Faced with this latest period of tumult and uncertainty in the tech industry (and the economy in general), how should CIOs respond? Almost certainly, some emerging companies with promising products or services, will run through their cash or otherwise fail to achieve reasonable scale.

Or they’ll fall victim to first mover dis-advantage and be eclipsed by a late-comers just as FaceBook trounced MySpace or Google sank AltaVista.

Establishment incumbents are equally vulnerable in times of sweeping change; witness the unhappy fates of Data General, Datapoint, Digital Equipment, Wang, and so many more.

Emerging companies offer too little visibility to provide their customers sufficient early warning to cut their losses. So here, the prudent response is to adopt a defensive stance from the first purchase forward:

1. Focus on “precise solutions to pressing problems,” as a top CIO once advised. In other words, avoid companies with grandiose roadmaps that go well beyond the focal issue -- the need for a unique solution to a functional shortfall in current software.

2. Postpone tight integration between the new product or service and the enterprise infrastructure or applications base until enough time passes to provide sufficient visibility. Yes that may require postponing full benefits or temporary installing costly work-arounds. But either will be more economic than a forced disengagement.

3. Avoid start-ups that seem overly intent on buying new business by cutting prices, concentrating far less on developing a profitable business model. Put another way: "Hope" is not a strategy. If the business doesn’t make economic sense it won’t exist for long.

4. Accept companies with a product or service that’s broadly attractive enough to make the business a reasonable acquisition targets should their efforts fail. Of course, this approach carries another form of risk since the acquirer may let the company wither or impose such Draconian maintenance costs or service degradation as to make the original service essentially unviable.

5. Develop an exit plan even before ink dries on the initial contract. If the company fails, what steps will be required to get clear of the encumbrance? Can some of these steps be taken in advance?

For establishment IT companies, three early warning signs have recurred in failure after failure over the past 30 years:

• The company has not succeeded in a program of continuous reinvention, as have Apple, IBM, Google, or Oracle.
• Management speaks of new strategies, products, and services but the company's field representatives are either unclear on or ignorant of them. (Field forces are remarkably impervious to change, be it absorbing new messages or taking on new products. But management's failure to deal with organizational resistance to change has often harbingered the firm's collapse.)
• The company's CEO is a founder who steadfastly postpones or resists grooming or even naming a successor. (The founder's resistance grows during a business down turn that . Postponement becomes more painful the longer the situation lasts, and threatens to tarnish his or her the founder's reputation. Postponement becomes more painful the longer the situation lasts until it's too late for a new CEO to navigate a turnaround.)

If these signals occur, it’s time to start making contingency plans for a graceful disengagement or at least reducing forward commitments.

Ernie von Simson is the senior partner in the CIO Strategy Exchange and the author of The Limits of Strategy an inside analysis of the success and failure of IT companies over the past 30 years.