OK Todd, I'll bite.
My experience (obviously) is of one who has implemented inkDROP for clients, not one who actually uses it in day-to-day production....so take what I say with that in mind. Naysayers might conclude that disqualifies me from even commenting on inkDROP....but I'd like to think I've seen more inkDROP implementations, both good and not-so-good, than anyone in a production environment is likely to see.
First, we should distinguish between the RAMpage option/feature that is inkDROP and device link profiles generally. What I THINK you want to know about is how device links are working generally and not inkDROP specifically. If it's specifically about inkDROP, all I can say is that it's one of the better implementations for using static device link profiles. In fact, RAMpage (in inkDROP v2) has done some amazing things in working around some of the limitations inherent in using static device links (DVLs). Protection of special colors, protection of imposition elements outside the page boundaries and coming up with an ingenious work-around for preserving TRUE 100% black overprint values are just a few of the things you get with inkDROP that go beyond the typical handling or conversions using static DVLs (I say "static" because there's a distinction of what can be done using static DVLs vs. "dynamic" DVL workflow products). But moving on....
First off, in terms of ink reduction/savings, this is NOT so much a function of inkDROP as it is of the software that is used to create the DVLs. It's fair to say that RAMpage inkDROP really has nothing to do with ink reduction or even color quality but has everything to do with how the DVLs are created. Bottom line, if you're looking for maximum ink reduction, prepare to spend some serious money on the software used to create the DVLs. Entry fee is at least $5,000 for decent DVL software and you can expect to pay easily 2-4x that for some of the higher end products.
Ink savings is a difficult thing to measure/track and it's been my impression that some folks that are interested in that aspect alone of inkDROP+DVLs really have no way to track it accurately. Having said that, I've heard from customers that have measured 30-40% ink savings on certain jobs. The downside of this kind of ink savings is that it requires the press operator to re-learn a lot of what he knows about ink key adjustment. Those levels of ink savings can even create problems at the press console in terms of fine ink control....set up the DVL for very aggressive ink savings and you might find the ink key adjustments on the console functioning more like on/off switches on the CMY ink channels! For aggressive ink savings, I would strongly suggest this only be used on relatively new presses with auto-scanning or in-line density control and closed-loop color controls.
In terms of *color* (from my perspective, what most should be thinking about!), inkDROP + DVLs can do wonders, especially if you're using the G7 method for GRACoL and SWOP. If the G7 plate/press curves can get you an 80% match to a "certified" GRACoL proof, DVLs can get you into the 90-95% color matching range PROVIDED you have good process control. In terms of "delta E" matching, you should see a drop from the typical 2-3 dE average you get with plate curves alone, down to the 1-1.5 dE range by using a DVL made from a custom ICC profile of your press. If plate curves put you in the ballpark, DVLs will usually get you someplace between 3rd base and home plate.
That's not to say color is always perfect. I've seen a few "failures" (results not much better than using plate curves). The times I've seen this, it usually comes down to in ability to replicate results from one press run to the next and..increasingly....use of paper stock that falls far outside GRACoL/SWOP specs for paper white L*a*b*. Specifically, I've seen cases where pastel colors simply do not match all that well and I believe it's generally a combination of how far off the paper white is from GRACol specs (usually optical brighteners are the culprit) and how the relative colorimetric ICC rendering is designed to work. Basically, the further off your paper is from the GRACoL specification, the more trouble you're going to have with lighter colors where paper tint becomes a major influence. In proofing this is rarely an issue since we typically use absolute colorimetric for accurately rendering the entire color space, including paper white (we tint it on the proof if we have to). Going to press however, absolute colorimetric is not an option so we're back to being at the mercy of our press stock for the final rendering of color.
Anyway, I've babbled on long enough.....I guess I would just say for those folks looking at inkDROP+DVLs as some sort of magic bullet for either color or ink savings...or both...to reset your expectations based on how well you already are able to control your color in the pressroom. If you're not using closed loop color or still using hand-held instruments, expect an improvement with DVLs...but don't expect miracles. For those with the good control in the pressroom AND using higher quality papers that are already very close to GRACoL/SWOP specs, you should expect to get very good results.