Will Today's Models Take Industry Into the Future?
But there are caveats. Modelers typically need to get reports from their fulfillment companies and plug the data into the right place in the model. Because of ambiguities in the data, this can be confusing and complex, requiring much time, knowledge, insight and aptitude. The modeler must identify patterns, make tallies and input the data.
Of course, the predictions will be accurate only if the assumptions are valid ... and there is room for error. The process can be daunting for small to midsize magazines with limited staff.
I envision more intuitive models that correct these shortfalls and take modeling to a new level. But more on that later. First, a rundown of what's available and how these firms are looking to the future:
Lighthouse model. Considered a good, basic model and great little workhorse by its developer, Greg Jones, Lighthouse remains applicable today because it is a straightforward spreadsheet model.
Its pluses are also its minuses. The model has flexibility, and users can tweak it to address quirks in their business. But tweaking can create bugs and get publishers into trouble. Ideally, it should be used without tweaking to take advantage of its basic horsepower. The model also is limited in its ability to handle odd or changing frequencies such as with a skiing magazine, which has a gap in the summer months.
When the model was sold to Kable Fulfillment Services, there were 600 licensees. Jones guesses there are 200-400 active users now. He thinks it may remain a strong alternative in the future for publishers who want a good, solid model, especially for launches.
Granite Bay. Originally Blue Dolphin, Granite Bay is a spreadsheet model developed by Don Nicholas in the 1990s that is much more complex and flexible than Lighthouse. Unlike "black box" models, Jones says, users can see the calculations, follow the logic and flow right in front of them and trace causal factors to better understand the economics of their business.
But the complexity makes it harder for the average user to work with, and it takes time to load and print results. Currently, 25 magazines use Granite Bay, including Harvard Business Review. Annual licensing costs range from $4,000 to $5,000 a title, with discounts available for multi-title publishers.
Webber. The Magazine Publishing Model, as per David Webber, is the only true weekly model. Any serious weekly uses it, including Newsweek, TV Guide, Business Week, Primedia and Rodale, from 600 to 700 titles. Webber considers the database model user-friendly as it looks like Excel and runs on any PC with any version of Windows. It's accurate and fast, delivering answers in seconds even though there might be more than 100 subscription sources. The model accepts up to 999.
Webber said his biggest competitor is the CDS model, which runs on DOS, while Kable and Palm Coast Data use the Webber model and sell it as their own. Webber said his is the only modeling company that does its own programming. He has worked on the model for 34 years, fixing bugs within an hour and changing the model to respond to industry changes. He updated the model to accept advance renewals four years ago.
More costly than many competitors, the Webber model can be licensed for $460 to several thousand a month with site licenses costing $17,000 to $80,000. Webber has written a new version in C++ but said his clients don't want to test it because they are happy with the current version.
CDS. This model comes from fulfillment giant Communications Data Services. Now under revision, the CDS model introduced a new reporting system last year, the Mars model. A new input system called Quasar is in development, while the calculation aspect of the model will be reconfigured in the future.
While Mars and Quasar are Windows-based and can be accessed in Windows XP, the CDS model still uses a DOS operating system. It takes help from a technician to make it work in Windows XP, said Terri Flam, director of circulation consulting at CDS.
Flam claims the model is accurate and handles unique situations well, such as transfers, price changes and advance renewals. In addition, the modeling staff analyzes historical data and REACT (a reporting system at CDS) changes and fills in information for clients, if they are CDS fulfilled, via the "enrichment process." Non-CDS fulfilled clients also can use the model.
Aimed at magazines with circulation exceeding 100,000, CDS model clients include Condé Nast, Meredith, Consumers Union, Forbes, Nickelodeon, Gruner+Jahr and Disney. Fifty-five users (90 titles) use it.
In the future, Flam wants more information brought in from fulfillment so the client has less to input, does less number crunching and more strategy. The yearly initialization fee is $1,200 for factored model feeds and $600 for un-factored (non-CDS fulfillment clients), plus an $800-a-month licensing fee, $500 a month for each additional title and monthly automatic update fees of $150 to $200.
CircPlanner PLUS. Publishers of all sizes, including those publishing up to 52 issues a year, use this Windows-based model from Kable Fulfillment Services Inc.'s Colorado group. According to Jodi Smith, a technical product representative for the Colorado group, it's the only model that analyzes historical information to get accurate assumptions. Updates are monthly and are handled automatically, with no need to re-key, whereas most update yearly. It's a circulation-driven model and doesn't contain a full P&L, notes Cindy Weathers, also a technical product representative.
Sixty publications from 15 publishers use the 6-year-old model, which is available only to Kable-fulfilled clients at this point.
IMT. Delivering a full magazine P&L and cash flow, IMT handles detailed, accurate what-if scenarios. According to owner and developer Jim Tucker, since IMT is spreadsheet-based, it is less flexible than a database model and can be hard to set up. Database models are easier to set up but not as good at projecting, he said.
Reader's Digest, Martha Stewart Living, IDG and The Economist use the IMT model. IMT works with bigger publishers that need customized solutions. Tucker claims the other models are tough to customize and more cookie-cutter. Costs vary by size, ranging from $5,000 to $15,000 annually per title.
Tucker predicts future models will understand true LTV of a subscriber analyzing transactional data, not just source data.
Ladd. The SMARTmodel is a complete circulation and financial model that models all circulation, does a full P&L and lets publishers consolidate P&Ls from multiple publications into one report. Users can factor in new accounts and revenue streams tied to circulation, such as book sales.
Few come close to doing that, said Jim Jankowski, vice president of product development of Ladd. Publishers use the model for strategic planning, to get the big picture, he said. On the other hand, loading data electronically can be better and easier with some of the other models.
Agora publishing, American Girl, Burda-Rizzoli, Christianity Today, Color of Sports, Entrepreneur, Curtco Media, Kalmbach Publishing, Smithsonian and Taunton Press use SMARTmodel. Depending on the number of titles and users, costs range from $2,400 to $24,000 annually.
In six to nine months a new version will be released that makes loading of information almost transparent. Jankowski anticipates that data will automatically be updated daily.
A model named "Hal" that speaks to you in English. I envision next-generation models that take the Ladd concept even further. Using today's powerful computer systems, these models would be built around fulfillment company data that automatically talk to and update the model daily. Rather than the modeler inputting order and cash flow patterns, response rates and the like, the model would automatically recalculate itself as new data arrive based on instructions from the modeler.
For example, using a menu format, the program would ask questions, the modeler would respond, and get immediate answers with a few clicks. One question the model would ask is, "Do you want me to build your renewal receipt pattern based on your last six, 12, 18 or 24 months of actual history at the fulfillment company?" The modeler would make a choice and away we go.
Same with response rates, pay-up rates and other flow patterns. And if you ran the model with one of these statistics set to look at, say, the past six months, you could change your mind, click a drop-down menu and select 12 months instead, and have a new result in moments. This type of change would take days to do today and requires much expertise to do correctly.
The model would know how to build the patterns and statistics and how to use them to create accurate projections. There would be no chance of the modeler misinterpreting the data from the fulfillment house reports or using it in the wrong way or in the wrong place. And because virtually all the statistics and patterns in the model theoretically could be updated every time the fulfillment house does an update (usually weekly), the model would be even more accurate.
Independent publishers with smaller staffs particularly would benefit from a more accurate modeling tool that requires little time, effort or knowledge from the modeler. Now, somebody just needs to build it.