Saturday, October 06, 2007

Cost of Simulations

The production of simulation-based learning environments can be an incredibly expensive proposition. Some of my friends in the video game development business tell me that a typical lower quality video game that is designed for around forty hours of game-play will, at minimum, cost between $2-20M dollars to produce (the cost can be much greater for higher quality games). Now consider the potential costs of creating a high quality simulation-based learning environment that provides forty hours of instruction in the context of a richly rendered 3D virtual environment.

It would seem that the production costs of high quality simulation-based learning environments might stand in the way of a future where students could learn and interact within detailed and highly functional online worlds. After all, there aren't many institutions that can afford to pay $2-20M for production of an entire online course that would need to support support forty hours of engaging contact with simulation-based content. Clark Aldrich's very informative blog on serious games and simulations recently listed the costs of producing simulation-based learning environments. What Clark's list really underscores is that any widespread use of simulation-based learning environments over multiple courses will be a costly proposition unless we can leverage a different set of dynamics for their construction.

Branching story:
Simulation in which students make a series of decisions via a multiple choice interface to progress through and impact an event.

Custom short (Less than 10 minutes) (perpetual site license): $30K
Custom medium (Between 10 minutes and 30 minutes) (perpetual site license): $100K
Custom long (Between 30 minutes and 2 hours) (perpetual site license): $500K
Off-the-shelf short (per user): $30
Off-the-shelf medium (per user): $100
Off-the-shelf long (per user): $500

Interactive spreadsheet:
Simulation in which students typically try to impact critical metrics by allocating resources along competing categories and getting feedback of their decisions through graphs and charts.

Custom short (Less than 1 hour) (perpetual site license): $30K+
Custom medium (Between 1 hour and 4 hours) (perpetual site license): $100K+
Custom long (Between 4 and 8 hours) (perpetual site license): $500K+
Off-the-shelf short (per user): $30*
Off-the-shelf medium (per user): $100*
Off-the-shelf long (per user): $500*

Small, easy-to-access game built to be simple and addictive, which often focuses on mastering an action and can provide awareness of more complicated issues.

Custom short (5 minutes) (perpetual site license): 10K
Custom medium (10 minutes) (perpetual site license): 15K
Custom long (30 minutes) (perpetual site license): $40K
Off-the-shelf short (per user): n/a
Off-the-shelf medium (per user): n/a
Off-the-shelf long (per user): n/a

Virtual labs:
A series of challenges/puzzles to be solved using on-screen representations of real-world objects and software

Custom short (30 minutes) (perpetual site license): $30K
Custom medium (1 hour)(perpetual site license): $75K
Custom long (4 hours)(perpetual site license): $150K
Off-the-shelf short (per user): $10
Off-the-shelf medium (per user): $30
Off-the-shelf long (per user): $100

Real-time, often 3D sims that encourages participants to repeat actions in high fidelity situations until the skills become natural in the real-world counterpart

Custom short (1 hour)(perpetual site license): $100K+
Custom medium (5 hours) (perpetual site license): $500K+
Custom long (20 hours) (perpetual site license): $1M+
Off-the-shelf short (per user): $100*
Off-the-shelf medium (per user): $400*
Off-the-shelf long (per user): $1000*

+ plus cost of facilitation
* including cost of facilitation

Can the dynamics of open source social software solve the problem of high cost simulation production for education? I think that it can. All we need is the right tools and the collective will to use them. After all, the MediaWiki-based Wikipedia doesn't cost anyone much money at all (in 2006, Wikipedia's Internet hosting costs came to $189,631). The cost of Wikipedia's real value - that of its content - has been distributed across its many contributors. What we need to solve the cost problem for simulation-based learning is a good 3D wiki-like technology that could be used to "evolve" multi-authored and highly functional simulation-based learning environments at low institutional cost.


Clark Aldrich said...

When asked to summarize what most teachers wanted out of simulations, I have said: Most of the teachers I know want approaches to teach the exact same material as before, produce better test scores in a way for which they can take credit, is easier for them then the current methods, but does not replace them or result in increased work in new areas (even if more than offset by the productivity gains of the new approach). And oh yes, and it has to be free.

You know better than I if this is fair, and if this will get us to a better place.

Anonymous said...

Half of the cost of developing a modern video game is for art. A lot of the cost drivers for video game art are less relevant for teaching tools:

1) It's probably not important to get a distinctive "look" for the tool, eliminating the need to iterate over minor details.

2) Using standard geometry & rigging for characters and objects is probably not a problem.

3) The environments are probably smaller

4) There's little need to texture surfaces that the user is unlikely to see.

5) There's little need to create similar environments that look very different (spaceship corridors vs. building corridors vs. tunnels)

R Hayes

Anonymous said...

I think you hit the nail on the head when you said we must “leverage a different set of dynamics” for simulation construction. Those dynamics will undoubtedly prove to be new technology with which to build simulations. As most companies construct them now, simulations rapidly accrue exorbitant costs, quickly leaping out of the scope of training budgets. But I recently read of a new technology titled SimPort that I believe will change the way the industry does simulations. SimPort is a framework developed by a Colorado-based training company, and it enables instructional designers to input a company’s unique features into an existing platform to create a customized simulation experience. The beauty of it – half the cost. Because the basic platform is already there, the simulation does not have to be built from scratch, freeing up some of those training dollars – not to mention a significant chunk of development time. And the framework means it’s easy to alter and maintain as needed. Instead of building an enormous branching tree, SimPort connects a number of smaller pieces – branching trees, spreadsheets, etc. – to create a comprehensive, interconnected simulation. As many in the training industry will agree, this customization makes for an extremely effective learning environment, a valid training experience. There’s been a lot of talk about the Big Four Fears recently (the four concerns about simulations that some executives cite: cost, development time, maintenance, and validity). Looks like with the advent of technologies like SimPort, these concerns will soon fade off the scene.

And, FYI: Website of the company who made SimPort can be found here:

M. Jones