Standard CBC-looking holdouts may be included in the ACBC survey, if that's what you would like to do.
If you own the CBC license, then you'll need to copy-and-paste the attribute level lists into the CBC interface (for attributes and levels). So, essentially, you are adding a CBC study within your SSI Web study (that already includes an ACBC section).
Next, tell CBC to create Fixed Tasks only (no random tasks), and specify exactly which levels to show in each task.
Then, put these fixed CBC tasks somewhere in your SSI Web study.
If you don't own the CBC license, then you can use "Free Format" question types in SSI Web, and with your own HTML programming, you can create CBC-looking tasks to serve as holdouts.
If you have sales data, then that serves as validation. But, often sales data don't line up (or should be expected to line up) with conjoint predictions. That's because there often are things that are different about the real world than the conjoint simulation world:
a) distribution: are all products equally available in the marketplace?
b) time on the market: has each product been available long enough to reach its potential
c) advertising/awareness: has each of the products achieved deep awareness? Conjoint makes all products known and available to everybody
d) effectiveness of sales force: even if a product is good, a bad sales effort could lead to diminished sales