If in a predictive dialing campaign (or in the Cookbook using predictive mode) when you use the overdial or no answer controls (see Overdial Settings) to reduce performance, the abandoned calls determined by Oceanic® will undershoot the abandoned call target that you set. The greater the amount you slug performance by, the greater the undershoot.
And yet in real life you will still be expecting the total number of abandoned calls to approximate the target you set in the Campaign Wizard. Because in real life a dialer will perform at the slugged rate you set, and still deliver a rate of abandoned calls similar to your target, and not to the reduced level determined by Oceanic®. We call this difference between the calls determined by Oceanic®, and those implicit in the target, dialer 'makeup' calls. For reporting purposes on the Campaign Summary, the two types of call abandonment figure are added together, and should approximate your target.
See Oceanic® Performance v. Dialer Performance to understand why dialers in practice may dial less efficiently that Oceanic®, and hence why a performance adjustment is required to Oceanic®, leading to makeup calls. If we didn't acknowledge these calls as abandoned, a campaign would treat them as other kinds of call outcome, live calls and so on; and more live calls would be recorded than expected in practice, overstating the length of a campaign.
Incidentally, Oceanic® has no way of telling you exactly how makeup calls occur; to do so would mean knowing, amongst other things, the inner workings of whatever predictive dialing engine is being modeled, and that is clearly quite impractical. And it is not necessary either, when it comes to benchmarking.