and I have heard of fees ranging from $500 to multiple tens of thousands of dollars. Unfortunately, there is less experience in licensing music and video, and the negotiations for well-known properties can be pretty wild. If the experience of early CDROM publishers holds in this arena as well, costs should stabilize as content owners gain more experience. Royalties are a way of letting the CDROM product pay its way as it sells, sharing some of your income with the content owner. This is an acceptable way to go if your balance sheet can stand sharing precious revenue off the top. The biggest danger in this area is in the possibility of total royalties exceeding your income, especially in cases where you want to license content from multiple owners. If you feel a 10% royalty is reasonable for one content owner, but you want to combine the content of 10 publishers in your product, you can't offer each a 10% royalty. Your return gets mighty thin when you are paying out 50-80% of net - or worse, gross - in royalties. This can be a tricky technical and legal contract area, so you might want to hook up with someone who has had some experience in these matters. The most basic rule is: Calculate the total royalty you can pay, then negotiate with each publisher on a portion of that amount, proportional to the amount of content each content owner provides. And there are a variety of arrangements which can be discussed outside the actual payment of dollars to the content owner. These involve such things as trading rights, trading converted content for rights, partnerships in development or marketing, and so on. Leave no possibility unexplored. Data conversion/compression The greatest potential sinkhole for your money in any electronic publishing project can be data conversion. This is one that can eat your lunch. This issue must play a role in your negotiating for rights as well. If the content for which you are dealing is not already in some usable electronic form, you are going to have to lay out hard dollars to get it in that condition. Therefore, the rights to such properties should have less value than the rights to data in pristine, marked-up digital files. Data conversion is the process of digitizing all the content required for the product. CDROM is a digital medium, right? So all text must be converted to ASCII, all still graphics to raster images, all sound to digital form. Video must be digitized frame by frame (DVI and CDI technologies excluded, sort of). No matter how you cut it, this is not cheap! - Print conversion: The original content is going to be in one of two forms: magnetic tapes or discs used for typesetting and/or composition for printing, or only on the printed page. Just because the typesetting/composition tapes are available is not reason for immediate rejoicing. If the format in which this information was captured is a) known and b) interpretable, it can be a real help in conversion. But many of the systems used to create print works are home- grown, arcane in the extreme, proprietary, inconsistent or a combination of all these disasters. Also, printers only care what ultimately gets shot for the printing plate, so last minute changes, corrections or additions may not make it to the master tape; they simply get pasted onto the mechanical. So you need to have a technician examine the nature of the electronic version of the information. It is not uncommon that the programming required to translate these tapes into useful form is so difficult that it is actually cheaper to simply re-key everything, as though the tapes didn't exist. If the content only exists on the printed page, as will frequently happen with out-of-print material, your choice is to scan the pages with optical character recognizers (OCR), or to literally re-type the entire work. Scanning is a technology to which there is less than meets the eye. OCR scanning may be appropriate for small jobs, but the problem is that scanners are simply unreliable for text, particularly text that includes technical information where there are many charts and tables, mixtures of letters and numbers, or special characters (like math symbols, Greek letters, etc.) Oh, I know, you are going to say "But XXX Corp. guarantees 99% accuracy for their OCR software." Ok, but consider that a computer screen can contain 2000 characters of text. 99% accuracy means that there might be "only" 20 typos per screen. Would you be impressed with such a presentation? But then, you say, all we have to do is clean it up, edit the scan. Ok, but even though most OCRs will flag unreadable characters, what do you do about those characters that the scanner misinterprets, that are not flagged. How about when it transcribes a "5" for an "S" or a "2" for a "7" and so on? Your editor must essentially re-read and verify the entire document, changing those errors that have been flagged, and searching for the errors that have not been flagged. This adds time and personnel cost. And few scanners really work at their rated speeds. A slow scanning system with large amounts of human overhead makes for a real problem with a large project. This leaves the option of re-typing, or re-keying. It is an inelegant, messy, ugly, low-tech solution to a high-tech problem. The only possible attribute that recommends it is that it works, and is often cost-effective. Specialty houses use a number of techniques to accomplish this task, the most significant of which is the utilization of low-cost offshore labor (we don't have to argue about political or social correctness here). The operators work at speeds in excess of 10,000 keystrokes per hour, and this is one of the few tasks where the number of people you throw at the problem truly impacts its completion time. Costs can range from as low as about $2.00 per page of original (depending on the number of characters per page - typeset is dense than typewritten), to $8-10 per page if you want SGML mark-up and other services. SGML stands for "Standard Generalized Markup Language," and is something you should be familiar with, if you are not now. It will be the subject of a future Developers Corner. A couple of cautions. Remember that a space is a character. Remember that character density varies from original to original. You will pay more for more service. Prices vary according to - among other things - the guaranteed accuracy rate (99.8% is not uncommon as a base, 100% accuracy costs more), and the speed with which you want the job completed. - Graphics: These are typically scanned easily. You will have to decide the resolution with which you want them scanned, and whether they are to be line-art, gray-scale or color images. A service bureau cost of $1-1.50 per image is not unusual. This is one area where it might pay to acquire a scanner and perform the task in-house, where you can more closely control the costs and the quality of the output. - Sound: Digital sound comes in a wide range of qualities, from speech grade (like the quality of telephone transmission), to compact disc audio grade (also known as Red Book audio). The grade you select also impacts the amount of sound your product can deliver, as in the amount of space it requires. One product I know contains something like 140 hours of speech-grade audio over the full text of a dictionary. At compact disc audio grade, the maximum amount of sound per disc is about 72 minutes. And there is a range in the middle of these two options. There are a number of in-house alternatives for producing lower grades, but you may need to hire a specialized studio for higher grades, at rates from $50-$150 per hour of studio time. - Video: This is, today, the single most difficult and potentially costly aspect of data conversion. Compression is always a factor in video conversion, because full-color, raster tv images are so incredibly data/space intensive. There are two families of video conversion/compression. The first captures each frame, converts and compresses it, for playback in the author-defined sequence. The second is called the delta technique (as in the Greek delta which denotes "change" in engineering and mathematical disciplines). This approach captures the entire first frame of a video sequence, then captures only those pixels which change from frame to frame throughout the rest of the sequence. The computing resources required to deal with this issue, so that the playback (access, decompression and display) is close to realistic, are substantial. Apple has made, and will continue to make, giant strides in this area. DVI and CDI have been working for years to perfect their approaches. No technique yet offers an acceptable technique for full- screen, tv-quality presentation. (And NTSC tv quality is a fairly low- quality compromise anyway.) The result, in DVI for example, of a compression/decompression task for full 30 frame-per-second (fps) video on about 30-40% of the screen, has been compared to the picture quality of a VHS vcr. Not top-drawer, but certainly acceptable. Other techniques, such as those used in CDROM-XA or CDTV development, can do a respectable job with 15 fps playback. Fifteen frames per second would not be satisfying for an Indiana Jones movie, but it can be perfectly acceptable for a technical presentation or a game. It is your decision, as the product designer, as to what is necessary and acceptable. Costs can be high. Full compression using DVI can cost thousands to convert a 20-30 second sequence. Using other workstation-based techniques might cost only time if you already have access to the necessary gear. - Compression: Even though the capacity of a CDROM seems vast at first glance, 600-650 megabytes starts to seem small in terms of large data collections and multimedia. And of course compression of data is only one aspect of your challenge in publishing: you must also be sure that your customer can de-compress the information at the time it is needed. Much compression today can be done in software, avoiding the need for your customer to acquire yet another piece of hardware to use your product. Even some video can be decom-pressed without hardware enhancement (CDTV and some straight CDROM techniques, for example). As your plans evolve, be sure to consider this impact on your customer base in selecting conversion/compression methodologies. As you can see, data conversion factors can weigh heavily in your product design and rights acquisition decisions. If you have access to electronic data or substantial labor pools which can perform text conversion, you can drastically lower these costs. Also if you have access to a fairly well- equipped computer lab you can lower your costs on graphic, sound and video conversion. These are important resources to figure into your product development costs. Software acquisition/indexing Software controls a CDROM production. It defines the way your customer can use the product and provides access to everything on the disc. As with any other computer-based development project, you have two alternatives for obtaining software to run the application/production: make it or buy it. If your CDROM project is a compilation of stand-alone material - such as clip art or public domain software - then you can easily use the standard facilities of the operating system as an access method. Appropriately labeled directories or folders can be as useful as an indexing approach. But if you are working with a database product, or a multimedia production, you must choose application software with which the product will be used. Making your own system might be a good option, particularly if you are technically-capable or have a staff that is. Sometimes, there is simply no commercial, off-the-shelf alternative which meets your needs. In the in- house environment - discs created to meet specific corporate goals for documentation, training, etc. - it appears that a substantial proportion of producers elect to use their own in-house programming resources to create the indexing, retrieval and presentation tools required. Buying - or rather licensing - a commercial package requires that you have just as good a handle on what you want the product to do as if you are going to build your own. Commercial CDROM authoring/retrieval systems cover the gamut of functionality and pricing. Some vendors will work with you to modify their code to meet a specific need, others cannot/will not. Some packages are highly flexible, others are extremely rigid. At last count, there were over 100 software vendors offering CDROM applications. Of these, perhaps two dozen dominate the market. Key points to consider in selecting a package: - Can the package handle the volume and type of your data? - What hardware system is required, and how easy or difficult is it to use in processing your data (indexing, for example) for production? - If your project is multimedia, can the package handle the data types you must include? - How useful is the interface, from your customers' perspective? Can the interface be modified to meet their specific needs? - Does the package perform - with your data - at a speed that will meet your customers' needs? - Can the package support networked access, an increasingly important issue, especially for CDROM products which carry a relatively high price tag? - Can the authoring portion of the package support pre-mastering, simulation and performance evaluation etc.? - Does the vendor have sufficient experience and resources to support you and your customers? Have they done this before? Do they have the financial stability to be there years from now? In short you must consider how effective the package will be for you to create the product in the first place AND how effective it will be for your customers to use. Pricing for software varies to the same degree and in much the same way as content licensing. You will certainly face some up-front cost - high or low - to acquire the system to index and produce your own disc. Then comes the question of royalties. Some vendors require them, some don't. There has evolved a fairly straight-line price/functionality curve in CDROM software. There are packages which can be acquired for only hundreds of dollars and require no royalties to be paid. But these systems are generally fairly rigid in design, are usually adapted to only one delivery platform, and often cannot handle multiple data types. On the other end of the scale, there are packages which can cost $25,000- $50,000 for the license, and ask $5-$75 per disc in royalty. You can expect that these systems are designed to handle a wide variety of needs, and to be compatible with several platforms. The middle ground in this range is a mixed bag. Do not discount the time required to assess mutliple software systems. The software is the window through which your customers must look to use your product. Premastering This activity is basically the process of taking your painstakingly produced product and reducing it to an image that will appear on the final CDROM. After all the data has been converted, compressed and indexed, premastering places all the elements in sequence, with appropriate blocking and other technical attributes required for the product to work. There are special pre-mastering systems, and some of the software packages mentioned above accomplish this task without having to go to a special system. This is also the stage at which you will probably want to simulate the performance of your product on hard disk before sending it off to be mastered. This is so you can do any tweaking of the product that you feel will improve its speed of access, appearance or other features. The result of premastering is a tape, a hard disk, a write-once optical disc or some other medium that the mastering facility can accept to create your "press run." The advent of write-once CDROM is a boon to publishers because it enables you to publish a one-off copy for testing, and if it is satisfactory, send it directly to the mastering house to mass produce copies. Premastering systems, if the function is not supported by your authoring/retrieval vendor, can cost anywhere from a couple of thousand dollars to $15,000-20,000. If you plan to be producing lots of disc products (which you should), this is not a superfluous investment. It can save you much more in mis-mastered discs in the long run. For a first effort, however, any number of vendors - software houses, service bureaus or the mastering facility itself - will contract for this service. Mastering CDROMs are pressed, just like the old LPs. A master is produced from which replicates are made. Much to the dismay of manufacturers, this end of the business has indeed become a commodity business. This is the least expensive part of the entire development process. Charges are based on three levels, generally. The master itself (which involves reading your input medium, translating to the master material, QC and so on) may cost $1500-2500, sometimes less. Often this price includes the pressing of a minimum number of replicates, usually 100 copies. Additional replicates generally cost $2.-3.00 each, depending on quantity. Finally, you probably don't want to distribute an unadorned disc, so the mastering facility will help in designing whatever logos or wording you wish printed on the top of the disc. Some mastering houses, like METATEC/Discovery Systems, have built great reputations on their ability to do high-quality screen printing on the disc, creating a truly colorful and unique look for your product. As always, this costs extra, but the costs are usually relatively small in comparison to the impact these finishing touches have on the marketability of the product. Packaging Depending on the way your product is to be marketed, you will have to consider some sort of packaging for the product. Some products must be accompanied with a full-sized manual that documents the use of the retrieval software. Those products with simpler interfaces can be delivered with less paper, in packages not unlike a videotape, or a bubble- pak. If your product is going to be sold by direct mail, or by a direct sales force, and basically on the value of the data (such as to a professional market) you will want to package it attractively, but the packaging itself will not likely make or break the sale. If you have a more retail market in mind, then you must consider the attention-getting properties of the package, and its ability to sell itself. Most mastering facilities - and some of the full-service service bureaus - will be able to help you acquire the packaging you require. Here, as in any manufacturing intensive process, volume dictates a large part of the price. If you want a 200-page documentation binder and 2-color printing for only 100 copies of your product, expect to pay upwards of $20-30 apiece (not to mention the cost of the writer - we don't come cheap!). On the other hand, if you can get by with a jewel box or flexible plastic sleeve and a colorful insert for multiple hundreds of copies, you might pay only $3-5 apiece. Some of the vendors who provide packaging for music discs have outrageous capabilities, with seemingly unlimited variations. These costs vary greatly, but I have seen outstanding samples priced at $3-6 in minimum quantities of 2000 or so. Marketing Somehow, in any list of tasks for product development, marketing always seems to come last. But in fact it is as important as considering all the other costs of development combined. You will not sell any copies of your product if no one knows it exists. You have to get the word out somehow. And the techniques for doing this successfully are as varied as the products and markets themselves. You might do yourself as much good with a couple of well-placed press releases as with a multi-thousand dollar advertising campaign. Then again, you might be in a market where no one will take you seriously unless they've seen a few full-color spreads in popular magazines. Unless you are a marketing wizard, you will probably want to enlist the aid of someone who can offer advice in this area. The variety of potential approaches to this problem means that costs are very difficult to calculate. You must think about them early on and plan for them, because all your hard work in creating the product could go for naught if you can't sell it. The only rule of thumb I have ever heard was from Mark Foster of Quanta Press a couple of years ago, when he said"You should plan on spending as much on marketing as you spend on data conversion." I have no idea what the causal relationship might be between data conversion and marketing. But his point was that data conversion is likely to be the single biggest cost in development, and you should be willing to commit to spending at least that much to get the product in front of customers. Because CDROM is a new way of communicating, I believe that it opens up new opportunities for marketing and for creating strategic partnerships in marketing, outside of and in addition to traditional sales and distribution channels. This subject has been part of past Developer's Corners, and will certainly be discussed in the future. SUMMARY I would hope you can see by now that content licensing, data conversion and software selection are the three biggest expenditures you are likely to face in the actual development of your product. Data conversion is a fixed cost, once you figure out what it is. But royalties for content and software will impact the return on each unit sold. So making decisions on these issues, and negotiating royalties bears a direct relationship on what you have to charge for each copy of your product, regardless of volume. Let's say you want to build a $50 CDROM, and have to pay out 15% to your content provider, plus $5 to the software vendor. In this case you will net $38.50 in real dollars per disc to pay back conversion and development costs, as well as on-going cost of sales, salaries and profit. In short, you will have to sell a whale of a lot of discs. (Further, if you work with a distributor, you must typically offer them a 40-50% just to handle your product. Selling your $50 disc through a distributor with a 40% discount nets you a gross of $30, and an after-royalty net - in this scenario - of $20.50.) You can see why so many CDROM products seem to be so high-priced, and why there are so few low-priced multimedia CDROM products on the market. If you are not now a publisher, but own some content or can make a good deal, and want to just put a product out there to see what happens, I recommend you do it. But don't expect to be entertaining Bill Gates and John Sculley on your private island with the income from your first product. If you are currently a publisher, many of these costs will be absorbed through the normal costs of doing business, or at least your business could be structured in this way. Therefore, after the first product, future CDROM releases could be incremental in cost, and represent "found money" on your balance sheet. The key to success in CDROM is that these expenses should not be undertaken with an eye to producing simply one product. The myth of the "Killer CDROM application" should be considered just that, a myth. There may be some home runs in CDROM just as there are in any other media. But the vast majority of products contribute cumulatively to the revenue stream, where initial costs are absorbed across the entire line, not that first single product. One attribute shared by CDROM publishers and publishers in any other medium is volume - volume of sales per title times the volume of titles produced. There are few if any successful one-title book publishers or movie producers. And it is not likely we will see any blockbuster one-title CDROM publishers. So do it! Apply lessons from your first experience to your second and third and fourth. The magic of CDROM is in what it can do. But the business CDROM takes as much commitment as any other.