If the startup is VC- or angel-funded, figure out what their runway is. You may just want to ask them what they think it is, the revenue assumptions they've us

If the startup is VC- or angel-funded, figure out what their runway is. You may just want to ask them what they think it is, the revenue assumptions they've used to understand that, and what kind access to funding they have beyond what's in the bank (don't count on this too much, however). With this information in hand, estimate the Probability that they'll last as long as you'll need them, the Cost of switching to a different product down the line, and the Value you derive from going with a startup vs. the "safe" option. Pick the startup that maximizes (and has a positive) V - (1-P)*C. (Feel free to substitute your own more elaborate ROI calculation.) You may be able to reduce your risk by adding a contract term to the effect of the following: if the startup goes out of business or discontinues the service, they'll agree to sell you at a specified price the relevant code, systems, license, etc. so you can continue operating the service in-house. Of course, this may not be relevant, depending on what exactly the startup's services are, and depending on your corporation's ability to support the service, and a lot of other legal issues worth examining with a lawyer.