Evan Schuman's StorefrontBacktalk

But bolting on capabilities built by outsiders to work with Target’s existing systems? That makes sense. It’s fast and practical. The downside: It’s a big outsourcing expense that makes the IT budget balloon very suddenly.

However, these are one-time expenses, or at least they can be. Once the development work is done, that balloon can deflate as fast as it blows up.

That kind of bursty IT spending sounds all wrong in terms of traditional systems development. Partly that’s because traditional development happened inside an IT department, so projects were constrained by IT staff size. Hand off whole projects to outsiders (and then watch them like hawks, because their bottom-line interest is their profits, not yours) and that constraint goes away. IT projects are constrained by budgets, not internal staff.

This approach doesn’t work for all IT projects—some foundational projects have to happen more slowly and aren’t really done until well after they’ve gone into production (think Target’s exceedingly bumpy first few months after it cut itself loose from Amazon in 2011—remember “Missoni Tuesday”?). But for well-defined bolt-ons, bursty IT spending is very possible.

It also may be very practical. Bringing IT development to a screeching halt after a big development push gives IT and the rest of the business a chance to absorb what’s been developed. That likely will mean working the technical bugs out of the in-store pickup system, for example—as well as finding the unexpected problems with the in-store part of the process. Users will have to figure out what they still aren’t getting from the predictive analytics. The E-commerce logistics product will have to be used to find out what isn’t quite right.

Then, once everything is stable and business-side users have identified what they still need—and new technology has created new options—it may be time for another burst of project spending.

That bursty, feast-or-famine approach isn’t the only way to develop merged channel (aka omnichannel) retail systems—Macy’s (NYSE:M), for one, runs more of a slow-and-steady merged channel race. (It turns out that Target is running that way, too.) But bursty IT spending could conceivably be done without gutting the IT department, and it might have given Target a project rhythm that would let it make big leaps forward followed by pauses to consolidate.

Which way—bursty or slow-and-steady—works best? We’re not likely to find out from Target. But even if some other chain chooses to go that route tomorrow, you can expect to wait at least five years for an answer. Short of a major catastrophe, that’s how long it would take to see whether such big leaps—in both system capability and budgets—would be a success.