Why do DSLR shutter life ratings vary so much between camera models?

Shutter life is a statistically distributed value with one model of camera giving widely varying results, as the graphs and tables on the site that you referred to show.

Here "shutter" is liable to include anything that moves mechanically when the shutter is operated that would fatally stop camera operation if it 'failed'. Failed could include dead or walking wounded but I'll avoid fine details.

More than anything else, shutter life is a design decision by the camera manufacturer.

Mechanical aspects which influence it may include mirror mass and speed, acceleration and deceleration and end stop impact forces, blind masses and speeds, accelerations, impacts, various torques applied to lever mechanisms, bearing types and sizes and quantities and loadings, and lubrication and bearing materials. And more.

Major camera manufacturers have been playing these games for many decades and are well aware of the tradeoffs of price and cost and noise and max speed and more. Ultimatly they are able to make marketing driven decisions to provide a given level of user experience.

The page you provided - Olekkin shutter life expectancy database suggests that the application is uneven. Some early Canon DSLRs (300, 350 ?) initially had shockingly low claimed shutter actuation lifetimes and these were substantially upgraded on subsequent models. A low end camera will usually have a claimed life of about 100,000 actuations and from memory a Nikon D3 claims 300,000 actuation lifetime.

However, the stats on the referenced website suggest some quite inconsistent results.
Lifetimes to about 30 failures based on tables at right for Nikons.
D5000 - 750,000+ !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
D3 - 150,000 - 200,000
D300 - 150,000 - 350,000

NB Sony seem to be very variable - eg Sony A900 - all over the map - 100,000 - 3,000,000+

But, the answer to "Why do manufacturers allow lifetimes to vary so widely?"is "because they can".